[Clavius]

I ask again, suddenly nobody is talking about diffraction anymore. Why? I know it's really cewl to have 24 megapixels but what's the point if you take physics into the equasion?

Think about the beloved DA 15 ltd. At an APS-C sensor size of 23.6 x 15.6 and an aperture of F8 it will become diffraction limited at 14,9 megapixels. But the DA 15 ltd only gets plenty sharp at F8! So if we want to shoot the nice DA 15 starbursts that we see from F11 and upwards with the K-3 and its 24 megapixels, diffraction should become a bit more prominent then with the K-5 and its 16 MP. If that's really disturbing or not is personal opinion.

Now, the same situation with a 36 x 24 sensor (and 15mm lens that is FF compatible of course) would result in diffraction only past 34.9 megapixels. That's a big difference. This was always the big advantage of larger sensors over smaller ones. Bigger sensor would always be better then the smaller sensors due to physics. But now that the 24mp camera turns out to be APS-C, nobody talks about this disadvantage of crop sensors anymore.

[Digitalis]

But the DA 15 ltd only gets plenty sharp at F8

I would say the DA15 only gets sharp at f/11* - but that is merely an opinion. Yes diffraction will be an issue, but it is a subtle effect; it isn't the brick wall that many perceive it to be.

Bigger sensor would always be better then the smaller sensors due to physics.

This is only true if you don't compensate for the differences in DOF between formats, DOF on 8X10 format is considerably smaller than it is on 35mm, with a 50mm lens at f/5.6 on 35mm format - to get roughly the same DOF as 35mm with a 300mm lens 8X10 you have to use f/32


100% center crops from the Pentax K5IIs - Sigma 8-16mm f/4.5-5.6 @ 12mm

*Or more accurately: diffraction brings the center sharpness down to the point where the corners don't look so bad.

[Clavius]

I would say the DA15 only gets sharp at f/11 - but that is merely an opinion. Yes diffraction will be an issue, it is a subtle thing; it isn't the brick wall that many perceive it to be.


100% center crops from the Pentax K5IIs - Sigma 8-16mm f/4.5-5.6 @ 12mm

Yes, it's completely dependent on perception. That's why I added that it is personal opinion. Diffraction isn't the worst thing that can happen to your photos.

Nevertheless, when people spend thousands of dollars on really very good lenses just for a minor resolution increase, and when people invest lots of money in cameras with no AA filter also for a little increase in resolution, but suddenly stop worrying about it when diffraction calculations come into play does puzzle me.

[awaldram]

Diffraction is relative to sensor resolution that is true

But a 15mm @f8 will not offer lower resolution on a 24mp sensor than the same lens on a 16mp sensor despite refraction

It's the same point being made with regard to low resolution glass it will never perform worse on a higher resolution sensor.

Optimum the sensor would always out resolve the lens even though this would mean diffraction would be happening at any aperture even f1.4

A 24mp aps-C sensor outperforms silver halide in grain.
A 36Mp FF sensor does not.

therefore it pretty obvious that aps-c is pretty much a mature technology approaching its limits (without a tech shift) whereas FF still has a way to go till it at the best it can be,
Once FF hit 50Mp+ things get interesting and I guess FF approaching the same maturity we would be seeing 100Mp FF sensors.

[Digitalis]

nevertheless, when people spend thousands of dollars on really very good lenses just for a minor resolution increase, and when people invest lots of money in cameras with no AA filter also for a little increase in resolution, but suddenly stop worrying about it when diffraction calculations come into play does puzzle me.

I understand, I was rather perplexed when I met a photographer who was so happy with his new toy, the D800E - even though he was his only lens: an APS-C specific zoom lens on it (Nikkor 17-55mm f/2.8G DX) he spent all that money on a new camera to only be able to see at best, half of its full potential.

But a 15mm @f8 will not offer lower resolution on a 24mp sensor than the same lens on a 16mp sensor despite refraction

If you printed both images at 300DPi @ 24"X36" the image from the higher resolution 24mp sensor would look better than the 16Mp image providing the same lens was used at the same aperture. But if you like to pixel peep, then there will be differences between the two files.

It's the same point being made with regard to low resolution glass it will never perform worse on a higher resolution sensor.

A higher resolution sensor will make the flaws in a lens even more obvious. And make correcting them even more tedious - Higher resolution sensors widen the gap between between good glass and excellent glass.

Optimum the sensor would always out resolve the lens even though this would mean diffraction would be happening at any aperture even f1.4

No thanks, to hell with that Idea.

[Clavius]

But a 15mm @f8 will not offer lower resolution on a 24mp sensor than the same lens on a 16mp sensor despite refraction

Who implied that? The lens will offer the same resolution, the results of 15mm F8 + 16mp and 15mm F8 +24mp only differ. And it won't be obvious to the naked eye. Only when you decide to use a heavy crop there will be visible issues.

[awaldram]

" I know it's really cewl to have 24 megapixels but what's the point if you take physics into the equasion?

Read more at: https://www.pentaxforums.com/forums/172-pentax-k-3/240893-school-me-2.html#ixzz2j6zxGceq
"
The point is a 24mp sensor will give more resolution despite diffraction on the 24mp aps-C sensor than it did on the 16Mp Sensor.
That is Physics for you
That final result may be diffraction limited but still it will have a greater lw/ph than the lower resolution sensor irrespective of lens IQ or other factors.

[Clavius]

That final result may be diffraction limited but still it will have a greater lw/ph than the lower resolution sensor irrespective of lens IQ or other factors.

Obviously! But can you do anything usefull with that greater lw/ph? Will those images really be "sharper"?

[rawr]

but suddenly stop worrying about it when diffraction calculations come into play does puzzle me.

Pentax K-3 users aren't the only ones that should be worried about diffraction...

36MP full-frame users should also be worried about it too, but concerns about diffraction don't seem to arise in the Nikon or Sony threads around here or on the web, despite folks sometimes spending many thousands of dollars on FF bodies (D800, A7r) with a pixel density close to 24MP APS-C. That puzzles me too.

Indeed, it is interesting to note that there is never much discussion of diffraction being a major issue for users of 24MP APS-C's like the D3200, D5200/5300, D7100, A77, A65, NEX-7, or indeed users of even smaller sensors like the 16MP micro-4/3's, and even amongst users of compact cameras like the Sony RX100 that sport very high pixel densities.

I guess the reason may be that, as Digitalis says, in all these cases the onset of diffraction is gradual, not immediate, and may only progressively eat away at sharpness past smaller apertures, depending on the lens. So it's not something people may usually see, especially if conditions aren't ideal and other things (motion blur, the use of crappy UV filters on a lens) may cause more problems for image sharpnness than diffraction.

[awaldram]

Obviously! But can you do anything usefull with that greater lw/ph? Will those images really be "sharper"?

errr.....Yes !

Your question was why is diffraction not considered in moving form 16 - 24 Mp on aps-c, we've established its an irrelevance, unless you have something different to compare it to i.e a FF K mount camera .

As no such animal currently exists whatever limits diffraction places on the 15mm @f8 on the k3 is purely speculation though you can say it will not pull the resolution into k5 16mp territory.

[Fastback67]

...therefore it pretty obvious that aps-c is pretty much a mature technology approaching its limits (without a tech shift) whereas FF still has a way to go till it at the best it can be,
Once FF hit 50Mp+ things get interesting and I guess FF approaching the same maturity we would be seeing 100Mp FF sensors.

This is exacty what I was referring to in my original post.

[awaldram]

Well in direct answer ot your question I don't see the end of the road yet for aps-c.

The manufacturers are finding it tougher partly because of technology limits and partly because of expectation.
If I was gambling on my seer ability I'd hazard 35Mp is about the limit for aps-C but with minor gains 50Mp could be a possibility.

As has been mentioned many camera sport 24Mp aps-c sensors the k3 is late to the party and reportedly carries a later version Sony sensor hence the interest in its ISO performance, Nikon rejected its older generation units in favor of Toshiba sourced Cmos (with its own problems)

[rawr]

In the case of some of the modern APS-C's like the K-3 or K-5IIs, I guess the boost in resolution from taking away the AA filter would also have a degree of offsetting impact on any decline in image resolution induced by diffraction. Although I haven't seen any tests, this might be visible when looking at the resolution curve of a lens when mounted on a K-5/D800 compared to a K-5IIs/D800E.

The impact of diffraction may also be offset by sensor type - eg whether the sensor is Bayer or non-Bayer (eg Foveon), or uses some sort of new micro-lenses. In future 24MP+ APS-C sensors might all be Foveon, who knows, with a completely different diffraction profile than current sensor tech.

[ChrisJ]

The one huge point being missed here (as far as I can see), that's the difference the Sony Exmor sensor gives. Pixel size is larger with these sensors, 95%+ of the surface area of the sensor is sensitive to light as opposed to only 25% on older sensors as still currently used by Canon. 75% of the surface area of their sensors is taken up with wiring and switching transistors required when the image is scanned off the sensor and sent to the A/D converter.

The Sony Exmor sensors have moved this wiring and transistors to layer on the back of the sensor so the pixels on these sensors are physically larger than those on older sensors, four times larger!

This really is a game changer because now you have pixel size that is as large or larger than FF sensors, noise and high ISO performance is at least as good as FF sensors. Diffraction when using small apertures is less of a problem, Moire is less of a problem, in short a lot of the reasons for having to use an AA filter are much reduced.

It wasn't long before Pentax engineers removed the AA filter, for the first time in the digital era images are sharp from camera, there is no sacrifice in resolution, clarity is visibly improved, for the first time we are looking at the image the sensor actually sees, not the image from blurred pixels caused by the AA filter that was needed on older sensors.

The K5-lls does have some Moire problems, back lit hair looks atrocious due to the Moire, with the K3 being able to turn the AA filter off and on is a real plus and unique to Pentax (I wonder if they've patented it?), it remains to be seen if 'waggling' the sensor really does work as well as it should. We'll see.

Chris