The Megapixel Race

ASAAR wrote:
> On Tue, 31 Oct 2006 08:11:16 +0100, Alfred Molon wrote:
>
>>>The diffraction spot *diameter* size in photography, for 555nm (green
>>>light),
>>>is equal to 1.22 x 0.555 x 2 x 'f/number' = 1.3542 x 'f/number' (in
>>>microns).
>>Does this mean that a camera with 6 micrometer pixels (as most APS-C
>>DSLRs) is diffraction limited at something like F4.5 (6/1.35) ?
>
> Someone recently posted a link (possibly in this thread) to a page
> titled "Optimum Aperture - Format size and diffraction". It has one
> table linking "Lens Aperture" to the "Diffraction Limited Spot Size
> (microns)" and another table having three columns, "Format Size"
> (film or sensor size), "Typical COF value (microns)" and "Limiting
> Aperture for a sharp 8 x 10 print". According to this table, full
> frame 35mm (36mm x 24mm) has a limiting aperture of f/22, APS-C
> (22.5mm x 15mm) has a limiting aperture of f/16, and 1/1.8" (7.1mm x
> 5.3mm) and 1/2.5" (5.7mm x 4.3mm) reach their limits at f/5 and f/4,
> respectively. Notice that these values relate the Circle of
> Confusion needed for sharp 8" x 10" prints to the film or sensor
> size, and isn't relating anything to film grain size or sensor pixel
> size.
>
> http://www.bobatkins.com/photography/technical/diffraction.html
>

This reasoning is more appropriate to film, though; in digital, if the
diffraction spot is smaller than your sensels, you're probably ok no
matter what the enlargement size is (well, assuming you have enough pixels).

Mind you, I did not read the link, so I go by what I think you said.
Please correct if wrong.
 >> Stay informed about: The Megapixel Race 

On Tue, 31 Oct 2006 14:19:18 +0100, acl wrote:

>> Someone recently posted a link (possibly in this thread) to a page
>> titled "Optimum Aperture - Format size and diffraction". It has one
>> table linking "Lens Aperture" to the "Diffraction Limited Spot Size
>> (microns)" and another table having three columns, "Format Size"
>> (film or sensor size), "Typical COF value (microns)" and "Limiting
>> Aperture for a sharp 8 x 10 print". According to this table, full
>> frame 35mm (36mm x 24mm) has a limiting aperture of f/22, APS-C
>> (22.5mm x 15mm) has a limiting aperture of f/16, and 1/1.8" (7.1mm x
>> 5.3mm) and 1/2.5" (5.7mm x 4.3mm) reach their limits at f/5 and f/4,
>> respectively. Notice that these values relate the Circle of
>> Confusion needed for sharp 8" x 10" prints to the film or sensor
>> size, and isn't relating anything to film grain size or sensor pixel
>> size.
>>
>> http://www.bobatkins.com/photography/technical/diffraction.html
>>
>
> This reasoning is more appropriate to film, though; in digital, if the
> diffraction spot is smaller than your sensels, you're probably ok no
> matter what the enlargement size is (well, assuming you have enough pixels).
>
> Mind you, I did not read the link, so I go by what I think you said.
> Please correct if wrong.

I don't see anything that needs correcting, although I'm not sure
that the information on Atkin's page is more appropriate for film
than for digital sensors. The sensor's sensels could be thought of
as a special case of film's grain, being uniform in size and with
regular spacing. Digital thus lacks film's random grain structure
which probably contributes to the creation of some kinds of
artifacts, such as moire, depending on the subject being
photographed. The Lens Aperture vs Diffraction Limited Spot Size
table only lists apertures from f/4 to f/128 and the corresponding
spot sizes range from 5.4 to 172 microns. Unless my very rough
calculations are off, I get sensel sizes of about 2.5 microns for my
1/2.5" sensor, 6.3 microns for a typical 8mp APS sensor, and 9
microns for a 12mp FF sensor.

So the diffraction spot won't be smaller than the sensels used by
my P&S for any of its apertures, since its maximum aperture is only
slightly greater than f/4. But the diffraction limited spot size
can easily be smaller than the sensel size for APS-C and FF cameras
given a large enough lens aperture. But if my estimates for sensel
sizes is approximately correct, the limiting aperture for 8mp APS-C
cameras falls between f/4 and f/5.6, and the limiting aperture for
12mp FF cameras falls between f/5.6 and f/8. Does this sound right?
If so, I'm not sure how useful any of this is if it indicates that a
1/2.5" sensor can't produce "reasonably sharp 8" x 10" prints",
since its diffraction limited spot sizes would be much greater than
its sensel size for all apertures based on this quote, from the same
Atkin's web page:

> So what has this to do with sharpness and stopping down? Well
> we've already seen that the COF defines the spot size which is
> acceptable for reasonably sharp 8x10 prints.

It might even be used to suggest that a 1/2.5" sensor wouldn't be
able to produce acceptably sharp 4" x 6" snapshots if apertures
greater than f/4 were used. Looking back again, that appears to be
exactly what is implied, since Atkins also says:

> So the table indicates that for an APS-C DSLR (like the EOS 20D,
> or in fact any Canon, Nikon, Minolta, Pentax or Sigma DSLR except
> for the Canon EOS 1D and 1Ds series), the smallest aperture that
> should be used is f16, assuming you want a reasonably sharp 8x10
> print. If you stop down more and still want a sharp print, you may be
> limited to 5x7 or 4x6 print size.
 >> Stay informed about: The Megapixel Race 

"David J. Littleboy" <davidjl.DeleteThis@gol.com> wrote in message
news:ei7p2f$teo$1@nnrp.gol.com...
SNIP
> In real life, 5D images look _great_ at f/16, still pretty good at
> f/22, but noticeably soft at f/32.

That makes some sense to me. The sensel pitch of the 5D is approx. 8.2
micron which corresponds to a diffraction spot diameter from an
aperture of f/6.1, the AA-filter will allow 'almost' double that to
f/12.1 before it becomes visible. So the theoretical estimate would
suggest an optimum around f/11, gradually getting worse as you select
smaller apertures.

--
Bart
 >> Stay informed about: The Megapixel Race 

In article <ei7p2f$teo$1@nnrp.gol.com>,
"David J. Littleboy" <davidjl.RemoveThis@gol.com> wrote:

> "ASAAR" <caught.RemoveThis@22.com> wrote:
> >
> > So the diffraction spot won't be smaller than the sensels used by
> > my P&S for any of its apertures, since its maximum aperture is only
> > slightly greater than f/4. But the diffraction limited spot size
> > can easily be smaller than the sensel size for APS-C and FF cameras
> > given a large enough lens aperture. But if my estimates for sensel
> > sizes is approximately correct, the limiting aperture for 8mp APS-C
> > cameras falls between f/4 and f/5.6, and the limiting aperture for
> > 12mp FF cameras falls between f/5.6 and f/8. Does this sound right?
>
> I think you need to redo your calculations. In real life, 5D images look
> _great_ at f/16, still pretty good at f/22, but noticeably soft at f/32.
>
> David J. Littleboy
> Tokyo, Japan

I could live with those numbers, but does Canon not still lack good
wide angle lens available?
--
"As democracy is perfected, the office represents, more and more closely,
the inner soul of the people. We move toward a lofty ideal. On some great
and glorious day the plain folks of the land will reach their heart's desire
at last, and the White House will be adorned by a downright moron."
- H. L. Mencken, in the Baltimore Sun, July 26, 1920.


Reality-Is finding that perfect picture
and never looking back.

www.gregblankphoto.com
 >> Stay informed about: The Megapixel Race 

In article <45478074$0$335$e4fe514c@news.xs4all.nl>, Bart van der Wolf
says...
>
> >> The diffraction spot *diameter* size in photography, for 555nm
> >> (green light), is equal to 1.22 x 0.555 x 2 x 'f/number' =
> >> 1.3542 x 'f/number' (in microns).
> >
> > Does this mean that a camera with 6 micrometer pixels (as most
> > APS-C DSLRs) is diffraction limited at something like F4.5 (6/1.35)
> > ?
>
> Yes, but the effect of residual Chromatic aberrations probably still
> exceeds the diffraction effects. It would require a smaller aperture
> in many cases to find the optimal balance between the two netative
> contributors to image resolution.
>
> What's more, reliably resolving a feature requires, according to the
> Nyquist frequency, 2 pixels per cycle. So when you use 12 micron /
> 1.3542 you'll get f/8.9 (which often corresponds quite well with the
> optimal trade off between diffraction and CA). Besides, the AA-filter
> should make sure there is little detail that is finer than that.
>
> Therefore, the resolution/quality will peak somewhere between f/4.5
> and f/8.9, probably closer to the latter due to the AA-filter.

Well, I remember I read somewhere the rule of thumb 1500/N for the
diffraction limit, where N is the F-number. Here is the site:
http://www.graflex.org/lenses/photographic-lenses-tutorial.html

According to this formula, at F8 the diffraction limit will be 188
lp/mm, which corresponds to a 2.7 micrometer pixel CCD. A CCD with 3
micrometer pixels (= 166 lp/mm) will be diffraction limited at F9
(1500/166), while one with 6 micrometer pixels will be diffraction
limited at F18.
--

Alfred Molon
------------------------------
Olympus 50X0, 7070, 8080, E300, E330, E400 and E500 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
Olympus E330 resource - http://myolympus.org/E330/
 >> Stay informed about: The Megapixel Race 

"Greg "_"" <grey_egg.TakeThisOut@greg_photo.com> wrote:
>>
>> I think you need to redo your calculations. In real life, 5D images look
>> _great_ at f/16, still pretty good at f/22, but noticeably soft at f/32.
>
> I could live with those numbers, but does Canon not still lack good
> wide angle lens available?

Depends what you mean by good. The 24/2.8, 24/1.4, and 35/1.4 are as good as
it gets. And you can use your 645 Mamiya or Pentax 35/3.5 as a shift lens.

Even the oft-insulted 17-40 produces fine images if you know what you are
doing. Boring shots, but the lens is doing just fine.

http://www.pbase.com/davidjl/image/59225146/large

http://www.pbase.com/davidjl/image/69033566/large

(The original of the latter one is full res so you can see the corners,
although only the lower left has any detail worth looking at.)

David J. Littleboy
Tokyo, Japan
 >> Stay informed about: The Megapixel Race 

In article <ei8o8h$6r8$1@nnrp.gol.com>,
"David J. Littleboy" <davidjl RemoveThis @gol.com> wrote:
>
> http://www.pbase.com/davidjl/image/69033566/large
>
> (The original of the latter one is full res so you can see the corners,
> although only the lower left has any detail worth looking at.)
>
> David J. Littleboy
> Tokyo, Japan

Not wide enough - not a cut I got an 11-18 Tamron SP lens two weeks
ago and I really like wide angles, If not careful I can shoot my feet!
Its not quite fish eye....but very wide. Getting the exact f'stop and
focus set to maximize Dof and sharpness is somewhat a chore. I also want
more Dof field and sharpness to boot. Shooting 4x5 these twenty years
has done this damage to me
--
"As democracy is perfected, the office represents, more and more closely,
the inner soul of the people. We move toward a lofty ideal. On some great
and glorious day the plain folks of the land will reach their heart's desire
at last, and the White House will be adorned by a downright moron."
- H. L. Mencken, in the Baltimore Sun, July 26, 1920.


Reality-Is finding that perfect picture
and never looking back.

www.gregblankphoto.com
 >> Stay informed about: The Megapixel Race 

"Greg "_"" <grey_egg.TakeThisOut@greg_photo.com> wrote:
> "David J. Littleboy" <davidjl.TakeThisOut@gol.com> wrote:
>>
>> http://www.pbase.com/davidjl/image/69033566/large
>>
>> (The original of the latter one is full res so you can see the corners,
>> although only the lower left has any detail worth looking at.)
>>
>> David J. Littleboy
>> Tokyo, Japan
>
> Not wide enough - not a cut I got an 11-18 Tamron SP lens two weeks
> ago and I really like wide angles, If not careful I can shoot my feet!

The 17-40 on FF is indistinguishable from the 11-18 on 1.5x APS-C in terns
of wideness. (I must have gotten tired of super-duper-super wide; while my
10-22 stayed put at 10mm all the time it was on the camera, the 17-40 gets
used throughout its range on the 5D.)

The one disappointment of the 17-40 on FF vs. the Canon 10-22 on APS-C, is
that while the 17-40 focuses ridiculously close, the 10-22 focuses insanely
close.

The Sigma 12-24 is a real 12mm on the 5D, and is very good in its lack of
rectilinear distortion. A good copy shot at f/16 produces amazing images on
a 5D or 1Dsmk2.

> Its not quite fish eye....but very wide. Getting the exact f'stop and
> focus set to maximize Dof and sharpness is somewhat a chore. I also want
> more Dof field and sharpness to boot. Shooting 4x5 these twenty years
> has done this damage to me

At f/11 and 11mm, your DOF should be from quite close to infinity even when
focussed at infinity. At 17mm on FF, at f/11, the DOF calculator at hand
claims 4 feet to infinity, and in my tests, the 17-40 just about does that.
(I was surprised, since my experience with DOF calculators for MF is that
they're way overoptimistic.)

David J. Littleboy
Tokyo, Japan
 >> Stay informed about: The Megapixel Race