Coolscan 8000 vs. 9000?

[MXP]

Are there any significant differences between these two scanners?

Is it same optic?

I know the 9000 is faster in scanning then the 8000.

What is the big difference between ICE 3 and 4?


Max

[MXP]

OK.
I found out the the 9000 has 16 bit for each color vs. 14 for the 8000.
Also the Dmax is a little higher. Maybe I should go for the 9000.

Max

"MXP" <***@post11.tele.dk> skrev i en meddelelse
news:42b73106$0$75101$***@dread14.news.tele.dk...
> Are there any significant differences between these two scanners?
>
> Is it same optic?
>
> I know the 9000 is faster in scanning then the 8000.
>
> What is the big difference between ICE 3 and 4?
>
>
> Max
>

[Lassi Hippeläinen]

MXP wrote:

> OK.
> I found out the the 9000 has 16 bit for each color vs. 14 for the 8000.
> Also the Dmax is a little higher. Maybe I should go for the 9000.

If those numbers are from marketing brochures, they are meaningless. They
can promise more D, because nominally there are more bits per sample, but
in real life the two additional bits in the least significant end are pure
noise.

-- Lassi

[Q.G. de Bakker]

Lassi Hippeläinen wrote:

> > OK.
> > I found out the the 9000 has 16 bit for each color vs. 14 for the 8000.
> > Also the Dmax is a little higher. Maybe I should go for the 9000.
>
> If those numbers are from marketing brochures, they are meaningless. They
> can promise more D, because nominally there are more bits per sample, but
> in real life the two additional bits in the least significant end are pure
> noise.

Nevertheless, if you can find an affordable 9000 get that one. If not, an
8000 is quite good.

[MXP]

I could get a new 8000 for about EUR 2000 and a 9000 is about EUR 2600.
In US you can maybe find it cheaper......but I am not living there......

Can anybody tell me if the standard scanning software from Nikon is good or
do I need SilverFast?

What print qualites can I achive from a perfect exposed 6x6 Velvia when it
is scanned
with a 8000 or 9000 compared to e.g. a 16MP SLR or 16-22MP Imacon digiback?
Printsize from A4 to A3.


Max


"Q.G. de Bakker" <***@tiscali.nl> skrev i en meddelelse
news:42b84b18$0$73073$***@news.wanadoo.nl...
> Lassi Hippeläinen wrote:
>
>> > OK.
>> > I found out the the 9000 has 16 bit for each color vs. 14 for the 8000.
>> > Also the Dmax is a little higher. Maybe I should go for the 9000.
>>
>> If those numbers are from marketing brochures, they are meaningless. They
>> can promise more D, because nominally there are more bits per sample, but
>> in real life the two additional bits in the least significant end are
>> pure
>> noise.
>
> Nevertheless, if you can find an affordable 9000 get that one. If not, an
> 8000 is quite good.
>
>

[Q.G. de Bakker]

MXP wrote:

> I could get a new 8000 for about EUR 2000 and a 9000 is about EUR 2600.
> In US you can maybe find it cheaper......but I am not living there......

Neither am i, else i had a new 9000 already...

> Can anybody tell me if the standard scanning software from Nikon is good
or
> do I need SilverFast?

I never used anything but Nikon Scan, so i don't know how it compares.
But i think the Nikon software is quite ok.

> What print qualites can I achive from a perfect exposed 6x6 Velvia when it
> is scanned
> with a 8000 or 9000 compared to e.g. a 16MP SLR or 16-22MP Imacon
digiback?
> Printsize from A4 to A3.

A 8000/9000 scan of, say, a 6x6 is 81 MP. And all are good, usefull.
So you can get really gigantic prints that look beautiful.

I believe scanned film requires more post-processing than direct digital
capture, and at print sizes of up to A3, a 16-22 MP digital SLR/back will do
very well too.
Not to mention the rather long time scanning takes (there, i just did
mention it. ;-))
So if that's your print size limit, perhaps workflow related considerations
would suggest direct digital capture to be the 'better' alternative?
On the other hand, going for direct digital capture often requires extra
investment (very short lenses), which may make scanning film themore
attractive option.
;-)

[MXP]

"Q.G. de Bakker" <***@tiscali.nl> skrev i en meddelelse
news:42b875a3$0$3890$***@news.wanadoo.nl...
> MXP wrote:
>
>> I could get a new 8000 for about EUR 2000 and a 9000 is about EUR 2600.
>> In US you can maybe find it cheaper......but I am not living there......
>
> Neither am i, else i had a new 9000 already...
>
>> Can anybody tell me if the standard scanning software from Nikon is good
> or
>> do I need SilverFast?
>
> I never used anything but Nikon Scan, so i don't know how it compares.
> But i think the Nikon software is quite ok.
>
>> What print qualites can I achive from a perfect exposed 6x6 Velvia when
>> it
>> is scanned
>> with a 8000 or 9000 compared to e.g. a 16MP SLR or 16-22MP Imacon
> digiback?
>> Printsize from A4 to A3.
>
> A 8000/9000 scan of, say, a 6x6 is 81 MP. And all are good, usefull.
> So you can get really gigantic prints that look beautiful.

I have just scanned a 6x6 using my Epson 3200. It ended up with a 250Mb
file (16 bit pr. color). I need more RAM for my computer when working in
Photoshop with that kind to file sizes. The 3200 scans are very soft and
trying
to correct with USM in Photoshop adds a lot of grain noise. I hate USM but
it is necessary when working with digital pictures. The film is perfectly
sharp.

>
> I believe scanned film requires more post-processing than direct digital
> capture, and at print sizes of up to A3, a 16-22 MP digital SLR/back will
> do
> very well too.
> Not to mention the rather long time scanning takes (there, i just did
> mention it. ;-))
> So if that's your print size limit, perhaps workflow related
> considerations
> would suggest direct digital capture to be the 'better' alternative?
> On the other hand, going for direct digital capture often requires extra
> investment (very short lenses), which may make scanning film themore
> attractive option.
> ;-)

My pictures is just for a living. So I can wait for the scans. But I like
the
smaller files from a digital SLR or back :-)
>
>

[David J. Littleboy]

"MXP" <***@post11.tele.dk> wrote:
>
> I have just scanned a 6x6 using my Epson 3200. It ended up with a 250Mb
> file (16 bit pr. color). I need more RAM for my computer when working in
> Photoshop with that kind to file sizes. The 3200 scans are very soft and
> trying
> to correct with USM in Photoshop adds a lot of grain noise. I hate USM but
> it is necessary when working with digital pictures. The film is perfectly
> sharp.

Try the following.

0) Apply a _very light_ NeatImage or Noise Ninja operation to your original
scan file.
1) Sharpen lightly (150%, 0.5, 2), or maybe more, or maybe none.
Experimentation required here.
2) Downsample to 1800 dpi
3) Sharpen aggressively.
4) Print at 300 dpi (6x).

> My pictures is just for a living. So I can wait for the scans. But I like
> the smaller files from a digital SLR or back :-)

I like downsampling scans. One really doesn't lose any significant
information, and the files are easier to work with.

David J. Littleboy
Tokyo, Japan

[MXP]

"David J. Littleboy" <***@gol.com> skrev i en meddelelse
news:d9a7rs$v0c$***@nnrp.gol.com...
>
> "MXP" <***@post11.tele.dk> wrote:
>>
>> I have just scanned a 6x6 using my Epson 3200. It ended up with a 250Mb
>> file (16 bit pr. color). I need more RAM for my computer when working in
>> Photoshop with that kind to file sizes. The 3200 scans are very soft and
>> trying
>> to correct with USM in Photoshop adds a lot of grain noise. I hate USM
>> but
>> it is necessary when working with digital pictures. The film is perfectly
>> sharp.
>
> Try the following.
>
> 0) Apply a _very light_ NeatImage or Noise Ninja operation to your
> original scan file.
> 1) Sharpen lightly (150%, 0.5, 2), or maybe more, or maybe none.
> Experimentation required here.
> 2) Downsample to 1800 dpi
> 3) Sharpen aggressively.
> 4) Print at 300 dpi (6x).

Good idea......I will try that.
Do you find that the prints get even better by doing this?
Normally I just send all the information (e.g. 200 Mb) to the
jet ink printer.....even if the picture is only A4 size or less.
Then the printer or printer driver has to handle all the information.

Is Epson 2880 dpi about the same as 300 or 600 "real" dpi?
I have never found out how the 2880 dpi should be interpreted.
A pixel or dot is for me a set of R, G and B.....or in the Epson case
a set of all 7 colors or how many they have. But I think Epson multiply
with no. of colors?


>
>> My pictures is just for a living. So I can wait for the scans. But I like
>> the smaller files from a digital SLR or back :-)
>
> I like downsampling scans. One really doesn't lose any significant
> information, and the files are easier to work with.
>
> David J. Littleboy
> Tokyo, Japan
>
>

[Bill Hilton]

>MXP writes ...
>
>Is Epson 2880 dpi about the same as 300 or 600 "real" dpi?

Not sure what you mean by "real" dpi, but the printer internally will
resample to 720 ppi so if you send it a larger file the software will
downsample and if smaller then it will upsample. Usually an even
multiple like 240 or 360 ppi is considered ideal, though with real
images (as opposed to test targets) 300 or whatever also works well.
Unlike Littleboy I wouldn't downsample just to have the printer
upsample though.

>I have never found out how the 2880 dpi should be interpreted.

This is the granularity of the dots laid down by the inkjets, and it
has nothing to do with the input file size (which should be expressed
in ppi or pixels per inch, vs dpi for the output). The output dpi is
chosen based on the paper type used, not the input file size rez
(though of course it makes little sense to print 2880 dpi with a very
small input file, say 100 ppi or so). For example, cheap uncoated
paper for text can typically handle only 180 dpi, most matte and
watercolor papers absorb some of the ink but not as much as uncoated
papers so can handle 720 or 1440 dpi, and papers with glossy coatings
can accept typically 1440-2880 dpi (higher on some models) because of
the coating.

[David J. Littleboy]

"MXP" <***@post11.tele.dk> wrote in message
news:42b9e5f8$0$82713$***@dread14.news.tele.dk...
>
> "David J. Littleboy" <***@gol.com> skrev i en meddelelse
> news:d9a7rs$v0c$***@nnrp.gol.com...
>>
>> "MXP" <***@post11.tele.dk> wrote:
>>>
>>> I have just scanned a 6x6 using my Epson 3200. It ended up with a 250Mb
>>> file (16 bit pr. color). I need more RAM for my computer when working in
>>> Photoshop with that kind to file sizes. The 3200 scans are very soft and
>>> trying
>>> to correct with USM in Photoshop adds a lot of grain noise. I hate USM
>>> but
>>> it is necessary when working with digital pictures. The film is
>>> perfectly sharp.
>>
>> Try the following.
>>
>> 0) Apply a _very light_ NeatImage or Noise Ninja operation to your
>> original scan file.
>> 1) Sharpen lightly (150%, 0.5, 2), or maybe more, or maybe none.
>> Experimentation required here.
>> 2) Downsample to 1800 dpi
>> 3) Sharpen aggressively.
>> 4) Print at 300 dpi (6x).
>
> Good idea......I will try that.
> Do you find that the prints get even better by doing this?

Yes. If only because I can get the sharpening and other adjustments the way
I want them.

> Normally I just send all the information (e.g. 200 Mb) to the
> jet ink printer.....even if the picture is only A4 size or less.
> Then the printer or printer driver has to handle all the information.

Yes. If you use Qimage, it will do much of this for you.

> Is Epson 2880 dpi about the same as 300 or 600 "real" dpi?

It depends on the image you are feeding the printer. It's a long story.

> I have never found out how the 2880 dpi should be interpreted.

The 2880 is advertizing blather.

> A pixel or dot is for me a set of R, G and B.....or in the Epson case
> a set of all 7 colors or how many they have. But I think Epson multiply
> with no. of colors?

Maybe, maybe not. The important thing is that if you feed your printer a
quality image at 360 ppi, it'll look better than if you feed it a quality
image at 240 ppi.

David J. Littleboy
Tokyo, Japan

[Gordon Moat]

MXP wrote:

> . . . . . . . . . .
>
> I have just scanned a 6x6 using my Epson 3200. It ended up with a 250Mb
> file (16 bit pr. color). I need more RAM for my computer when working in
> Photoshop with that kind to file sizes. The 3200 scans are very soft and
> trying
> to correct with USM in Photoshop adds a lot of grain noise. I hate USM but
> it is necessary when working with digital pictures. The film is perfectly
> sharp. . . . . . . . .

Using PhotoShop, there are some other methods for sharpening images. Rather
than using USM on the entire image, you can apply it selectively to the
channels that seem soft. You can also use a Gaussian Blur or Smart Blur on one
or more channels, then sharpen. Another choice is using the Maximum, Minimum,
or High Pass filters an a different layer, then blending the layers to get
sharper results in the final image. There are usually at least three methods
of doing anything in PhotoShop, though some combination might be needed for
each particular image.

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[Bill Hilton]

>MXP writes ...
>
>The 3200 scans are very soft and trying
>to correct with USM in Photoshop adds a lot of grain noise.

Google on "edge sharpening" and you'll probably find steps for how to
avoid this ... basically run the 'find edges' filter on a b/w copy of
the image, invert it and run a couple of filters to smooth it and then
use this as a mask with the edges active and the clear toned areas
masked out ... then you can run USM very aggressively (say 400%, radius
.6 - 1, threshold 0 since the mask acts as the threshold) and sharpen
the edges without overly sharpening the other areas. I have an action
that does this in LAB mode if you want to email me for a copy. Or
there are a couple of companies selling plug-ins that do this.

Bill

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote:
>
> A 8000/9000 scan of, say, a 6x6 is 81 MP. And all are good, usefull.
> So you can get really gigantic prints that look beautiful.

You can't enlarge film any more with a scan than you can with any other
means.

To my eye, film begins to break down when you enlarge anything over 8x,
whatever the means you use to enlarge.

> I believe scanned film requires more post-processing than direct digital
> capture, and at print sizes of up to A3, a 16-22 MP digital SLR/back will
> do
> very well too.
> Not to mention the rather long time scanning takes (there, i just did
> mention it. ;-))

The 16-22MP cameras/backs are still outrageously expensive.

> So if that's your print size limit, perhaps workflow related
> considerations
> would suggest direct digital capture to be the 'better' alternative?
> On the other hand, going for direct digital capture often requires extra
> investment (very short lenses), which may make scanning film themore
> attractive option. ;-)

Other than the 1Dsmk2, I really don't recommend direct digital. Most of the
MF backs lack an AA filter, and aliasing artifacts are really nasty with
Bayer sensors.

My current opinion here is that if you crop 6x6 to 1:1.414, then you won't
see much difference between your best 4000 dpi and the 1Dsmk2. The film
_should_ edge out the 1Ds2, but in practice, it won't.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> > A 8000/9000 scan of, say, a 6x6 is 81 MP. And all are good, usefull.
> > So you can get really gigantic prints that look beautiful.
>
> You can't enlarge film any more with a scan than you can with any other
> means.

That's right. And you can get really gigantic prints that look beautiful
from film too.
;-)

The worry however might be if the scanner can get as much out of film as you
can any other way. Not whether the scanner can get more out of film than
there is to begin with.
And, yes, it can do the first.
I think the 8000/9000 4000 ppi resolution is pretty much right.

> To my eye, film begins to break down when you enlarge anything over 8x,
> whatever the means you use to enlarge.

Well, yes, and no.
'No', i do not find because 8x isn't the limit.
Considering that a larger print is usually seen from a larger distance (else
you're so close you can't see the print properly, and you tend to back off),
larger prints look quite well.
For 'indoor' prints 40x is quite doable, without getting horrid results.
Smaller for smaller 'indoors'. I don't know many that 'indoors' that can
accomodate much larger prints. ;-)

And 'yes', because it does indeed not matter how you get to that size.

> The 16-22MP cameras/backs are still outrageously expensive.

I couldn't agree more.

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote:
>
>> To my eye, film begins to break down when you enlarge anything over 8x,
>> whatever the means you use to enlarge.
>
> Well, yes, and no.
> 'No', i do not find because 8x isn't the limit.
> Considering that a larger print is usually seen from a larger distance
> (else
> you're so close you can't see the print properly, and you tend to back
> off),
> larger prints look quite well.

I disagree. For the stuff I'd like to do, people will walk up and look
closer. At which point, I'd like to have the detail they're looking for.
People don't put their noses on large prints, but they'll come within 12 or
15 inches, I'd think. And I want the detail to be there.

> For 'indoor' prints 40x is quite doable, without getting horrid results.
> Smaller for smaller 'indoors'. I don't know many that 'indoors' that can
> accomodate much larger prints. ;-)

Hmm. 40x isn't what I call a "photograph". That's a "poster". And 40x from
film is seriously horrid to my eye. Did you mean something different here?
That's printing a 4000 dpi scan at 100 dpi, and that's a seriously ugly
print.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> > For 'indoor' prints 40x is quite doable, without getting horrid results.
> > Smaller for smaller 'indoors'. I don't know many that 'indoors' that can
> > accomodate much larger prints. ;-)
>
> Hmm. 40x isn't what I call a "photograph". That's a "poster". And 40x from
> film is seriously horrid to my eye. Did you mean something different here?
> That's printing a 4000 dpi scan at 100 dpi, and that's a seriously ugly
> print.

It isn't horrid, really.

I have some 7 ft prints (inkjet, from Nikon 8000 scanned 6x6 negs) hanging
in a conference room in a local hotel, and belive it or not, they do look
good.

And you can come close and even see detail you simply do not see at 'normal'
viewing distances. Though you will then also see that coming close is
perhaps not the best way of seeing these prints.

It's like looking at negs through a magnifier: you will see more detail than
you will see in a print, but you will also see how close that detail is to
grain. Looking at a print at normal distances, you notice the grain less.
But you also not see the finest detail captured.

Anyhow, given the size and the setting, these things do look good, despite
the less-than-300-dpi inkjet prints*.
What seriously makes them look less good is the fact that they are done on
translucent film and have light shone through them from behind. But that's
another story. ;-)

* Done at a local company specialised in designing and building
trade-show/industry-show stands. Wasn't nearly as expensive as a lab would
ask. The benefits of local economy, i guess. ;-)).

[Bill Hilton]

I have an 8000 and agree that 16 bit vs 14 bit and slightly different
dmax are not overly important, though I'd rather have the 9000 (but
I've definitely bought my last film scanner).

>I could get a new 8000 for about EUR 2000 and a 9000 is about EUR
>2600. In US you can maybe find it cheaper

I think the 9000 is $2,000 or so in the USA, didn't know they still
sold the 8000 new.

>Can anybody tell me if the standard scanning software from Nikon is good
>or do I need SilverFast?

Nikon Scan is adequate, I feel. But then I haven't used SilverFast so
I could be wrong.

>What print qualites can I achive from a perfect exposed 6x6 Velvia when
>it is scanned with a 8000 or 9000 ...

Should print quite nicely at 20x20" or maybe even 24x24" to most
people's taste, though it wouldn't be a bad idea to get a scan done if
you can find someone to do it and run the print yourself to see if it's
OK for your tastes. Helps if you know some Photoshop, especially
sharpening and tonal adjustments.

> ... compared to e.g. a 16MP SLR or 16-22MP Imacon digiback?

Haven't run those tests yet (g) but I do have an 11 Mpix dSLR and
prints aren't quite as good as I can get from my 645 Velvia scans, much
less my 6x7 cm scans ... from what I hear the 16 Mpix Canon 1Ds MII is
pretty much on par with medium format prints except for resolution but
it's $8,000. Everyone who has used the 22 Mpix Phase One back sez it's
better than any medium format film but then it's $30,000 just for the
back. Have never seen prints for comparison but that's what one hears
from reliable sources. The Nikon film scanner makes a lot of sense
while you wait for digital back prices to go down unless you're
shooting many hundreds of rolls of film.

Bill

[MXP]

"Bill Hilton" <***@aol.com> skrev i en meddelelse
news:***@g43g2000cwa.googlegroups.com...
>I have an 8000 and agree that 16 bit vs 14 bit and slightly different
> dmax are not overly important, though I'd rather have the 9000 (but
> I've definitely bought my last film scanner).

Do you think the 9000 is the last we will see from Nikon?
Do 4000 dpi pull out all the details of your chromes?

>
>>I could get a new 8000 for about EUR 2000 and a 9000 is about EUR
>>2600. In US you can maybe find it cheaper
>
> I think the 9000 is $2,000 or so in the USA, didn't know they still
> sold the 8000 new.

There is only one 8000 left.

>
>>Can anybody tell me if the standard scanning software from Nikon is good
>>or do I need SilverFast?
>
> Nikon Scan is adequate, I feel. But then I haven't used SilverFast so
> I could be wrong.
>
>>What print qualites can I achive from a perfect exposed 6x6 Velvia when
>>it is scanned with a 8000 or 9000 ...
>
> Should print quite nicely at 20x20" or maybe even 24x24" to most
> people's taste, though it wouldn't be a bad idea to get a scan done if
> you can find someone to do it and run the print yourself to see if it's
> OK for your tastes. Helps if you know some Photoshop, especially
> sharpening and tonal adjustments.

It is always hard work to get the exact colors out on the print.

>
>> ... compared to e.g. a 16MP SLR or 16-22MP Imacon digiback?
>
> Haven't run those tests yet (g) but I do have an 11 Mpix dSLR and
> prints aren't quite as good as I can get from my 645 Velvia scans, much
> less my 6x7 cm scans ... from what I hear the 16 Mpix Canon 1Ds MII is
> pretty much on par with medium format prints except for resolution but
> it's $8,000. Everyone who has used the 22 Mpix Phase One back sez it's
> better than any medium format film but then it's $30,000 just for the
> back. Have never seen prints for comparison but that's what one hears
> from reliable sources. The Nikon film scanner makes a lot of sense
> while you wait for digital back prices to go down unless you're
> shooting many hundreds of rolls of film.
>
> Bill
>
22 Mpix is not much more than 16 Mpix. You need 4x to get double resolution?
But maybe the quality pr. pixel is a lot higher on the Phase One.

I must find out what I will do. A Nikon D200 with a 32MP sensor for
$2000 is not going to happen this year......

Max

[David J. Littleboy]

"MXP" <***@post11.tele.dk> wrote:
> "Bill Hilton" <***@aol.com> skrev:

>>I have an 8000 and agree that 16 bit vs 14 bit and slightly different
>> dmax are not overly important, though I'd rather have the 9000 (but
>> I've definitely bought my last film scanner).
>
> Do you think the 9000 is the last we will see from Nikon?

I hope not, but I don't understand why Nikon ever made an MF scanner at all.
Completely weird, since they haven't made an MF camera in over 50 years.

> Do 4000 dpi pull out all the details of your chromes?

Pretty much. If you shoot high contrast test targets, you can get more
detail onto a chrome than the 8000/9000 can scrape off, but real images
don't have that sort of detail: vein patterns on a leaf are 1:1.6 contrast,
not 1:2000.

> 22 Mpix is not much more than 16 Mpix. You need 4x to get double
> resolution?
> But maybe the quality pr. pixel is a lot higher on the Phase One.

The MF digital backs largely don't have an antialiasing filter (a component
that is mathematically required for correct imaging), so they have a bogus
sharpness.

> I must find out what I will do. A Nikon D200 with a 32MP sensor for
> $2000 is not going to happen this year......

But a 12MP D200 for US$2000 should happen very soon.

David J. Littleboy
Tokyo, Japan

[Neil Gould]

Recently, David J. Littleboy <***@gol.com> posted:

> "MXP" <***@post11.tele.dk> wrote:
>> "Bill Hilton" <***@aol.com> skrev:
>
>>> I have an 8000 and agree that 16 bit vs 14 bit and slightly
>>> different dmax are not overly important, though I'd rather have the
>>> 9000 (but I've definitely bought my last film scanner).
>>
>> Do you think the 9000 is the last we will see from Nikon?
>
> I hope not, but I don't understand why Nikon ever made an MF scanner
> at all. Completely weird, since they haven't made an MF camera in
> over 50 years.
>
Nikon is a respected name in other areas of photography, such as
reproduction for pre-press (I kept a Nikon copy camera lens as a
decoration over the mantle in my studio). As mid-to-high end film scanners
are primarily replacements for pre-press darkroom equipment, it would have
been surprising if Nikon bowed out of that market.

>> Do 4000 dpi pull out all the details of your chromes?
>
> Pretty much. If you shoot high contrast test targets, you can get more
> detail onto a chrome than the 8000/9000 can scrape off, but real
> images don't have that sort of detail: vein patterns on a leaf are
> 1:1.6 contrast, not 1:2000.
>
I don't think that getting film detail is so much a resolution issue as
how well the scanner can handle other aspects. There are more differences
between scanner technologies than there are between the resolution of
these scanners. If one is only considering CCD imaging devices, they will
all give more similar results than one would get from PMT devices, for
example. Beyond that, a lot (if not more) depends on the image being
scanned, the software used, and the skill and patience of the operator.

Neil

[Q.G. de Bakker]

David J. Littleboy wrote:

> I hope not, but I don't understand why Nikon ever made an MF scanner at
all.
> Completely weird, since they haven't made an MF camera in over 50 years.

They sell quite well, i understand.
So not so weird after all?

> > 22 Mpix is not much more than 16 Mpix. You need 4x to get double
> > resolution?
> > But maybe the quality pr. pixel is a lot higher on the Phase One.
>
> The MF digital backs largely don't have an antialiasing filter (a
component
> that is mathematically required for correct imaging), so they have a bogus
> sharpness.

While on the other hand, the name these 'anti-alias filters', or 'low pass
filters' were used to go under was 'soft focus filters'.
You don't have sharpness at all using these devices. They make up by
software sharpening.
So how's that?
;-)

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42b9830f$0$36646$***@news.wanadoo.nl...
> David J. Littleboy wrote:
SNIP
>> The MF digital backs largely don't have an antialiasing
>> filter (a component that is mathematically required for
>> correct imaging), so they have a bogus sharpness.
>
> While on the other hand, the name these 'anti-alias filters',
> or 'low pass filters' were used to go under was 'soft focus
> filters'.

That's not correct. Not only would such a filter be very expensive,
what's worse is that you would see little if an AA-filter were to be
used as a softfocus filter.
AA filters produce 4 focused images in close proximity behind the
filter as light passes through (2 rotated layers of) a bi-refringent
material (Lithium Niobate is often used). Soft focus filters typically
create a mix of focused and non-focused image content.

This
<http://www.canon.com/technology/detail/digi_35mm/lo_filter/index.html>
is one schematic example of what an AA-filter looks like, and it's the
distance between filter and the sensor array that needs to be matched
with the sensel pitch. The average spread between the two focused
images equals the sensel pitch, so 2 pixels get some of the image
detail, thus effectively *only* blurring the finest detail, some/most
(depending on lens MTF and image contrast) of which can be restored by
postprocessing.

> You don't have sharpness at all using these devices. They
> make up by software sharpening.
> So how's that?
> ;-)

It's not terribly accurate ... ;-)

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> > While on the other hand, the name these 'anti-alias filters',
> > or 'low pass filters' were used to go under was 'soft focus
> > filters'.
>
> That's not correct. Not only would such a filter be very expensive,
> what's worse is that you would see little if an AA-filter were to be
> used as a softfocus filter.
> AA filters produce 4 focused images in close proximity behind the
> filter as light passes through (2 rotated layers of) a bi-refringent
> material (Lithium Niobate is often used). Soft focus filters typically
> create a mix of focused and non-focused image content.
>
> This
> <http://www.canon.com/technology/detail/digi_35mm/lo_filter/index.html>
> is one schematic example of what an AA-filter looks like, and it's the
> distance between filter and the sensor array that needs to be matched
> with the sensel pitch. The average spread between the two focused
> images equals the sensel pitch, so 2 pixels get some of the image
> detail, thus effectively *only* blurring the finest detail, some/most
> (depending on lens MTF and image contrast) of which can be restored by
> postprocessing.

So in the end it *is* a soft focus filter, albeit a very expensive one...
Nothing new there. Though the length people go to give spin to it, explain
it away, is impressive.
(And, by the way, 'cheap' soft focus filters also only blur the finest
detail).
;-)

What these things do is limit the image resolution provided by the lens.
They do indeed not produce "bogus sharpness", just less sharpness.
But granted, it is "mathematically correct" softeness.

To get the highest resolution/best sharpness, these filters have to go.
The thing about this trying to make people believe that these are not soft
focus filters is where the "bogus" bit in this story can be found.
And then there are all the lens manufacturers too, trying to convince us we
need extra high resolution lenses for direct digital capture...

Yes, without filter you do get aliasing.
It's an either-or situation: sharpness or aliasing. A real life limitation
of direct digital capture.



> > You don't have sharpness at all using these devices. They
> > make up by software sharpening.
> > So how's that?
> > ;-)
>
> It's not terribly accurate ... ;-)
>
> Bart
>

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42b9b48a$0$66160$***@news.wanadoo.nl...
SNIP
> So in the end it *is* a soft focus filter, [...]

Soft focus? This is with only capture deconvolution sharpening:
http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> > So in the end it *is* a soft focus filter, [...]
>
> Soft focus? This is with only capture deconvolution sharpening:
> http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg

Yes, Bart: soft focus.
What these thingies doe is get rid of the high frequency detail.
That is what soft focus filters do too.
Depsite the presentations ('spin'), it's as simple as that. Not exactly
rocket science. ;-)

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42bad6b4$0$8315$***@news.wanadoo.nl...
> Bart van der Wolf wrote:
>
>> > So in the end it *is* a soft focus filter, [...]
>>
>> Soft focus? This is with only capture deconvolution sharpening:
>> http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg
>
> Yes, Bart: soft focus.
> What these thingies doe is get rid of the high frequency detail.
> That is what soft focus filters do too.

I'm afraid they don't, thank goodness.
They combine focused with unfocused (spherical aberrations).
There is still high spatial frequency detail, but with lowered
contrast.

But perhaps you have a good link to corroborate your statement?
Here are some from me (hint: Softar):
<http://www.schneideroptics.com/filters/filters_for_still_photography/soft_focus/>
even available in a software version:
<http://www.schneider-kreuznach.com/filter_e/sfilter_sf.htm>
or (hint: bold typeface):
<http://www.mir.com.my/rb/photography/companies/canon/fdresources/fdlenses/85mmsoft.htm>
or, with a clear distinction between diffusion and softfocus:
<http://www.f32.net/Associates/CookeOptics.php>

> Depsite the presentations ('spin'), it's as simple as that.

So what's wrong with the presentations?
Just qualifying them as spin seems a little evasive, so unlike you ;-)

> Not exactly rocket science. ;-)

Exactly, although you give different signals (just hints, no
objectively verifiable facts).

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> > Yes, Bart: soft focus.
> > What these thingies doe is get rid of the high frequency detail.
> > That is what soft focus filters do too.
>
> I'm afraid they don't, thank goodness.
> They combine focused with unfocused (spherical aberrations).
> There is still high spatial frequency detail, but with lowered
> contrast.

Yes, with lowered contrast.
The only (!!!) way you can get resolution down, the only (!!!) way you can
get rid of high (or any other) frequencies.

They are, and act exactly like, soft focus filters.
Despite all the spin... ;-)

> Just qualifying them as spin seems a little evasive, so unlike you ;-)

If it looks like, smells like, tastes like, but most importantly acts like a
soft focus filter, Bart, it *is* a soft focus filter.
Now they try to change the way they look, the way they smell, the way they
taste... But they cannot change what they do, because that is precisely what
they need them to do.
So no matter by whatever name you may wish to call them, they are and remain
soft focus filters.

Now let's see more results of the energy you put in finding articles that do
not even deny flat out (how could they? They just avoid mentioning the
common name. Spin...) saying they are not soft focus filters... ;-)

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42bb2f92$0$82426$***@news.wanadoo.nl...
> Bart van der Wolf wrote:
>
>> > Yes, Bart: soft focus.
>> > What these thingies doe is get rid of the high frequency detail.
>> > That is what soft focus filters do too.
>>
>> I'm afraid they don't, thank goodness.
>> They combine focused with unfocused (spherical aberrations).
>> There is still high spatial frequency detail, but with lowered
>> contrast.
>
> Yes, with lowered contrast.
> The only (!!!) way you can get resolution down, the only (!!!) way
> you can
> get rid of high (or any other) frequencies.

Not really, just lowering the sampling pitch will achieve real
resolution reduction. In fact (adequately scanned) film exhibits
usually much higher resolution, but with lower modulation, compared to
digital captures.

> They are, and act exactly like, soft focus filters.
> Despite all the spin... ;-)
>
>> Just qualifying them as spin seems a little evasive, so unlike you
>> ;-)
>
> If it looks like, smells like, tastes like, but most importantly
> acts like a
> soft focus filter, Bart, it *is* a soft focus filter.

Wrong, their behavior is quite different (assuming you refer to soft
focus filters instead of diffusion filters, which still is a bit of a
quess at this point).

SNIP

> So no matter by whatever name you may wish to call them, they are
> and remain
> soft focus filters.

I'm not trying to redefine accepted terminology, you seem to be.

> Now let's see more results of the energy you put in finding
> articles...

Do you need more to convince you of the difference between diffusion
filters and soft focus filters?
How about providing some yourself, not that I need convincing but it
might be a change from me doing the work (which you seem to distrust
anyway).

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> > Yes, with lowered contrast.
> > The only (!!!) way you can get resolution down, the only (!!!) way
> > you can
> > get rid of high (or any other) frequencies.
>
> Not really, just lowering the sampling pitch will achieve real
> resolution reduction.

Bart, the resolution that needs to be brought down is that in the image the
lens produces. Not in the captured image.

> > If it looks like, smells like, tastes like, but most importantly
> > acts like a
> > soft focus filter, Bart, it *is* a soft focus filter.
>
> Wrong, their behavior is quite different (assuming you refer to soft
> focus filters instead of diffusion filters, which still is a bit of a
> quess at this point).

Trying to introduce some topic drift? Spin? ;-)

The result of what these fancy AA/low pass filters do is indistinguishable
from what soft focus filters do.
So you may call them by whatever name you like, you may believe whatever
intricate explanation of how these things work you want to believe.
But don't be fooled thinking that they are different.

> > So no matter by whatever name you may wish to call them, they are
> > and remain
> > soft focus filters.
>
> I'm not trying to redefine accepted terminology, you seem to be.

???
You apparently miss the entire point of this bit of the thread.

> Do you need more to convince you of the difference between diffusion
> filters and soft focus filters?
> How about providing some yourself, not that I need convincing but it
> might be a change from me doing the work (which you seem to distrust
> anyway).

Why don't you have a long hard think, not about the fancy explanations of
how these devices work, not about trying to understand the different names
that are being used for these devices, but about what these devices end up
doing to the image the lens produces. Then perhaps the penny will drop. ;-)

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42bbbc69$0$82424$***@news.wanadoo.nl...
> Bart van der Wolf wrote:
SNIP
>> > If it looks like, smells like, tastes like, but most
>> > importantly acts like a soft focus filter, Bart, it *is*
>> > a soft focus filter.
>>
>> Wrong, their behavior is quite different (assuming you
>> refer to soft focus filters instead of diffusion filters,
>> which still is a bit of a quess at this point).
>
> Trying to introduce some topic drift? Spin? ;-)

???

It neither looks like, smells like, tastes like, nor acts like a soft
focus filter, and you still insist on calling it what it isn't. When
I'm trying to guess what makes you believe something wrong, you call
it spin?
Hmm, interesting attitude.

> The result of what these fancy AA/low pass filters do is
> indistinguishable from what soft focus filters do.

LOL, if you still want to believe so, feel free. I've explained the
differences, shown examples and provided links, but nobody is forcing
you to accept the verifiable truth.

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> >> Wrong, their behavior is quite different (assuming you
> >> refer to soft focus filters instead of diffusion filters,
> >> which still is a bit of a quess at this point).
> >
> > Trying to introduce some topic drift? Spin? ;-)
>
> ???

Turning this into something about the diference betwen soft focus and
difuison filters?
;-)

> It neither looks like, smells like, tastes like, nor acts like a soft
> focus filter, and you still insist on calling it what it isn't. When
> I'm trying to guess what makes you believe something wrong, you call
> it spin?
> Hmm, interesting attitude.

Bart, we've reached the end of this line.
That you don't (want to?) see what these things are, fine.

> LOL, if you still want to believe so, feel free. I've explained the
> differences, shown examples and provided links, but nobody is forcing
> you to accept the verifiable truth.

You have done nothing of the kind, Bart.
Only pointed towards (among other things) examples of the exact same spin i
keep referring to.

(And now you say "I've explained": you haven't said anything really, have
you? Just pointed to where you found thingies on the net. ;-))

All that remains is say again what you have snipped away:

"Why don't you have a long hard think, not about the fancy explanations of
how these devices work, not about trying to understand the different names
that are being used for these devices, but about what these devices end up
doing to the image the lens produces. Then perhaps the penny will drop. ;-)"

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote:
>
> All that remains is say again what you have snipped away:
>
> "Why don't you have a long hard think, not about the fancy explanations of
> how these devices work, not about trying to understand the different names
> that are being used for these devices, but about what these devices end up
> doing to the image the lens produces. Then perhaps the penny will drop.
> ;-)"

No, you should think about what these devices end up doing to the image the
_sensor_ produces. What they do is assure that there are no aliasing
artifacts in the image. If you want a correct, artifact-free images from
your sensor, you must use a low-pass filter of some sort in front of the
sensor.

That's the math, and no amount of squawking about "what it does to the image
from the lens" changes that. Digital sensors are limited in what they can
reproduce by their spacing, and you have to use them within that "speed
limit", or accept the consequences.

If you insist that the filter has "got to go", then you have to argue that
the aliasing and aliasing artifacts are _acceptable_.

But that argument doesn't fly. At least for professional work.

The reason it doesn't fly is that even if you accept the snap-to-grid effect
(jaggies, the bogus sharpness you seem to be fond of), you can't predict
when unacceptable Moire will occur. So you can't use it in a professional
context where you need to be assured that the images produced will be
usable. What some people are finding with the filter-less backs is exactly
that: they can't predict when the system will screw them, and they are
unusable in a professional context.

As a practical matter, if you look at the images from the 8MP AA-filtered
sensors, you'll see that they capture about as much information 24x36 mm of
35mm film. Rather than whimper about the "blur filter", it would seem that a
more sensible, rational, logic attitude would be to find that performance
flipping amazing, simply because it is.

So your whole rant here is wrong mathematically, professionally, and
practically. That's three strikes, QG.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> No, you should think about what these devices end up doing to the image
the
> _sensor_ produces.

That is the next step in the image capture process.
We are still trying to create a break-through in the "soft focus denial"
stage.

> What they do is assure that there are no aliasing
> artifacts in the image. If you want a correct, artifact-free images from
> your sensor, you must use a low-pass filter of some sort in front of the
> sensor.

Yes.
Who's contesting that?
(The use of the word "correct", by the way, is contestable. ;-))

> [...]
> If you insist that the filter has "got to go", then you have to argue that
> the aliasing and aliasing artifacts are _acceptable_.

Reread the thread, and you notice how i (more than once) mention that it is
an either-or situation.
You either use the filter and do not get these artifacts, or you do get more
detail but also artifacts.
It's a choice. Nothing wrong with that, is there?

The amount and degree of undesirable artifacts depend on the 'pattern'
presented to the sensor.
It is by no means predetermined that not using that soft focus thingy always
results in unacceptable images.
Just like it is not predetermined that you always need more resolution than
you get using such a thing.

> But that argument doesn't fly. At least for professional work.

The argument that aliasing artifacts are acceptable?
Depends entirely on how much of those there are.

And don't forget that aliasing artifacts do occur even with these filters in
place.
It isn't as simple as you present it here, is it?

> The reason it doesn't fly is that even if you accept the snap-to-grid
effect
> (jaggies, the bogus sharpness you seem to be fond of),

No, David, the "jaggies" are a side effect of the real (!) and extra
"sharpness" you do get when not using a filter.
There is nothing "bogus" about that extra "sharpness".

> you can't predict
> when unacceptable Moire will occur.

Right!

> So you can't use it in a professional
> context where you need to be assured that the images produced will be
> usable.

And where you can instantly see, the moment a image has been captured,
whether it would be better to use that thing or not?

> What some people are finding with the filter-less backs is exactly
> that: they can't predict when the system will screw them, and they are
> unusable in a professional context.


> As a practical matter, if you look at the images from the 8MP AA-filtered
> sensors, you'll see that they capture about as much information 24x36 mm
of
> 35mm film. Rather than whimper about the "blur filter", it would seem that
a
> more sensible, rational, logic attitude would be to find that performance
> flipping amazing, simply because it is.

'Whimper' is what Bart does.
I have no problem accepting that these filers are resolution limiting soft
focus filters. Why would i?
I'm sure i have said that before.

'Whimper' is what you do when you proclaim as gospel truth (mathematical,
professional and practical no less) that the extra resolution (at first
denied; perhaps still?) is never ever something one could use.
If i haven't mentioned that before, note that i have now. ;-)

Why is it so hard for some to accept things the simple, sometimes ugly, way
they are?
Where does this aversion, revulsion even, stem from?

> So your whole rant here is wrong mathematically, professionally, and
> practically. That's three strikes, QG.

I admire your confidence, David.
;-)

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote:
> David J. Littleboy wrote:
>
>> No, you should think about what these devices end up doing to the image
> the
>> _sensor_ produces.
>
> That is the next step in the image capture process.
> We are still trying to create a break-through in the "soft focus denial"
> stage.

If you want more resolution in digital capture, you need more pixels.

>> What they do is assure that there are no aliasing
>> artifacts in the image. If you want a correct, artifact-free images from
>> your sensor, you must use a low-pass filter of some sort in front of the
>> sensor.
>
> Yes.
> Who's contesting that?
> (The use of the word "correct", by the way, is contestable. ;-))

You are, in that last line of yours. There's nothing contestable about it at
all.

Failing to use a low-pass filter means that the position of elements in the
image is incorrect by +/- 1/2 a pixel. What you get at the pixel level is a
snap-to-grid effect. Every pixel is wrong.

>> [...]
>> If you insist that the filter has "got to go", then you have to argue
>> that
>> the aliasing and aliasing artifacts are _acceptable_.
>
> Reread the thread, and you notice how i (more than once) mention that it
> is
> an either-or situation.
> You either use the filter and do not get these artifacts, or you do get
> more
> detail but also artifacts.

No. You don't get more detail. You get incorrectly placed detail.

> It's a choice. Nothing wrong with that, is there?

It's not a choice.

> The amount and degree of undesirable artifacts depend on the 'pattern'
> presented to the sensor.
> It is by no means predetermined that not using that soft focus thingy
> always
> results in unacceptable images.

Here's the either/or: When you don't get a disaster, you just get an image
with the detail positioned incorrectly on the pixels.

>> But that argument doesn't fly. At least for professional work.
>
> The argument that aliasing artifacts are acceptable?
> Depends entirely on how much of those there are.
>
> And don't forget that aliasing artifacts do occur even with these filters
> in
> place.
> It isn't as simple as you present it here, is it?
>
>> The reason it doesn't fly is that even if you accept the snap-to-grid
> effect
>> (jaggies, the bogus sharpness you seem to be fond of),
>
> No, David, the "jaggies" are a side effect of the real (!) and extra
> "sharpness" you do get when not using a filter.

There's no such thing as "real sharpness". If you don't low-pass filter, the
pixel grid forces detail onto the pixel grid. Which means that the detail is
put in the wrong position in the final image.

So that apparent extra sharpness is completely bogus. It's a random guess by
the sensor. If you see seven clear lines in an area in an image, the normal
assumption would be for there to have been seven lines in the subject. But
if you look at the Foveon images of test charts (on dpreview), the 9 lines
turn into 7, and then 5 lines. The sensor produces sharp patterns at the
pixel spacing of the sensor.

It's a complete joke.

> There is nothing "bogus" about that extra "sharpness".

It's completely bogus because it's created by putting the detail in the
wrong place, i.e. forcing the detail onto the period of the sensor.

> I have no problem accepting that these filers are resolution limiting soft
> focus filters. Why would i?

You remain incorrect in the idea that they limit resolution. The apparently
resolved detail (the seven bogus lines in the test chart) isn't resolved at
all, it's made up by the sensor.

>> So your whole rant here is wrong mathematically, professionally, and
>> practically. That's three strikes, QG.
>
> I admire your confidence, David.

It comes from being right here. Arguing against using an AA filter is
untennable.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> If you want more resolution in digital capture, you need more pixels.

Come on, try harder.
;-)

> >> What they do is assure that there are no aliasing
> >> artifacts in the image. If you want a correct, artifact-free images
from
> >> your sensor, you must use a low-pass filter of some sort in front of
the
> >> sensor.
> >
> > Yes.
> > Who's contesting that?
> > (The use of the word "correct", by the way, is contestable. ;-))
>
> You are, in that last line of yours. There's nothing contestable about it
at
> all.

Let's see.
You introduce a device with no other intention as change the image another
device produces, and think that the end result still can be called artifact
free?
Think again.

And no, i'm not contesting that lowering the image resolution *artificially*
will reduce the risk of some other artifacts occuring.

Where do you see that?
You're just arguing for the sake of it, aren't you? For you make no sense at
all, and that's not how we know you.

> Failing to use a low-pass filter means that the position of elements in
the
> image is incorrect by +/- 1/2 a pixel. What you get at the pixel level is
a
> snap-to-grid effect. Every pixel is wrong.

Complete and utter nonsense.

It *may* happen to *some* pixels.
You yourself mentioned something about unpredictability. Now you forget?

> No. You don't get more detail. You get incorrectly placed detail.

That's not true. Plain and simple.

> > It's a choice. Nothing wrong with that, is there?
>
> It's not a choice.

Not if your religion doesn't allow it, David. No.

> Here's the either/or: When you don't get a disaster, you just get an image
> with the detail positioned incorrectly on the pixels.

That appears to be your new found believe, yes.
But it is pure nonsense.

> > No, David, the "jaggies" are a side effect of the real (!) and extra
> > "sharpness" you do get when not using a filter.
>
> There's no such thing as "real sharpness". [...]

There you go...
;-)

> If you don't low-pass filter, the
> pixel grid forces detail onto the pixel grid. Which means that the detail
is
> put in the wrong position in the final image.

If you don't low pass filter, present the sensor with an patterned image,
such that each distinct point in that pattern falls exactly on one sensor
cell, you get at least double the resolution you get when using a soft focus
device.
Sure, it would be a truly amazing feat to do that. But it shows that your
'mathematical' thinking leaves a lot to be desired.

You cannot predict what will be presented to the sensor, nor how any pattern
will match the sensor patern.
Sometimes it will produce ugliness in some places, but equally often it will
produce perfectly acceptable things in other places.
There is no law of nature, no religious dogma, no mathematical model even
that supports your newly found strictness.

> So that apparent extra sharpness is completely bogus. [...]

Again, pure nonsense.

> > I have no problem accepting that these filers are resolution limiting
soft
> > focus filters. Why would i?
>
> You remain incorrect in the idea that they limit resolution. The
apparently
> resolved detail (the seven bogus lines in the test chart) isn't resolved
at
> all, it's made up by the sensor.

And yet again...

> > I admire your confidence, David.
>
> It comes from being right here. Arguing against using an AA filter is
> untennable.

The Real Truth, 'ey David?
;-)

[Stacey]

Q.G. de Bakker wrote:

> David J. Littleboy wrote:
>
>> If you want more resolution in digital capture, you need more pixels.
>
> Come on, try harder.
> ;-)

Interesting you snipped this out..

------
So that apparent extra sharpness is completely bogus. It's a random guess by
the sensor. If you see seven clear lines in an area in an image, the normal
assumption would be for there to have been seven lines in the subject. But
if you look at the Foveon images of test charts (on dpreview), the 9 lines
turn into 7, and then 5 lines. The sensor produces sharp patterns at the
pixel spacing of the sensor.

-----

See if you can follow this simple example. Look between 12 and 18 on both
examples and then explain how the example with the AA filter is "softer".
You're trying to compare a filter placed in FRONT of the lens to the
multiple aa filters placed over the sensor pixels.


http://www.geocities.com/kievgurl/samples/AAFilter.html

Then agin I doubt you're interested in learning anything and want to argue
this just to argue. I'm glad David posted this as I did learn something new
today! :-)


--

Stacey

[Q.G. de Bakker]

Stacey wrote:

> Interesting you snipped this out..
>
> ------
> So that apparent extra sharpness is completely bogus. It's a random guess
by
> the sensor. If you see seven clear lines in an area in an image, the
normal
> assumption would be for there to have been seven lines in the subject. But
> if you look at the Foveon images of test charts (on dpreview), the 9 lines
> turn into 7, and then 5 lines. The sensor produces sharp patterns at the
> pixel spacing of the sensor.
>

Stacey, that is describing aliasing.
The hidden assumption here is that without filter aliasing will always
occur, in every part of the image.
Furthermore, it is ignoring the fact that the AA filters reduce the
resolution to at least half that of the pixels spacing.

[Boat]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42bdde45$0$13884$***@news.wanadoo.nl...
>> the sensor. If you see seven clear lines in an area in an image, the
> normal
>> assumption would be for there to have been seven lines in the subject.
>> But
>> if you look at the Foveon images of test charts (on dpreview), the 9
>> lines
>> turn into 7, and then 5 lines. The sensor produces sharp patterns at the
>> pixel spacing of the sensor.
>>
>
> Stacey, that is describing aliasing.
> The hidden assumption here is that without filter aliasing will always
> occur, in every part of the image.
> Furthermore, it is ignoring the fact that the AA filters reduce the
> resolution to at least half that of the pixels spacing.

The problem seems pretty clear now, even to us slow'uns. It's a question of
what happens when the detail exceeds the sensor's ability to capture it. Or
similarly, how an image displays into fewer pixels. You can sum or average
the "signals" into the pixel they happen to fall nearest, or you can moosh
it into the one, two or more nearest. Maybe the problem is simply that real
pixels have non-zero size.

Unfiltered, we readily agree artifacts can occur in some percentage of
shots. Not all aliasing ruins the shot; we only care about the ones made
unusable. On these, you might as well not have clicked the shutter.
Extending this logic, say the camera just doesn't fire. It appears to have
fired, but didn't and I have no immediate way of knowing. Say I have a
second camera, one that always fires, but gosh darn it, every shot is
visibly less sharp in the very finest details.

When I reach into the gear bag, which one will I come out with? How often
will the "bad" one misfire? One in ten? More? Less?

[Stacey]

Q.G. de Bakker wrote:

>
> Stacey, that is describing aliasing.
> The hidden assumption here is that without filter aliasing will always
> occur, in every part of the image.

No, it's knowing that at any point it -can- occur and ruin the image. It
does occure often enough to be a problem or they wouldn't have developed
this "fix" for it.

> Furthermore, it is ignoring the fact that the AA filters reduce the
> resolution to at least half that of the pixels spacing.

Looking at the sample I posted between the sigma and a 10D, I'm not seeing
this resolution loss.

And again trying to say aa filters over the pixels are the same as a soft
focus filter placed over the lens is absurd.

--

Stacey

[Q.G. de Bakker]

"Stacey" <***@yahoo.com> schreef in bericht
news:***@individual.net...
> Q.G. de Bakker wrote:
>
> >
> > Stacey, that is describing aliasing.
> > The hidden assumption here is that without filter aliasing will always
> > occur, in every part of the image.
>
> No, it's knowing that at any point it -can- occur and ruin the image. It
> does occure often enough to be a problem or they wouldn't have developed
> this "fix" for it.
>
> > Furthermore, it is ignoring the fact that the AA filters reduce the
> > resolution to at least half that of the pixels spacing.
>
> Looking at the sample I posted between the sigma and a 10D, I'm not seeing
> this resolution loss.
>
> And again trying to say aa filters over the pixels are the same as a soft
> focus filter placed over the lens is absurd.
>
> --
>
> Stacey

[Q.G. de Bakker]

Stacey wrote:

> > Stacey, that is describing aliasing.
> > The hidden assumption here is that without filter aliasing will always
> > occur, in every part of the image.
>
> No, it's knowing that at any point it -can- occur and ruin the image. It
> does occure often enough to be a problem or they wouldn't have developed
> this "fix" for it.

You apparently do not know that aliasing also can occur when using an AA
filter?

Yes, it is a problem that occurs often enough for this thing to make sense.
But it still is that 'either-or' thing.
Without filter, aliasing will occur and not appear periodically with
increasing detail frequencies. Up to the point detail frequency is the same
as the pixel pitch (ignoring Bayer and other patterns for simplicity's
sake). With image detail exactly the same as the pixel pitch, and a rather
lucky placement of one pattern over the other, there will be no aliasing,
and each single sensor element will record true (not bogus, as David keps
believing) image detail.
With filter, the same periodic aliasing problem occurs, though with less
contrast, thus less visible (and if contrast is reuced ar enough: not
visible), but (!) the filter is limiting the recordable image detail
frequency to at least half that which the sensor could record. That is the
only (!) way it can achieve that feat.

Now you can try standing on your head, juggling three dead herrings with
your left hand, holding a copy an English-Danish dictionary in the other,
while singing the third and fifth verse of your favourite musical tune,
nothing will change that.

> > Furthermore, it is ignoring the fact that the AA filters reduce the
> > resolution to at least half that of the pixels spacing.
>
> Looking at the sample I posted between the sigma and a 10D, I'm not seeing
> this resolution loss.

Well good for you!

> And again trying to say aa filters over the pixels are the same as a soft
> focus filter placed over the lens is absurd.

Yes, becuase it is not over the lens. ;-)
No matter how you look at it, Stacey (and Bart) this AA/Low Pass thing *is*
a soft focus filter.

[Stacey]

Q.G. de Bakker wrote:

> but (!) the filter is limiting the recordable image detail
>frequency to at least half that which the sensor could record. That is the
>only (!) way it can achieve that feat.

Have you looked at actual sample images or just theorizing what is happening
from your limited viewpoint on this subject? Garbage infomation being
stored is just that, garbage information.


>> Looking at the sample I posted between the sigma and a 10D, I'm not
>> seeing this resolution loss.
>
> Well good for you!
>


So did you even bother to look at the examples? Of course not, you already
know everything... Nothing new here.

--

Stacey

[Lassi Hippeläinen]

Q.G. de Bakker kirjoitti:
<...>
> No matter how you look at it, Stacey (and Bart) this AA/Low Pass thing *is*
> a soft focus filter.

Obviously you don't know how soft focus filters work...

-- Lassi

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42be79ec$0$86124$***@news.wanadoo.nl...
> Stacey wrote:
SNIP
>> And again trying to say aa filters over the pixels are the
>> same as a soft focus filter placed over the lens is absurd.
>
> Yes, becuase it is not over the lens. ;-)
> No matter how you look at it, Stacey (and Bart) this AA/Low
> Pass thing *is* a soft focus filter.

As we clearly disagree, how about a couple of independent sources to
support your claim?

Bart

[Boat]

"Bart van der Wolf" <***@no.spam> wrote in message
news:42bfc936$0$31293$***@news.xs4all.nl...
>
> "Q.G. de Bakker" <***@tiscali.nl> wrote in message
>> No matter how you look at it, Stacey (and Bart) this AA/Low
>> Pass thing *is* a soft focus filter.
>
> As we clearly disagree, how about a couple of independent sources to
> support your claim?

First principles, perhaps, will have to suffice. Whatever you choose to call
it, its job is to blur the image. Even if they performed the job the same
way through the same means, anti-alias is a better name for the thingee in
front of my sensor. Its purpose is to minimize aliasing artifacts. Soft
focus describes quite well its role as a facial pore reducing device when in
front of the lens.

[Q.G. de Bakker]

Boat wrote:

> > As we clearly disagree, how about a couple of independent sources to
> > support your claim?
>
> First principles, perhaps, will have to suffice. Whatever you choose to
call
> it, its job is to blur the image.

That's exactly it.

I wonder what Bart will make of that...

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42c0153b$0$8317$***@news.wanadoo.nl...
> Boat wrote:
>
>> > As we clearly disagree, how about a couple of independent sources
>> > to
>> > support your claim?
>>
>> First principles, perhaps, will have to suffice. Whatever you
>> choose to
> call
>> it, its job is to blur the image.
>
> That's exactly it.
>
> I wonder what Bart will make of that...

Don't behave like a silly child, I've shown results and links to
relevant facts.
How about some links from your side for a change, or are you just
talk?

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> Don't behave like a silly child, I've shown results and links to
> relevant facts.
> How about some links from your side for a change, or are you just
> talk?

Bart, how difficult is it to understand what "Boat" summarized so
succinctly?
But no, you need other people saying something for you to get your head
round a simple concept?
Who's the silly child, i wonder...

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42c038c3$0$14026$***@news.wanadoo.nl...
> Bart van der Wolf wrote:
>
>> Don't behave like a silly child, I've shown results and links to
>> relevant facts.
>> How about some links from your side for a change, or are you just
>> talk?
>
> Bart, how difficult is it to understand what "Boat" summarized so
> succinctly?
> But no, you need other people saying something for you to get your
> head
> round a simple concept?
> Who's the silly child, i wonder...

Apparently all talk. You have no factual links to back up your
contention of soft focus filters acting as AA-filters, do you?

Bart

[MXP]

"Stacey" <***@yahoo.com> skrev i en meddelelse
news:***@individual.net...
> Q.G. de Bakker wrote:
>
>> David J. Littleboy wrote:
>>
>>> If you want more resolution in digital capture, you need more pixels.
>>
>> Come on, try harder.
>> ;-)
>
> Interesting you snipped this out..
>
> ------
> So that apparent extra sharpness is completely bogus. It's a random guess
> by
> the sensor. If you see seven clear lines in an area in an image, the
> normal
> assumption would be for there to have been seven lines in the subject. But
> if you look at the Foveon images of test charts (on dpreview), the 9 lines
> turn into 7, and then 5 lines. The sensor produces sharp patterns at the
> pixel spacing of the sensor.
>
> -----
>
> See if you can follow this simple example. Look between 12 and 18 on both
> examples and then explain how the example with the AA filter is "softer".
> You're trying to compare a filter placed in FRONT of the lens to the
> multiple aa filters placed over the sensor pixels.
>
>
> http://www.geocities.com/kievgurl/samples/AAFilter.html
>
> Then agin I doubt you're interested in learning anything and want to argue
> this just to argue. I'm glad David posted this as I did learn something
> new
> today! :-)
>
>
> --
>
> Stacey

It looks likes color fringle. I would have belived that the lens had a
problem. So when
people test lenses using a digital SLR and complain about color fringle they
often argue
that it is because the lens is not good enough for the sensor (missing color
correction
that the film is not able to show). In these cases the problem could be an
"AA problem"?

Max

[Boat]

"David J. Littleboy" <***@gol.com> wrote in message
news:d9j8vc$ed0$***@nnrp.gol.com...

> So that apparent extra sharpness is completely bogus. It's a random guess
> by the sensor. If you see seven clear lines in an area in an image, the
> normal assumption would be for there to have been seven lines in the
> subject. But if you look at the Foveon images of test charts (on
> dpreview), the 9 lines turn into 7, and then 5 lines. The sensor produces
> sharp patterns at the pixel spacing of the sensor.

Thanks, David. That finally makes clear what you two were jabbering about.
Great discussion; you two are to be congratulated.

[Stacey]

Boat wrote:

> "David J. Littleboy" <***@gol.com> wrote in message
> news:d9j8vc$ed0$***@nnrp.gol.com...
>
>> So that apparent extra sharpness is completely bogus. It's a random guess
>> by the sensor. If you see seven clear lines in an area in an image, the
>> normal assumption would be for there to have been seven lines in the
>> subject. But if you look at the Foveon images of test charts (on
>> dpreview), the 9 lines turn into 7, and then 5 lines. The sensor produces
>> sharp patterns at the pixel spacing of the sensor.
>
> Thanks, David. That finally makes clear what you two were jabbering about.
> Great discussion; you two are to be congratulated.

Here's a sample page showing the two different examples he was talking
about.

http://www.geocities.com/kievgurl/samples/AAFilter.html
--

Stacey

[johnboy]

"Stacey" <***@yahoo.com> wrote in message
news:***@individual.net...
> Here's a sample page showing the two different examples he was talking
> about.
>
> http://www.geocities.com/kievgurl/samples/AAFilter.html

With NO AA filter is what I'd want... if I cared one tiny bit about digital,
but I don't.

[Bart van der Wolf]

"Stacey" <***@yahoo.com> wrote in message
news:***@individual.net...
SNIP
> Here's a sample page showing the two different examples
> he was talking about.
>
> http://www.geocities.com/kievgurl/samples/AAFilter.html

And here's 'some' more info using a different test target:
http://www.normankoren.com/Tutorials/MTF2.html#Nyquist

Now obviously, most people don't spend their days shooting targets but
it illustrates what can happen without proper precautions.

Bart

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42bc1419$0$82398$***@news.wanadoo.nl...
SNIP
> All that remains is say again what you have snipped away:

I snipped that away because I had already answered it.

But since you asked, I'll do it again:
- Soft focus filters produce a sharp image, with (sperical aberration)
unsharpness added. That usually leads to reduced shadow contrast, and
wide area 'glow' of light areas next to darker areas.
- AA-filters or Low-pass filters reduce the modulation of the finest
details, causing a slight blur. Due to prior knowledge of the type of
blur, it can be partially restored (something that is impossible with
Soft focus filters, too much is lost).

> "Why don't you have a long hard think, not about the fancy
> explanations of how these devices work,

Nothing fancy about it, just facts.

> not about trying to understand the different names that are
> being used for these devices,

No, I'm trying to understand why you obviously don't grasp the
differences, it isn't all that hard, really. Of course it might help
if you did say what you mean instead of telling others what to do ...

> but about what these devices end up doing to the image the
> lens produces. Then perhaps the penny will drop. ;-)"

Maybe you should take your own advice? :-o

Bart

[Q.G. de Bakker]

Bart van der Wolf wrote:

> But since you asked, I'll do it again:
> - Soft focus filters produce a sharp image, with (sperical aberration)
> unsharpness added. That usually leads to reduced shadow contrast, and
> wide area 'glow' of light areas next to darker areas.
> - AA-filters or Low-pass filters reduce the modulation of the finest
> details, causing a slight blur. Due to prior knowledge of the type of
> blur, it can be partially restored (something that is impossible with
> Soft focus filters, too much is lost).

Bart, what you again fail to see is that you are saying the same thing
twice, albeit using other words (and those words, 'spin', still are what's
fooling you).
Both filters in effect reduce MTF of higher spatial frequencies.
Nothing more to it.

[Bart van der Wolf]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42bd4770$0$98068$***@news.wanadoo.nl...
SNIP
> Both filters in effect reduce MTF of higher spatial frequencies.
> Nothing more to it.

The changes to the MTF are totally different, and the Soft focus
filter in addition loses more modulation in lowest frequencies.
Not the same result at all, but I'm glad you finally explained
what you were suggesting.

Bart

[Philip Homburg]

In article <hkQve.292$***@newssvr11.news.prodigy.com>,
Boat <boat042-***@yahoo.com> wrote:
>"Bart van der Wolf" <***@no.spam> wrote in message
>news:42bfc936$0$31293$***@news.xs4all.nl...
>>
>> "Q.G. de Bakker" <***@tiscali.nl> wrote in message
>>> No matter how you look at it, Stacey (and Bart) this AA/Low
>>> Pass thing *is* a soft focus filter.
>>
>> As we clearly disagree, how about a couple of independent sources to
>> support your claim?
>
>First principles, perhaps, will have to suffice. Whatever you choose to call
>it, its job is to blur the image. Even if they performed the job the same
>way through the same means, anti-alias is a better name for the thingee in
>front of my sensor. Its purpose is to minimize aliasing artifacts. Soft
>focus describes quite well its role as a facial pore reducing device when in
>front of the lens.

There is difference. The properties of an ideal AA filter are dictated
by mathematics. An ideal 'brick-wall filter' is transparent for all
frequencies below Nyquist and is completely opaque for all frequencies
at the sampling rate or above.

Soft focus filters are used because they result in pictures that are
more pleasing to the human eye. I don't think that people developing
soft focus filters ever accidentally created a product that can be used
effectively as an AA filter.


--
That was it. Done. The faulty Monk was turned out into the desert where it
could believe what it liked, including the idea that it had been hard done
by. It was allowed to keep its horse, since horses were so cheap to make.
-- Douglas Adams in Dirk Gently's Holistic Detective Agency

[Neil Gould]

Recently, Philip Homburg <***@pch.home.cs.vu.nl> posted:

> In article <hkQve.292$***@newssvr11.news.prodigy.com>,
> Boat <boat042-***@yahoo.com> wrote:
>> "Bart van der Wolf" <***@no.spam> wrote in message
>> news:42bfc936$0$31293$***@news.xs4all.nl...
>>>
>>> "Q.G. de Bakker" <***@tiscali.nl> wrote in message
>>>> No matter how you look at it, Stacey (and Bart) this AA/Low
>>>> Pass thing *is* a soft focus filter.
>>>
>>> As we clearly disagree, how about a couple of independent sources to
>>> support your claim?
>>
>> First principles, perhaps, will have to suffice. Whatever you choose
>> to call it, its job is to blur the image. Even if they performed the
>> job the same way through the same means, anti-alias is a better name
>> for the thingee in front of my sensor. Its purpose is to minimize
>> aliasing artifacts. Soft focus describes quite well its role as a
>> facial pore reducing device when in front of the lens.
>
> There is difference. The properties of an ideal AA filter are dictated
> by mathematics. An ideal 'brick-wall filter' is transparent for all
> frequencies below Nyquist and is completely opaque for all frequencies
> at the sampling rate or above.
>
> Soft focus filters are used because they result in pictures that are
> more pleasing to the human eye. I don't think that people developing
> soft focus filters ever accidentally created a product that can be
> used effectively as an AA filter.
>
I didn't see a claim that an AA filter has the same properties or operates
in the same way as soft-focus filters. What Q.G. has argued, and Boat
summarized is that the underlying function is the same -- to blur some
aspect of the image (specifically high-frequency information).

Some people, such as David L., feel that AA filters are indispensable.
From his perspective of what constitutes a "good image", that feeling is
completely predictable. Others feel that AA filters are an option that
help some images but may hurt some others. If everyone gets the images
that they want using AA or not, why should anyone else really care?

Regards,

Neil

[Bart van der Wolf]

"Neil Gould" <***@myplaceofwork.com> wrote in message
news:F10we.437$***@newssvr19.news.prodigy.com...
SNIP
> I didn't see a claim that an AA filter has the same properties
> or operates in the same way as soft-focus filters.

QG> "The result of what these fancy AA/low pass filters do
QG> is indistinguishable from what soft focus filters do".

> What Q.G. has argued, and Boat summarized is that the
> underlying function is the same -- to blur some
> aspect of the image (specifically high-frequency information).

I don't think anyone contested that with regards to AA-filters.
One may have different opinions about the need for an AA-filter,
and one may fear blurring due to the AA-filter, but the result of
both types of filter are not the same. Soft focus filters are
designed (in various ways) to introduce spherical aberrations
(which also affects lower spatial frequencies), unlike AA-filters.

[...]
> If everyone gets the images that they want using AA or not,
> why should anyone else really care?

Perhaps not everyone gets the images they want, or people
want to optimize what they can get. It may help to understand
the inevitable boundaries, and optimize the rest.

As I've shown, AA-filters blur only the finest detail, but
some/most of that can be restored:
<http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg>
so the trade-off isn't as bad as some may suggest.

Bart

[Neil Gould]

Recently, Bart van der Wolf <***@no.spam> posted:

> "Neil Gould" <***@myplaceofwork.com> wrote in message
> news:F10we.437$***@newssvr19.news.prodigy.com...
> SNIP
>> I didn't see a claim that an AA filter has the same properties
>> or operates in the same way as soft-focus filters.
>
>> "The result of what these fancy AA/low pass filters do
>> is indistinguishable from what soft focus filters do".
>
I read that statement as a matter of function, the end result, not the
process by which the result is achieved. The arguments against that
statement appear focus on the process and ignore the end result.

>> What Q.G. has argued, and Boat summarized is that the
>> underlying function is the same -- to blur some
>> aspect of the image (specifically high-frequency information).
>
> I don't think anyone contested that with regards to AA-filters.
> One may have different opinions about the need for an AA-filter,
> and one may fear blurring due to the AA-filter, but the result of
> both types of filter are not the same. Soft focus filters are
> designed (in various ways) to introduce spherical aberrations
> (which also affects lower spatial frequencies), unlike AA-filters.
>
Your argument appears to be one of "how much" blurring, not "whether it
blurs or not".

> [...]
>> If everyone gets the images that they want using AA or not,
>> why should anyone else really care?
>
> Perhaps not everyone gets the images they want, or people
> want to optimize what they can get. It may help to understand
> the inevitable boundaries, and optimize the rest.
>
I think it's safe to say that no one *always* gets the images they want...
;-)

> As I've shown, AA-filters blur only the finest detail, but
> some/most of that can be restored:
> <http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg>
> so the trade-off isn't as bad as some may suggest.
>
Well, if your example of the details being "restored" has to do with
"sharpening", then I'd disagree that you are achieving a "restoration" at
all. Instead, those "details" are being _created_ by changing the content
of pixels near a given threshold. It's an automated version of pixel-level
editing, and as such is an interpretation, not a restoration. In other
words, considering the signal (aka original scene), those details were
captured "wrong", and they're still "wrong" after that manipulation, but
if it achieves the image that you wan't then it works for you. That was my
point about why should anyone care? A good image is a good image.

Neil

[Bart van der Wolf]

"Neil Gould" <***@myplaceofwork.com> wrote in message
news:MJ9we.2012$***@newssvr31.news.prodigy.com...
> Recently, Bart van der Wolf <***@no.spam> posted:
>
>> "Neil Gould" <***@myplaceofwork.com> wrote in message
>> news:F10we.437$***@newssvr19.news.prodigy.com...
>> SNIP
>>> I didn't see a claim that an AA filter has the same properties
>>> or operates in the same way as soft-focus filters.
>>
>>> "The result of what these fancy AA/low pass filters do
>>> is indistinguishable from what soft focus filters do".
>>
> I read that statement as a matter of function, the end result, not
> the
> process by which the result is achieved.

That's possible. I read the "properties OR" part as describing the
process, which may or may not have been what was meant (those short
answers leave too much room for interpretation IMO).

SNIP
> Your argument appears to be one of "how much" blurring, not
> "whether it blurs or not".

Not really. Soft focus filters add/mix blur to/with a mostly in-focus
image. It's distinctly different from just reducing the amplitude of
the highest frequencies, in fact they remove more from the lowest
frequencies. Quite different behavior.

SNIP
>> As I've shown, AA-filters blur only the finest detail, but
>> some/most of that can be restored:
>> <http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg>
>> so the trade-off isn't as bad as some may suggest.
>>
> Well, if your example of the details being "restored" has
> to do with "sharpening" [...]

No, sharpening in the sense of edge contrast enhancement like USM was
not applied.
Richardson Lucy restoration is one of many methods used to restore
blurred (PSF convolved) images to their original state, it adapts
based on Baysian statistics (which are useful for Poisson
distributions, like Photons exhibit). It uses prior knowledge about
the blur from the Point Spread Function (PSF) of the optical chain.
<http://www.mathworks.com/access/helpdesk/help/toolbox/images/deblurr7.html#8550>

To put it in simple terms, it kind of pulls in luminance from
surrounding pixels for each pixel in the image, proportional to the
blur that displaced the luminance. That's why some features 'shrink'
in size, and gain or lose luminance.

The restoration process is not perfect because information was lost at
capture time (e.g. 12-bit capture and quantzation noise added, and
non-image data was added at capture and read-out, ADC,
demosaicing/gamma-adjustment/colorbalancing, etc.). Other restoration
methods can be used, but they may restore the noise more effectively
than the signal, which is generally a bad avenue to take for images.

Bart

[Neil Gould]

Recently, Bart van der Wolf <***@no.spam> posted:

> "Neil Gould" <***@myplaceofwork.com> wrote in message
>
>> Your argument appears to be one of "how much" blurring, not
>> "whether it blurs or not".
>
> Not really. Soft focus filters add/mix blur to/with a mostly in-focus
> image. It's distinctly different from just reducing the amplitude of
> the highest frequencies, in fact they remove more from the lowest
> frequencies. Quite different behavior.
>
I guess the distinction that you're making doesn't clear it up for me. It
appears that you're describing the cut-off frequency and/or slope of the
roll-off, rather than describing something that "behaves quite
differently"... they "behave" the same, but have different
characteristics.

> SNIP
>>> As I've shown, AA-filters blur only the finest detail, but
>>> some/most of that can be restored:
>>> <http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jpg>
>>> so the trade-off isn't as bad as some may suggest.
>>>
>> Well, if your example of the details being "restored" has
>> to do with "sharpening" [...]
>
> No, sharpening in the sense of edge contrast enhancement like USM was
> not applied.
> Richardson Lucy restoration is one of many methods used to restore
> blurred (PSF convolved) images to their original state, it adapts
> based on Baysian statistics (which are useful for Poisson
> distributions, like Photons exhibit). It uses prior knowledge about
> the blur from the Point Spread Function (PSF) of the optical chain.
>
<http://www.mathworks.com/access/helpdesk/help/toolbox/images/deblurr7.htm
l#8550>
>
I see. That is a better approach than sharpening, but I don't see how it
could be free of potential errors, as well. It is changing the content of
pixels based on presumptions about the image that may or may not be true,
since the detail has already been lost. Again, we're talking about
trade-offs, and I don't see those trade-offs as being the points of
contention in this discussion... no one is claiming that they don't exist.
;-)

Neil

[Bart van der Wolf]

"Neil Gould" <***@myplaceofwork.com> wrote in message
news:wnewe.2034$***@newssvr31.news.prodigy.com...
SNIP
> Again, we're talking about trade-offs, and I don't see those
> trade-offs as being the points of contention in this
> discussion... no one is claiming that they don't exist.

Correct, but overstating the (partially recoverable) blur versus
increased aliasing, can tip the balance in the wrong direction if one
makes an 'informed' choice. ;-) Buyer beware!

Bart

[Nicholas O. Lindan]

"Bart van der Wolf" <***@no.spam> wrote

> Not really. Soft focus filters add/mix blur to/with a mostly in-focus
> image. It's distinctly different from just reducing the amplitude of
> the highest frequencies, in fact they remove more from the lowest
> frequencies. Quite different behavior.

If a blurred ['low pass'] image is added to the original image one gets a
'soft focus' effect, as mentioned about. The net effect is:

s ......
i .
g .
n ..........
a
l f r e q u e n c y

In audio this it is a distant/romantic/muzak sound for music
[just cutting the treble makes the music sound dull] but
this filter is bad for speech comprehension. In the visual
realm it is bad for reading: dirty eyeglasses.

I am sure every vendor has its own special take on 'soft focus'.
The above is a simple analogy of soft-focus.

Interestingly, if the blurred image is -subtracted- from the
original image the result is 'unsharp masking'. In audio
it makes voice sound clearer: very sharp pronunciation without
much sibilance; music gets the jangles.

The frequency analogy above is not what really goes on with
image filtering. A spatial filter is a convolution which is
not a 1:1 match to what high-pass / low-pass temporal [er, audio]
filters do.

> As I've shown, AA-filters blur only the finest detail, but
> some/most of that can be restored:

Theoretically a convolution can be undone. Practically it is
limited by increased noise in the recovered image.

--
Nicholas O. Lindan, Cleveland, Ohio
Consulting Engineer: Electronics; Informatics; Photonics.
To reply, remove spaces: n o lindan at ix . netcom . com
Fstop timer - http://www.nolindan.com/da/fstop/index.htm

[Bart van der Wolf]

"Nicholas O. Lindan" <***@sig.com> wrote in message
news:Dlfwe.14344$***@newsread3.news.atl.earthlink.net...
SNIP
> Theoretically a convolution can be undone. Practically it
> is limited by increased noise in the recovered image.

Correct, although the dampened RL processing I applied has an improved
S/N behavior when compared to the original algorithm. Also, images
from digicams have two properties that make them very suited for such
post-processing, low noise (high S/N ratio) and a small PSF kernel
(the signal is spread to mostly the immediate surroundings). That's
also why film scans overall tend to benefit less.

Here is another application:
http://clarkvision.com/imagedetail/image-restoration1/

Bart

[Philip Homburg]

In article <F10we.437$***@newssvr19.news.prodigy.com>,
Neil Gould <***@myplaceofwork.com> wrote:
>Recently, Philip Homburg <***@pch.home.cs.vu.nl> posted:
>> There is a difference. The properties of an ideal AA filter are dictated
>> by mathematics. An ideal 'brick-wall filter' is transparent for all
>> frequencies below Nyquist and is completely opaque for all frequencies
>> at the sampling rate or above.
>>
>> Soft focus filters are used because they result in pictures that are
>> more pleasing to the human eye. I don't think that people developing
>> soft focus filters ever accidentally created a product that can be
>> used effectively as an AA filter.
>>
>I didn't see a claim that an AA filter has the same properties or operates
>in the same way as soft-focus filters. What Q.G. has argued, and Boat
>summarized is that the underlying function is the same -- to blur some
>aspect of the image (specifically high-frequency information).

And why is that relevant to this discussion? Both filters have different
designs and goals. Bad lenses also blur high frequencies, but that doesn't
make them particularly good soft focus filters (or AA filters for that
matter).

>Some people, such as David L., feel that AA filters are indispensable.
>From his perspective of what constitutes a "good image", that feeling is
>completely predictable. Others feel that AA filters are an option that
>help some images but may hurt some others. If everyone gets the images
>that they want using AA or not, why should anyone else really care?

Because ideal AA filters do not exist and are certainly not used in
combination with Bayer pattern sensors. So it becomes an interesting
trade-off.

In most cameras you can't just install or remove the AA. So the choice of
AA filter is something to think about when selecting a camera for a
particular subject.


--
That was it. Done. The faulty Monk was turned out into the desert where it
could believe what it liked, including the idea that it had been hard done
by. It was allowed to keep its horse, since horses were so cheap to make.
-- Douglas Adams in Dirk Gently's Holistic Detective Agency

[Bart van der Wolf]

"Philip Homburg" <***@pch.home.cs.vu.nl> wrote in message
news:***@inews_id.stereo.hq.phicoh.net...
SNIP
> Because ideal AA filters do not exist and are certainly not
> used in combination with Bayer pattern sensors. So it
> becomes an interesting trade-off.

IMHO that's an important point, "trade-off" being the key word. Some
subjects (e.g. involving fabric or brick walls/streets, tile roofs and
several other regular patterns, especially when shot at an angle) are
destined to produce trouble when an AA-filter is omitted (especially
with Bayer CFA sensors), others may get away without such a precaution
(only smooth texture).
It's also useful to know that the various Raw converters handle the
artifacts in different ways, but they too have to strike a balance
between detail extraction and artifacts.

In fact, the ship example I showed also has aliasing artifacts
(despite the AA-filter) in the tackle running through the large block.
The devastation by colored and other artifacts was restricted to
managable proportions, allowing controlled luminance re-sharpening.

> In most cameras you can't just install or remove the AA. So the
> choice of AA filter is something to think about when selecting a
> camera for a particular subject.

And in general, micro-lenses also help to optimize the resistance
against aliasing and other artifacts (besides their positive influence
on sensitivity and potential for reducing vignetting). It's influence
on image quality is e.g. visible in reduced stairstepping tendency in
diagonal/slanted edges.

Bart

[johnboy]

"Bart van der Wolf" <***@no.spam> wrote
>
> "Q.G. de Bakker" <***@tiscali.nl> wrote
>> Yes, Bart: soft focus.
>> What these thingies doe is get rid of the high frequency detail.
>> That is what soft focus filters do too.
>
> I'm afraid they don't, thank goodness.
> They combine focused with unfocused (spherical aberrations).

Not necessarily spherical aberrations.

> There is still high spatial frequency detail, but with lowered contrast.

Depends, but you are generally right. Another example is the Tiffen
'contrast' filters. I use the quote because we normally think of colored
filters while these are neutral - intended really for Digital Video to help
suppress blooming highlights, and to help fill in shadows. I have used them
in terrible high-noon light on the 'blad with success.

Then there's the Imagon, another different animal.

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote in message
news:42b9b48a$0$66160$***@news.wanadoo.nl...
> Bart van der Wolf wrote:
>
>> > While on the other hand, the name these 'anti-alias filters',
>> > or 'low pass filters' were used to go under was 'soft focus
>> > filters'.
>>
>> That's not correct. Not only would such a filter be very expensive,
>> what's worse is that you would see little if an AA-filter were to be
>> used as a softfocus filter.
>> AA filters produce 4 focused images in close proximity behind the
>> filter as light passes through (2 rotated layers of) a bi-refringent
>> material (Lithium Niobate is often used). Soft focus filters typically
>> create a mix of focused and non-focused image content.
>>
>> This
>> <http://www.canon.com/technology/detail/digi_35mm/lo_filter/index.html>
>> is one schematic example of what an AA-filter looks like, and it's the
>> distance between filter and the sensor array that needs to be matched
>> with the sensel pitch. The average spread between the two focused
>> images equals the sensel pitch, so 2 pixels get some of the image
>> detail, thus effectively *only* blurring the finest detail, some/most
>> (depending on lens MTF and image contrast) of which can be restored by
>> postprocessing.
>
> So in the end it *is* a soft focus filter, albeit a very expensive one...
> Nothing new there. Though the length people go to give spin to it, explain
> it away, is impressive.

It's not spin, it's mathematics. To sample correctly, you must bandlimit the
signal to the passband of the sampling system. Anything above the passband
of the sensor is mapped by the sampling process to a frequency in the
passband and trashes your image.

It's _not_ "blurring detail", it's removing information that can't be
rendered by the sensor.

> What these things do is limit the image resolution provided by the lens.
> They do indeed not produce "bogus sharpness", just less sharpness.
> But granted, it is "mathematically correct" softeness.
>
> To get the highest resolution/best sharpness, these filters have to go.

No! They have to be there to get correct images.

> The thing about this trying to make people believe that these are not soft
> focus filters is where the "bogus" bit in this story can be found.

That's mathematically illiterate and wrong.

> And then there are all the lens manufacturers too, trying to convince us
> we
> need extra high resolution lenses for direct digital capture...

What you need is high MTF in the passband. Higher limiting resolution
usually is a good indication that the lens has higher contrast at lower
frequencies than a lens with a lower limiting resolution.

> Yes, without filter you do get aliasing.
> It's an either-or situation: sharpness or aliasing. A real life limitation
> of direct digital capture.

Digital capture is limited to the range of frequencies the sensor spacing
allows you to capture. That's the limitation. You must bandlimit the input
to that range.

There's also the added problem that Bayer demosaicing fails in a seriously
ugly manner in the presence of Moire.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> > So in the end it *is* a soft focus filter, albeit a very expensive
one...
> > Nothing new there. Though the length people go to give spin to it,
explain
> > it away, is impressive.
>
> It's not spin, it's mathematics. To sample correctly, you must bandlimit
the
> signal to the passband of the sampling system. Anything above the passband
> of the sensor is mapped by the sampling process to a frequency in the
> passband and trashes your image.

See that it *is* spin?
What you describe above (apart from the 'why' bit) is exactly what soft
focus filters do.

> It's _not_ "blurring detail", it's removing information that can't be
> rendered by the sensor.

It is lowering contrast of the higher spatial frequency detail.
It lowers MTF of that detail.
It 'soft focusses' the image produced by the lens.

That it does so because the ordered array of sensitive cells cannot cope
with that type of information is interesting (sort of), but has no bearing
on the fact that what these things are is soft focus filters.

> > To get the highest resolution/best sharpness, these filters have to go.
>
> No! They have to be there to get correct images.

Yes, it i indeed that either-or thing i mentioned. Isn't it.
You give up high resolution to avoid regular and irregular moire patterns.

But "To get the highest resolution/best sharpness, these filters have to
go".
They really have to!

> > The thing about this trying to make people believe that these are not
soft
> > focus filters is where the "bogus" bit in this story can be found.
>
> That's mathematically illiterate and wrong.

It is neither.
These are soft focus filters, and mathematics (and use of fancy words) is
used to make people forget that.

> > And then there are all the lens manufacturers too, trying to convince us
> > we
> > need extra high resolution lenses for direct digital capture...
>
> What you need is high MTF in the passband.

:-)

Indeed. And beyond that they do not need to produce anything.
In fact, if they had excellent MTF up to the cut-off frequency, and scoed 0%
above, they would be even better.
So that's why Zeiss is promoting their 300 lp/mm lenses as specially suited
for digital. To have that 300 lp/mm reduced to, what?, 20 lp/mm?

> Higher limiting resolution
> usually is a good indication that the lens has higher contrast at lower
> frequencies than a lens with a lower limiting resolution.

Perhaps, yes.
Fact remains that anything above the limiting resolution has to be
'soft-focussed' away to avoid sensor related problems.

> > Yes, without filter you do get aliasing.
> > It's an either-or situation: sharpness or aliasing. A real life
limitation
> > of direct digital capture.
>
> Digital capture is limited to the range of frequencies the sensor spacing
> allows you to capture. That's the limitation. You must bandlimit the input
> to that range.

Indeed.
But don't get me wrong, i can accept that limitation.
(I interject this here, because even i get the feeling that i may be on a
anti-digital rant. I'm not! ;-))
I just find it strange (beyond amusement) that so much effort is put into
giving spin to the fact that in direct digital capture resolution has to be
limited, and that this is done using soft focus filters (albeit expensive
ones). I wonder, is that really so hard to stomach?

> There's also the added problem that Bayer demosaicing fails in a seriously
> ugly manner in the presence of Moire.

That too.

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote:
>>
>> No! They have to be there to get correct images.
>
> Yes, it i indeed that either-or thing i mentioned. Isn't it.
> You give up high resolution to avoid regular and irregular moire patterns.

No. You don't give up any resolution. You never had that resolution in the
first place.

A digital sensor is only capable of resolving up to a certain frequency.

> But "To get the highest resolution/best sharpness, these filters have to
> go".
> They really have to!

No. To correctly resolve at the highest resolution the system is capable of
they have to be there.

> I just find it strange (beyond amusement) that so much effort is put into
> giving spin to the fact that in direct digital capture resolution has to
> be
> limited,

Stop right there: the resolution is limited by the spacing of the pixels.
You can't resolve any more than the spacing of the pixels. Trying to do so
results in artifiacts. If you want quality images, you have to have an AA
filter.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> Stop right there: the resolution is limited by the spacing of the pixels.
> You can't resolve any more than the spacing of the pixels.

Obviously, but...

> Trying to do so
> results in artifiacts.

?
You mean using images with 'detail frequency' the same or nearly the same as
the pixel-pitch will create artifacts.
That's not "trying" to resolve "more" than the spacing of the pixels, is it?
;-)

> If you want quality images, you have to have an AA
> filter.

... the AA filter removes resolvable (by the sensor !) resolution. The
resolution/high frequency detail that would create these undesired effects.
That resolution is quite well resolved by the sensor in some bits, creates
ugly things in other bits of the image, depending on where on the ordered
array it 'falls'.
The fabled Nyquist thingy these soft focus devices are used to reduce
resolution to is about half that of what the sensor can resolve (ignoring
the Bayer thingy for a moment). So there's what you throw away...
Given an ideal situation, you can resolve quite a bit more than the limit
these AA filters impose, without any artifact at all. Mind you, it would be
a exceptionally ideal situation. ;-)

[MXP]

"David J. Littleboy" <***@gol.com> skrev i en meddelelse
news:d9f0om$apr$***@nnrp.gol.com...
>
> "Q.G. de Bakker" <***@tiscali.nl> wrote:
>>>
>>> No! They have to be there to get correct images.
>>
>> Yes, it i indeed that either-or thing i mentioned. Isn't it.
>> You give up high resolution to avoid regular and irregular moire
>> patterns.
>
> No. You don't give up any resolution. You never had that resolution in the
> first place.
>
> A digital sensor is only capable of resolving up to a certain frequency.
>
>> But "To get the highest resolution/best sharpness, these filters have to
>> go".
>> They really have to!
>
> No. To correctly resolve at the highest resolution the system is capable
> of they have to be there.
>
>> I just find it strange (beyond amusement) that so much effort is put into
>> giving spin to the fact that in direct digital capture resolution has to
>> be
>> limited,
>
> Stop right there: the resolution is limited by the spacing of the pixels.
> You can't resolve any more than the spacing of the pixels. Trying to do so
> results in artifiacts. If you want quality images, you have to have an AA
> filter

Why does this not happen with film? .....is it because of the random
pattern film grains has?

If so......why not make a sensor with random pattern? ......the RAW
converter
just have to be more advanced?
.
>
> David J. Littleboy
> Tokyo, Japan
>
>

[Q.G. de Bakker]

MXP wrote:

> Why does this not happen with film? .....is it because of the random
> pattern film grains has?

That's precisely it, yes.
Film has a pattern, but it isn't ordered. So you get less (practically none)
of the problems you get when two ordered patterns overlay.

This, by the way, is also the reason why direct digital capture without soft
focus filter can work: overlay a very random pattern on a strictly ordered
pattern, and there will not be very much unpleasantness.

> If so......why not make a sensor with random pattern? ......the RAW
> converter
> just have to be more advanced?

Well, as long as the randomness of the pattern is known, it is as easy to
process as a strictly ordered pattern.
However, you will want an even subject to be recorded evenly, don't you?
Film doesn't do that. It shows grain. A random digital array will do the
same. Now if they can get the pattern size/pixel-pitch an order of magnitude
smaller than it is today... Who knows... ;-)

[Gregory Blank]

In article <d9a7rs$v0c$***@nnrp.gol.com>,
"David J. Littleboy" <***@gol.com> wrote:

>
> But a 12MP D200 for US$2000 should happen very soon.

How do you figure they just released the D2x for 5k.

--
LF Website @ http://members.verizon.net/~gregoryblank

"To announce that there must be no criticism of the President,
or that we are to stand by the President, right or wrong,
is not only unpatriotic and servile, but is morally treasonable
to the American public."--Theodore Roosevelt, May 7, 1918

[David J. Littleboy]

"Gregory Blank" <***@gregblankphoto.com> wrote:
> "David J. Littleboy" <***@gol.com> wrote:
>>
>> But a 12MP D200 for US$2000 should happen very soon.
>
> How do you figure they just released the D2x for 5k.

The sensor in the D2x can't cost any more than any other 1.5x sensor, so
there's no reason not to release a "20D killer" with that sensor...

(By soon I meant "a year or so".)

David J. Littleboy
Tokyo, Japan

[Stacey]

David J. Littleboy wrote:

>
> "Gregory Blank" <***@gregblankphoto.com> wrote:
>> "David J. Littleboy" <***@gol.com> wrote:
>>>
>>> But a 12MP D200 for US$2000 should happen very soon.
>>
>> How do you figure they just released the D2x for 5k.
>
> The sensor in the D2x can't cost any more than any other 1.5x sensor, so
> there's no reason not to release a "20D killer" with that sensor...
>


Except for the profits..
--

Stacey

[Sander Vesik]

Stacey <***@yahoo.com> wrote:
> David J. Littleboy wrote:
>
> >
> > "Gregory Blank" <***@gregblankphoto.com> wrote:
> >> "David J. Littleboy" <***@gol.com> wrote:
> >>>
> >>> But a 12MP D200 for US$2000 should happen very soon.
> >>
> >> How do you figure they just released the D2x for 5k.
> >
> > The sensor in the D2x can't cost any more than any other 1.5x sensor, so
> > there's no reason not to release a "20D killer" with that sensor...
> >
>
>
> Except for the profits..

Uhh... Most of the cost or price in the camera doesn't come form the
sensor. not in 35mm style digital cameras anyways.

--
Sander

+++ Out of cheese error +++

[Stacey]

Sander Vesik wrote:

> Stacey <***@yahoo.com> wrote:
>> David J. Littleboy wrote:
>>
>> >
>> > "Gregory Blank" <***@gregblankphoto.com> wrote:
>> >> "David J. Littleboy" <***@gol.com> wrote:
>> >>>
>> >>> But a 12MP D200 for US$2000 should happen very soon.
>> >>
>> >> How do you figure they just released the D2x for 5k.
>> >
>> > The sensor in the D2x can't cost any more than any other 1.5x sensor,
>> > so there's no reason not to release a "20D killer" with that sensor...
>> >
>>
>>
>> Except for the profits..
>
> Uhh... Most of the cost or price in the camera doesn't come form the
> sensor. not in 35mm style digital cameras anyways.
>
And again why would they release a $1500 camera when they can get 3X that
much for the ones they have for sale now?
--

Stacey

[David J. Littleboy]

"Stacey" <***@yahoo.com> wrote:
> Sander Vesik wrote:
>>>
>>> Except for the profits..
>>
>> Uhh... Most of the cost or price in the camera doesn't come form the
>> sensor. not in 35mm style digital cameras anyways.
>>
> And again why would they release a $1500 camera when they can get 3X that
> much for the ones they have for sale now?

You flunk Econ 101. The market for US$5000 bodies is extremely limited. The
market for US$1500 bodies is large and growing. If they make US$3500 on each
US$5000 body they sell, that's a lot less than making US$500 on 100 times as
many bodies.

The business plan is to sell to the high-end pros, the camera pools and the
rich doctors first. When that market saturates, come out with a $1500 camera
for the rest of us. Canon's problem is that they probably can't sell a
cheaper body with the 1Ds2 sensor in it for less than US$2500 or US$3000,
and a factor of two in price for a mere 17% increase in resolution may be a
hard sell to the mass amateur market.

David J. Littleboy
Tokyo, Japan

[Q.G. de Bakker]

David J. Littleboy wrote:

> You flunk Econ 101. The market for US$5000 bodies is extremely limited.
The
> market for US$1500 bodies is large and growing. If they make US$3500 on
each
> US$5000 body they sell, that's a lot less than making US$500 on 100 times
as
> many bodies.

$500 profit selling a body for $1500 means they make $4000 profit, not
$3500, when they sell it for $5000, right?

Have i flunked too, or will you take a make-up exam?
;-)

[David J. Littleboy]

"Q.G. de Bakker" <***@tiscali.nl> wrote:
> David J. Littleboy wrote:
>
>> You flunk Econ 101. The market for US$5000 bodies is extremely limited.
> The
>> market for US$1500 bodies is large and growing. If they make US$3500 on
> each
>> US$5000 body they sell, that's a lot less than making US$500 on 100 times
> as
>> many bodies.
>
> $500 profit selling a body for $1500 means they make $4000 profit, not
> $3500, when they sell it for $5000, right?
>
> Have i flunked too, or will you take a make-up exam?

You flunk, too. They're different bodies: one is ridiculously heavy with a
nice viewfinder, and one is a reasonable weight with a yucky viewfinder.

The $5000 body is based on a pro body, and probably costs them $1500 to
make.
The $1500 body is based on an amateur body, and probably costs them $1000 to
make.

(Actually, come to think of it, my mfg cost estimates are probably US$500
too high. Oh, well, I pass econ, but flunk math.)

David J. Littleboy
Tokyo, Japan

[Gregory Blank]

In article <d9d5bh$pld$***@nnrp.gol.com>,
"David J. Littleboy" <***@gol.com> wrote:

> "Gregory Blank" <***@gregblankphoto.com> wrote:
> > "David J. Littleboy" <***@gol.com> wrote:
> >>
> >> But a 12MP D200 for US$2000 should happen very soon.
> >
> > How do you figure they just released the D2x for 5k.
>
> The sensor in the D2x can't cost any more than any other 1.5x sensor, so
> there's no reason not to release a "20D killer" with that sensor...
>
> (By soon I meant "a year or so".)
>
> David J. Littleboy
> Tokyo, Japan

I suppose that would be great, but I would like to
see a FF sensor come from Nikon in that time frame
priced at 3k with maybe 20-24 mp. (Ah well just wishful thinking :)

--
LF Website @ http://members.verizon.net/~gregoryblank

"To announce that there must be no criticism of the President,
or that we are to stand by the President, right or wrong,
is not only unpatriotic and servile, but is morally treasonable
to the American public."--Theodore Roosevelt, May 7, 1918

[Gordon Moat]

"David J. Littleboy" wrote:

> "Gregory Blank" <***@gregblankphoto.com> wrote:
> > "David J. Littleboy" <***@gol.com> wrote:
> >>
> >> But a 12MP D200 for US$2000 should happen very soon.
> >
> > How do you figure they just released the D2x for 5k.
>
> The sensor in the D2x can't cost any more than any other 1.5x sensor, so
> there's no reason not to release a "20D killer" with that sensor...
>
> (By soon I meant "a year or so".)
>
> David J. Littleboy
> Tokyo, Japan

It could happen, though the cost savings to allow a price drop would result
in a cheaper body. Pick up a D2X, and then pick up a cheaper (Nikon or
Canon) digital SLR, and you will notice a difference with the D2X. Not only
does it seem more solid, the viewfinder is better and brighter. You are not
just getting a sensor for the money.

Obviously, prices will continue to drop, or capabilities will increase at
the same price points. Considering that the near 6 µm cell site size seems
to be somewhat of a limit, the best a full 24 mm by 36 mm chip could do
would be just over 20 to 24 MP. If Nikon, or Sony, could develop that 5.4
µm cell site to a larger full frame chip, then they could get 24 to 28 MP.
That would be the upper limit, though I wonder when/if we would ever see
such products.

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[MXP]

"Gordon Moat" <***@attglobal.net> skrev i en meddelelse
news:***@attglobal.net...
> "David J. Littleboy" wrote:
>
>> "Gregory Blank" <***@gregblankphoto.com> wrote:
>> > "David J. Littleboy" <***@gol.com> wrote:
>> >>
>> >> But a 12MP D200 for US$2000 should happen very soon.
>> >
>> > How do you figure they just released the D2x for 5k.
>>
>> The sensor in the D2x can't cost any more than any other 1.5x sensor, so
>> there's no reason not to release a "20D killer" with that sensor...
>>
>> (By soon I meant "a year or so".)
>>
>> David J. Littleboy
>> Tokyo, Japan
>
> It could happen, though the cost savings to allow a price drop would
> result
> in a cheaper body. Pick up a D2X, and then pick up a cheaper (Nikon or
> Canon) digital SLR, and you will notice a difference with the D2X. Not
> only
> does it seem more solid, the viewfinder is better and brighter. You are
> not
> just getting a sensor for the money.
>
> Obviously, prices will continue to drop, or capabilities will increase at
> the same price points. Considering that the near 6 µm cell site size seems
> to be somewhat of a limit, the best a full 24 mm by 36 mm chip could do
> would be just over 20 to 24 MP. If Nikon, or Sony, could develop that 5.4
> µm cell site to a larger full frame chip, then they could get 24 to 28 MP.
> That would be the upper limit, though I wonder when/if we would ever see
> such products.
>
> Ciao!
>
> Gordon Moat
> A G Studio
> <http://www.allgstudio.com>
>
>
How is Imacon able to resolve 8000 dpi in their scanners?
I know it is a 3 line CCD running across the film but the spacing
must be according to 8000 dpi? ......which means that it should be
possible to make 8000 x 12000 (24x36) sensor?

[Gordon Moat]

MXP wrote:

> "Gordon Moat" <***@attglobal.net> skrev i en meddelelse
> news:***@attglobal.net...
> > "David J. Littleboy" wrote:
> >
> >> "Gregory Blank" <***@gregblankphoto.com> wrote:
> >> > "David J. Littleboy" <***@gol.com> wrote:
> >> >>
> >> >> But a 12MP D200 for US$2000 should happen very soon.
> >> >
> >> > How do you figure they just released the D2x for 5k.
> >>
> >> The sensor in the D2x can't cost any more than any other 1.5x sensor, so
> >> there's no reason not to release a "20D killer" with that sensor...
> >>
> >> (By soon I meant "a year or so".)
> >>
> >> David J. Littleboy
> >> Tokyo, Japan
> >
> > It could happen, though the cost savings to allow a price drop would
> > result
> > in a cheaper body. Pick up a D2X, and then pick up a cheaper (Nikon or
> > Canon) digital SLR, and you will notice a difference with the D2X. Not
> > only
> > does it seem more solid, the viewfinder is better and brighter. You are
> > not
> > just getting a sensor for the money.
> >
> > Obviously, prices will continue to drop, or capabilities will increase at
> > the same price points. Considering that the near 6 µm cell site size seems
> > to be somewhat of a limit, the best a full 24 mm by 36 mm chip could do
> > would be just over 20 to 24 MP. If Nikon, or Sony, could develop that 5.4
> > µm cell site to a larger full frame chip, then they could get 24 to 28 MP.
> > That would be the upper limit, though I wonder when/if we would ever see
> > such products.
> >
> > Ciao!
> >
> > Gordon Moat
> > A G Studio
> > <http://www.allgstudio.com>
> >
> >
> How is Imacon able to resolve 8000 dpi in their scanners?
> I know it is a 3 line CCD running across the film but the spacing
> must be according to 8000 dpi? ......which means that it should be
> possible to make 8000 x 12000 (24x36) sensor?

They use a very different CCD, and in fact there are some that are 10600
resolution. The problem of getting the cell size really small, is that they
cell becomes not very light sensitive. Scanners use a powerful light source of
a known intensity and colour. The scan also happens at a much slower pace than
would capturing an image with a camera. The slower pace of the scanner linear
CCD array helps it capture that detail.

FillFactory in Belgium has a few papers you can read about this. Kodak also
have some of this information in the chip area of their website (not in the
digital photography area). While it would be possible to make a chip as you
describe, it would be limited to a scanning type of situation, with a known
intense light source.

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com/technology.html>

[deloid]

"Bill Hilton" <***@aol.com> wrote in message
news:***@g43g2000cwa.googlegroups.com...
>I have an 8000 and agree that 16 bit vs 14 bit and slightly different
> dmax are not overly important, though I'd rather have the 9000 (but
> I've definitely bought my last film scanner).
Bill,

What did you mean by "last scanner". After the 8000 dies you going straight
digital?

deloid

[Bill Hilton]

>> I'd rather have the 9000 (but
>> I've definitely bought my last film scanner).

> deloid asks ...
>
>What did you mean by "last scanner". After the 8000 dies you going
>straight digital?

That's a good question, what would I do if the 8000 died ... I didn't
consider that option :) I'd try to get it fixed but if it were truly
dead I'd probably get a 9000 after all.

I've already gone to digital as a replacement for 35 mm (8 and 11 Mpix
Canon bodies) but for landscapes I still prefer film and the 645 or 6x7
cameras with Velvia. The problem with shooting landscapes with medium
format is you're competing with large format and it's hard to stay
competitive with that.

But I think everyone who shoots digital RAW and also scans will
sympathize when I say that scanning is a pain in the butt compared to
RAW conversions :) I don't think the current Canon 16 Mpix body will
fully replace the 6x7 (though I haven't actually shot it) but I think
in 12-24 months Canon will likely come out with a 20-24 Mpixel dSLR and
at that point I'll probably convert for good to all-digital. Or if the
Pentax 645 digital body is 22 Mpix instead of 18 I might get that
instead of waiting for Canon since I have seven lenses for that system,
but I don't have high hopes for Pentax right now.

Bill

[Gordon Moat]

Bill Hilton wrote:

> . . . . . . . . . . . .
>
> > ... compared to e.g. a 16MP SLR or 16-22MP Imacon digiback?
>
> Haven't run those tests yet (g) but I do have an 11 Mpix dSLR and
> prints aren't quite as good as I can get from my 645 Velvia scans, much
> less my 6x7 cm scans ... from what I hear the 16 Mpix Canon 1Ds MII is
> pretty much on par with medium format prints except for resolution but
> it's $8,000. Everyone who has used the 22 Mpix Phase One back sez it's
> better than any medium format film but then it's $30,000 just for the
> back.

I have used the older PhaseOne backs a couple times (rental), and they
worked quite nicely. No complaints about resolution, noise, nor aberrations.
Colour could have been better, but I am probably more picky than most
people.

> Have never seen prints for comparison but that's what one hears
> from reliable sources. The Nikon film scanner makes a lot of sense
> while you wait for digital back prices to go down unless you're
> shooting many hundreds of rolls of film.
>
> Bill

It would surprise me if digital back prices (new) went down much. It might
happen eventually, but I think it is too early to tell how many companies
will still be in that market. Without competition, the choices in the future
might be limited, without much price break.

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[Bill Hilton]

> Gordon Moat writes ...
>
>It would surprise me if digital back prices (new) went down much.

? I think the 16 Mpix Phase One back is around $16,000 and supposedly
Pentax and maybe others will have a full camera (not just a back) with
18 Mpix for $7,500 in a few months ... I'm pretty sure that will put
price pressure on the P1 back.

Ditto for the 22 Mpix back that sells for $30,000 ... the Mamiya ZD is
supposed to go for around $12,500 and due "any day now" with 22 Mpix
... I really doubt the Phase One back prices will stay that high once
there's viable competition with the same # of pixels and half the cost.

[Gordon Moat]

Bill Hilton wrote:

> > Gordon Moat writes ...
> >
> >It would surprise me if digital back prices (new) went down much.
>
> ? I think the 16 Mpix Phase One back is around $16,000 and supposedly
> Pentax and maybe others will have a full camera (not just a back) with
> 18 Mpix for $7,500 in a few months ... I'm pretty sure that will put
> price pressure on the P1 back.

These are more than just MP in comparing the various backs. Some of them
do much better with colour than others, and some are better with avoiding
noise than others. Lower noise at 16 MP might be better than greater noise
at 18 MP.

I agree that if there are more players in the market, then the competition
could affect prices. Kodak just acquired creo, who owned Leaf. Since Kodak
already has a chip business supplying PhaseOne and Pentax, it seems like
it would be silly to keep Leaf going. Leaf could end up shut down, having
engineers transferred to Kodak's chip business, or sold off.

After Shriro (Hasselblad) acquired Imacon, it seems that they are moving
towards only Hasselblad backs. Of course, that is the bigger market, and
could help Hasselblad lens and body sales. PhaseOne supply for many
different cameras, including Hasselblad, though they only have a
manufacturer direct agreement with Rollei.

Sinar and Jenoptik EyeLike are two other companies in this market. Sinar
are at the top in capability and price. EyeLike suffer from a lack of
distribution, though some key contracts seem to keep their back market
alive. Both companies have many other products not dependant on digital
back sales.

The funny thing is that there are now more companies making backs than
there are companies making new medium format cameras. Somehow I don't
think that is a situation that can continue into the future.

>
>
> Ditto for the 22 Mpix back that sells for $30,000 ... the Mamiya ZD is
> supposed to go for around $12,500 and due "any day now" with 22 Mpix

I got to briefly try one out. The weight, handling, and ergonomics are
great, and Mamiya has done well with this. One look through the large and
bright viewfinder tells you that working with such a camera could be a
good experience. It is far different than the very squinty viewfinders on
most digital SLRs. In comparison, I wish the Canon top of the line had
more magnification in the finder, though the full frame is better for
viewing than APS sized or half frame digital SLRs. Both Kodak full frame
models are let down by low magnification, and a body based upon a low end
design.

I don't entirely understand the market target for Mamiya with the ZD.
There are not nearly enough "features" nor focus points to attract anyone
away from Canon. My feeling, somewhat echoed by the MAC rep, is that the
Mamiya digital back would be a better choice for those who already own
Mamiya lenses.

>
> ... I really doubt the Phase One back prices will stay that high once
> there's viable competition with the same # of pixels and half the cost.

I think the chip prices need to drop for them to reduce the costs much.
Other than that, PhaseOne would need to have some other products that
compliment the digital back, and provide some revenue stream. The Rollei
6008 Digital package is one example, though it seems they (Rollei) might
be loosing money on that just to get some lens sales (profits).

We hear a little of the same things from the smaller camera digital SLR
makers. Many are loosing money on each body just to try to get market
share. The idea is that accessories and lens sales could pick up the
difference, and ensure overall profits. One thing that disturbs me about
this is a recent industry report on lens sales, indicating a 26% drop in
the last four years. Can any company still generate profits in the
photography industry?

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[Tony Polson]

Gordon Moat <***@attglobal.net> wrote:

>After Shriro (Hasselblad) acquired Imacon, it seems that they are moving
>towards only Hasselblad backs. Of course, that is the bigger market, and
>could help Hasselblad lens and body sales.


Without wishing to dispute your point about digital backs for MF
cameras, Imacon developed the Digital Modul-R for Leica, and are also
working with Leica on the M Digital.

[Gordon Moat]

Tony Polson wrote:

> Gordon Moat <***@attglobal.net> wrote:
>
> >After Shriro (Hasselblad) acquired Imacon, it seems that they are moving
> >towards only Hasselblad backs. Of course, that is the bigger market, and
> >could help Hasselblad lens and body sales.
>
> Without wishing to dispute your point about digital backs for MF
> cameras, Imacon developed the Digital Modul-R for Leica, and are also
> working with Leica on the M Digital.

Very true, though I was mostly addressing the medium format market. The
Leica deal was in place prior to Hasselblad acquiring Imacon. In a way it
makes sense for companies to outsource the engineering. I have heard of
Sinar offering similar development services, though they are very secretive
about for whom they do development work.

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[johnboy]

"Tony Polson" <***@nospam.net> wrote

> Without wishing to dispute your point about digital backs for MF
> cameras, Imacon developed the Digital Modul-R for Leica, and are also
> working with Leica on the M Digital.

Gentlemen's bet that Leica is deceased by 2007?

[Gordon Moat]

johnboy wrote:

> "Tony Polson" <***@nospam.net> wrote
>
> > Without wishing to dispute your point about digital backs for MF
> > cameras, Imacon developed the Digital Modul-R for Leica, and are also
> > working with Leica on the M Digital.
>
> Gentlemen's bet that Leica is deceased by 2007?

18 months from now . . . . . what would you like to bet for?

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[Gregory Blank]

In article <***@attglobal.net>,
Gordon Moat <***@attglobal.net> wrote:

> johnboy wrote:
>
> > "Tony Polson" <***@nospam.net> wrote
> >
> > > Without wishing to dispute your point about digital backs for MF
> > > cameras, Imacon developed the Digital Modul-R for Leica, and are also
> > > working with Leica on the M Digital.
> >
> > Gentlemen's bet that Leica is deceased by 2007?
>
> 18 months from now . . . . . what would you like to bet for?

Might I suggest an Imacon scanner? ~:-0

--
LF Website @ http://members.verizon.net/~gregoryblank

"To announce that there must be no criticism of the President,
or that we are to stand by the President, right or wrong,
is not only unpatriotic and servile, but is morally treasonable
to the American public."--Theodore Roosevelt, May 7, 1918

[Gordon Moat]

Gregory Blank wrote:

> In article <***@attglobal.net>,
> Gordon Moat <***@attglobal.net> wrote:
>
> > johnboy wrote:
> >
> > > "Tony Polson" <***@nospam.net> wrote
> > >
> > > > Without wishing to dispute your point about digital backs for MF
> > > > cameras, Imacon developed the Digital Modul-R for Leica, and are also
> > > > working with Leica on the M Digital.
> > >
> > > Gentlemen's bet that Leica is deceased by 2007?
> >
> > 18 months from now . . . . . what would you like to bet for?
>
> Might I suggest an Imacon scanner? ~:-0

Nice, but I don't want him to feel too bad about having to spend so much
money. ;-)

Ciao!

Gordon Moat
A G Studio
<http://www.allgstudio.com>

[Gregory Blank]

In article <***@g43g2000cwa.googlegroups.com>,
"Bill Hilton" <***@aol.com> wrote:

>Everyone who has used the 22 Mpix Phase One back sez it's
> better than any medium format film but then it's $30,000 just for the
> back.

Having 30K invested in something is reason enough not to
admit a standard 3k film camera will beat it terms of resolution.

--
LF Website @ http://members.verizon.net/~gregoryblank

"To announce that there must be no criticism of the President,
or that we are to stand by the President, right or wrong,
is not only unpatriotic and servile, but is morally treasonable
to the American public."--Theodore Roosevelt, May 7, 1918