[Zephos]

I watched this video and I'm suddenly not too eager to buy old glass that was made for full frame cameras. I hope Pentax continues to offer APS-C optimized lenses, now that the FF is coming. Here's the link:

You Tube

You Tube

There's also some interesting stuff about how an APS-C camera can be better that FF when it comes to cropping.

[VoiceOfReason]

I have a mix of ff and apsc lenses. I think the better thing is to get good quality, regardless of format optimization, and make sure it works with what you use and that you're happy with what you have and the result it gives.

[MadMathMind]

"Put it on a Nikon or Sony crop body with a 1.5x crop factor"...

Guess he doesn't know about Pentax either.


I also think his explanation of why he dislikes FF lenses on crop sensors. He cites a 24-120 Nikon lens giving 13 MP on a FF body and 9 MP on a crop sensor. But he fails to account for the fact that the image circle of the FF sensor is large--so of course the lens has a higher resolution in terms of MPix. But it's like saying that a 4000x3000 image will be sharper than a 2000x1000 one because the first one is 12 MP and the second is just 2 MP.

The image circle of a full frame lens is 2.25x times the size that of an APS-C one. If the lens were performing equally well across the frame, that would mean that the full-frame should pull just over 20MP. But it does just 13MP. That means that outside the APS-C area, the lens has just an extra 4 MP of resolution.

I would like to think that an expert wouldn't fall into this trap of using pixels and DXO scores alone to measure sharpness. He is right that pixels = resolution, but willfully neglects that resolution != sharpness to make his point.

And of course, the second part about cropping and detail level is a bit vacuous. If everything is viewed at 100%, he is right about pixel count. But few people do that. When you have to take everything to 1920x1080 for screen viewing or print at 5x7 at 300dpi (1500x2100), then it's not going to matter much. All the extra is going to be lost in interpolation during downsizing.


Discussion of crop factor is such a disaster. I blame Canon for this. If they didn't use smaller sensors, it wouldn't really be necessary because all APS-C bodies would be the same. As it is, what you see on Canon and Nikon bodies with the same lens can differ quite a bit as lenses get longer (e.g., a 200mm lens on Nikon is equivalent in FoV to 187mm on Canon) and thus warrants all this discussion that invariably confuses people.

[aleonx3]

It is the opposite, I rather have a FF lens on APS-C camera than a lens optimized for APS-C sensor.

[biz-engineer]

Just doing some research on diffraction limited resolution and MTF charts. Generally the lens peak resolution is at 2 stops from wide open. With 24 Mpixels APS-C sensor it's best to avoiding diffraction by opening at f/5.6 or f/4 . Conclusion: to benefit from the 24Mpix resolution, lenses should be f2.8 or faster.

Lenses that are best for the K-3: Pentax DA*16-50, Tamron/Sigma 17-50, + limited lenses that peak 50lpmm @ f/5.6 .
On the other hand, Pentax 17-70 , Pentax 18-135, Pentax 18-55, Pentax 16-85 are lenses that have best resolution by stopping down to f/8 where diffraction appears, hence good enough for the K-5, K30, K50 etc 16Mpix sensor.
In addition, faster lenses provide more light, i.e ability to shoot at lower iso, hence less noise.

Fast lenses provide higher IQ right away. I started with kit lenses but I believe I should have started immediately with high end glass. Low end glass is cheaper but ultimately tends to be a waste of money and also is harder to sell on second hand market.

[kiwi_jono]

There are some benefits using some crop lenses on a crop body (versus FF lens) but the biggest advantages of crop lenses are size, weight and cost.

As far as sharpness goes it all depends on the specifics of the lens involved. For example I recently got a Tamron SP 70-270 f2.8 (a FF lens) and its the sharpest lens I own - no question.

Another point is where he is talking about the FF crop mode versus crop camera. Sure you may end up a with a lower resolution in FF crop mode but in this case the pixels are still bigger than crop camera which has advantages over the APS-C body in low light / high ISO performance. Its not just about maximum resolution in my view.

[blende8]

Basically the smaller the sensor, the better resolving the lens must be.
An APS-C lens must have a higher resolution for a APS-C sensor than a FF lens for a FF sensor, if both sensors have the same number of MP.
That we do not notice this when we mount a FF lens on an APS-C camera is because only the "sweet spot" = center of the lens is utilized for the image.

This is the same when using MF lenses on a FF body. It normally works, but you do not get a resolution benefit.

[Sagitta]

Unless you're printing your 24MP image at full size and full resolution, all this is mostly pointless anyway as you're going to be shrinking your image to print or your printer is going to be flushing resolution anyway to get to its print size.

That miniscule difference in IQ is going to be unnoticeable to the average viewer except for pixel peepers viewing the thing online.

[Stone G.]

Basically the smaller the sensor, the better resolving the lens must be. An APS-C lens must have a higher resolution for a APS-C sensor than a FF lens for a FF sensor, if both sensors have the same number of MP.

True - and that's irrespective of whether we talk analogue film or digital sensors: If we print/view to the same size, the resolution that matters is lines/(image height). Thus, lens resolution in terms of lines/milimeter is less demanding for the larger formats.

That we do not notice this when we mount a FF lens on an APS-C camera is because only the "sweet spot" = center of the lens is utilized for the image.

This is a comon misconception. The entrance pupil for, say, a 50mm f/2 lens is the same (25mm) for both an APS-C and an FF-optimized lens and when used wide open the entire pupil - and hence, the front element - is used to form image points on the film/sensor. Even when the lenses are stopped down, we use substantial parts of the pupil and the front element near the rim to form off-axix image points - see for example: Misconceptions in photographic optics

[biz-engineer]

Why optimized for APS-C lenses matter?
Optimized for APS-C does not exist. APS-C systems are FF system geometry with smaller sensor in it.

IQ wise, independant from format , due to optical bandwidth of glass, the lens center resolutions ranges from 40lpmm (average glass/zooms) to 50lpmm (good glass, prime lenses), that is to say 20% improvement in sharpness. But from f/5.6 (e.g HD Pentax 16-85) to f/2.8 lens, that is 400% more brightness, e.g with f/5.6 you use iso 400 and with f/2.8 you use iso 100.
Better go for FF fast glass.

Comparison:
System (A): Pentax K-S2 + 18-50 versus
System (B): Pentax FF with 24-70 f/2.8.

System (B) provides 4 times more light + 2 times better sensor = 8 x better than System (A)

APS-C is inferior system.

[PJ1]

He lost me the moment he said you multiply the aperture by the crop factor. He obviously doesn't understand aperture. Using his mathematics, adapting a 70-200/f2.8 to a Pentax Q will give you a 70-200/f16 wide open! His mathematics are rubbish because that is not the way aperture works. Aperture gives an certain intensity of light within a given image circle. No matter what piece of that image circle you take (what the angle of view or "crop factor" is) the intensity and therefore the aperture remain the same. If that was not the case, we would need different exposure times for different parts of the image circle.

[Sandy Hancock]

Ah grasshopper, did you not understand that depth of field is the only parameter of importance in photography?

He lost me the moment he said you multiply the aperture by the crop factor. He obviously doesn't understand aperture. Using his mathematics, adapting a 70-200/f2.8 to a Pentax Q will give you a 70-200/f16 wide open! His mathematics are rubbish because that is not the way aperture works. Aperture gives an certain intensity of light within a given image circle. No matter what piece of that image circle you take (what the angle of view or "crop factor" is) the intensity and therefore the aperture remain the same. If that was not the case, we would need different exposure times for different parts of the image circle.

[ppppsssstttt]

Ugh I hate this discussion, but I'll participate anyways

The only advantage for an optimized APSC lens is that you can have wider angles on an APSC camera. If you stick with full frame lenses, good luck finding wider than 24mm without being a fisheye. But you we have the 15mm and 14mm DA lenses without being fishy.

Everything else is a push. 28mm is 28mm, f/1.8 is f1.8. The amount of light gathered at the sensor per area is the same. Your ability to stop motion and use a higher shutter speed does not change.

It only gets slightly confusing if you're comparing equivalent field of view and depth of field between the two systems.

Otherwise, the resolution of the lens is.. the resolution of the lens. And the detail in your picture is limited by your limiting step. Are you limited by the sensor or the lens? Or are you limiting yourself?

[Sandy Hancock]

Equivalence = thread death

Next....

[blende8]

This is a comon misconception.
...

Ok, well, let me correct, then:
"... because only the "sweet spot" = center of the image is utilized."