[0144]

This was based on something that Klaus said about 15 megapixel cameras (APS-C) hitting diffraction at f5.6. In his lens testing, he did not see improvements in lens sharpness over f5.6 on the 50D and attributed it to diffraction setting in at that point. I can only imagine what he would see with the 7D.

you should read his tests

for example, canon EF 100mm USM macro(same goes with other lenses i checked)

on 350D(8Mpix) => on 50D(15Mpix)

F4 2003 => 2409
F8 1992 => 2372
F16 1663 => 1973

sooo... where does diffraction make 15Mpix camera equal to 8Mpix one? is the diffraction limit really as close as we're thinking?

i'll do some testing some time, but for now i think there are actually 2 kind of diffraction limits:
1. when lens if rendering not as sharp as on previous stop
2. when extra megapixels give no resolution advantage on same stop.

we can easily see #1 in all tests, but i haven't seen #2 yet.

[infoomatic]

I think this is just a typing mistake ... 12 MP for K-X replacement or the rumored camera between K-7 and K-X. 21 MP on APS-C makes no sense.

[asw66]

I don't pretend to be an expert on either optics or sensor design, but I do know that clever engineers are sometimes capable of coming up with solutions that defy expectations. Even if it is true as a general rule that a greater pixel density means more noise, skilled engineers can sometimes find the exceptions to those general rules.

I think we can all agree that designing a sensor and then processing its signal to produce visible images is a really really complex task. In the midst of all this complexity, the general rule about pixel density might take a back seat to other considerations. We really don't know the details of the precise technical challenges involved, so there's no point in being dogmatic about such things.

Rather than fixate on specific numbers, I suggest that we should be glad that there is an accumulating groundswell of evidence that a new body will be announced soon. Pentax have been hitting one home run after another lately, so we have good grounds for thinking that the new body will be thoroughly drool-worthy, whatever precise form it takes.

[Eruditass]

Did you read the article? Everything being equal, more MP do not imply higher noise provided you are looking at the same image size. Individual pixel noise will rise but that doesn't matter because there are more pixels to average each other out. On the contrary, the noise will look finer-grained.

In what way has Ned Bunnel justified the view of 15MP being a "sweet spot" for APS-C?

I'm going to quote this to make sure people see it. Yes, 100% crops will look worse, but the whole image itself will look better:

For equal, normalized SNR, a high-resolution camera is still better than a low-resolution camera. While it is always possible to simulate a low-resolution camera using a higher resolution camera (since downsampling is easy), it is not possible to simulate a high-resolution camera from a lower resolution one other than by interpolating or inventing data.

This is if the sensor manufacturers don't skimp on the circuits to improve read noise in favor of having more pixels.

---

Last edited by Eruditass; 06-07-2010 at 06:45 AM.

[Class A]

The issue is early diffraction. At present, 7D shooter can't shoot over f4 without diffraction setting in.

Diffraction happens independently of pixel pitch. It is caused by the aperture size and the wave-like nature of light (it can be particle-like as well, but let's not indulge in the Copenhagen interpretation of the nature of light ).

If you increase the pixel-pitch, i.e., sensor resolution you'll be able to see the diffraction effects when pixel-peeping. However, when comparing at the same output size, the higher-resolution sensor will always look better / sharper.

At f/4, you'd need 52MP to see the diffraction effect and a lens that is diffraction limited at f/4. The vast majority of lenses are not diffraction limited at f/4 but exhibit optical aberrations overshadowing the diffraction effect.

The 7D is often criticised for its IQ. I believe it wouldn't be the camera for me but I don't doubt for a split second that "too many pixels" are to blame. It seems Canon has an issue with using two different channels for the green component of the matrix and matching them up takes away from the resolution. IIRC, GordonBGood discussed this ages ago on DPReview.

This was based on something that Klaus said about 15 megapixel cameras (APS-C) hitting diffraction at f5.6.

For green light you need 27MP to capture diffraction effects at f/5.6 on APS-C. See Do Sensors “Outresolve” Lenses? (Tab. 3).

I personally like high-resolution sensors because they (everything else being equal) yield finer-grained noise and have better cropability. Ever found an image in an image and wished you had enough resolution to crop into it? Moreover, in order to increase focal length, I'd rather have 1.4 x resolution I can crop into than put a 1.4 teleconverter between lens and the camera. There is no teleconverter which works as perfectly as cropping.

[Digitalis]

i'll do some testing some time, but for now i think there are actually 2 kind of diffraction limits:
1. when lens if rendering not as sharp as on previous stop
2. when extra megapixels give no resolution advantage on same stop.

we can easily see #1 in all tests, but i haven't seen #2 yet.

I have seen #2 plenty of times. on my 4X5 large format camera I use the Rodenstock APO-Sironar 180mm f/5.6 HR - a diffraction limited lens. though it cannot cover full frame 4X5. In my testing, Diffraction becomes noticeable at f/22 and higher f stops on T-max ISO 100 film. optimal aperture is around f/13 though as Michael Reichmann noticed in LL on a 39 MP digital back the tolerances for diffraction are much lower than that of film

- it would be interesting to use the 180mm APO-Sironar on the Pentax 645, the flange distance is certainly long enough.

[Christopher M.W.T]

Jesus, look people I love this forum and I really respect most of you but these discussions just start turning into pure tech wank and ***** measuring on specifications, numbers, optics, physics etc....just accept that modern engineers can pull off some amazing miracles, enjoy the damn camera and stop been armchair critics.

[lithos]

Jwank

What the hell? Wank is unblocked? Why was I not informed?

Is ****** still blocked?

[ogl]

you should read his tests

for example, canon EF 100mm USM macro(same goes with other lenses i checked)

on 350D(8Mpix) => on 50D(15Mpix)

F4 2003 => 2409
F8 1992 => 2372
F16 1663 => 1973

sooo... where does diffraction make 15Mpix camera equal to 8Mpix one? is the diffraction limit really as close as we're thinking?
.

it's equal to 10 MP Not 8. Of course, we can put even 52 MP sensor in APS-C camera. It could be.
But it's theory. The truth is that 15 MP at APS-C is reasonable limit.
I use all apertures from f/1.8 till f/8. Very rare even f11.

28-30 MP is the reasonable limit for FF. IMO. For full frame MF ~ 80-100 MP.

Pay attention at photozone's tests - for 4/3 - the best apertures - f/2.8-4. Is it good? No way.

[Christopher M.W.T]

What the hell? Wank is unblocked? Why was I not informed?

Is ****** still blocked?

Well now you know better

[mikemike]

LOL!

Actually I own a boot-shaped German beer glass with the Asahi Pentax logo on it (the crowned eye); this was given to representatives on a 60's Photokina. So the story is not very far off

Pictures or its vaporware!!!

[Rondec]

A lot of it does depend on print size. However, whether or not you are printing at a particular size, I think everyone tends to pixel peep a little bit and then you see the effects of diffraction. Of course, if you aren't printing really big or cropping aggressively you won't see it in real life, but it is still there below the surface.

[thibs]

Pictures or its vaporware!!!

It exists, I saw a couple ones on Ebay 4-5 years ago

[Bossy]

Wow...
I was actually at the PMA in Melbourne last sunday and I talked with guy at the Pentax counter.
They seemed to be really restricted in giving information. I asked them about the possibility of a FF coming soon but they answered with a pale face.
Anyway, I've got to see and actually try 645D on my hand so I was extremely satisfied..

[Christopher M.W.T]

Wow...
I was actually at the PMA in Melbourne last sunday and I talked with guy at the Pentax counter.
They seemed to be really restricted in giving information. I asked them about the possibility of a FF coming soon but they answered with a pale face.
Anyway, I've got to see and actually try 645D on my hand so I was extremely satisfied..

Pale face? hmmm be more specific on that one lol