which camera quiz

[Roger Clark]

Roger,
Thanks for doing the test, AFAK when adding 1.4X TC light drops by a full stop but magnification increases by 1.4X so I would expect the MTF to drop by 2/1.4=1.43 relative to the prime, if the TC was perfect, in your test it drops by 12% so can you elaborate on the method you extracted MTF? perhaps images of the test strip if possible?

Dan,
This baffles me why you would expect the equation for MTF to be dependent on light level. It matters not that you lose a stop in light. One compensates by a longer exposure so you record the same light level. Even so, the equations for MTF are relative and do not depend on exposure or light level (except in cases of excessive noise will mask low MTF).

Second, if the lpm dropped 1.4x when adding a TC, that would mean zero additional detail on the subject. If that were the case, people would not see more detail on their subjects when using a TC. Clearly that is not the case. I use a TC a lot because I DO get more detail on the subject. And that is exactly what the MTF curves I derived for my camera setup show. I see little degradation in IQ with a 1.4x TC on super telephotos, and a little with a 2X but still well worth it. That again is what the MTF curves say I should observe.

Also taking the above data, when adding 1.4X TC at nyquist limit MTF is about 5%, so the part that is blur filter limited is between MTF10 and MTF5 which I argue and I think you agree is negligible for practical purposes. so it is very clear that for a tighter pixel camera Nyquist is way past MTF0 point, and unless the 7D has an unusually strong blur filter that I doubt based on the samples I have seen, it will be diffraction limited. And this was with 300 f/2.8 lens which I believe if not sharper, is as sharp as 500 and 600 primes. Based on this data there isn't much improvement in the small pixel camera when you add the TC. Of course pixels resolving diffraction blur are still better than interpolated pixels, so I believe if you want to upsample the photos, it has an advantage but not otherwise.

Yes, I would agree that between MTF10 and MTF5 is negligible difference. That was not my point. You said MTF 50 was the metric and there is little information at higher lpm. I countered and said that is where the fine detail is. One could block average 2x2 pixels and downsize all images with a 300 f/2.8 and 2X TC on a 5D2. That would make a hard cut MTF50 filter. Would the resulting image show the same fine detail? No. I would consider such images degraded. Try the same experiment with ISO 100 film and the same lens+TC and scan the film on a drum scanner at 21 megapixels, and you won't see much difference.

The idea of MTF50 being the limit for optical quality dates back to the film area. But digital image quality has moved way beyond film. It's time for the old concept to be modified. Digital also has much higher signal-to-noise ratios than film, so the use of unsharp mask and more sophisticated methods, like Richardson-Lucy, boosts the low MTF at the highest frequencies making them more important. One couldn't do that with film because the S/N was not there.

Regarding my method. It is a pretty standard bar chart, measure the minimum and maximum at multiple points and compute the MTF. The zero and 100% levels were determined by large black and white areas nearby the test target.

I have heard people comment the 7D blur filter is stronger. I would predict that use of a TC on a 7D will help resolve more detail on a subject with a good lens, but it will be a small gain (smaller than on cameras like the 5D2). The use of a 2x TC will not likely improve detail on a subject over a 1.4x TC (maybe very very slightly). I think that is what people are observing.

Roger

[Roger Clark]

I would have said, the pixel pitch and the blur filter (and of course, the two are correlated; the strength ie MTF50 point of the OLPF is set by the pixel pitch). But if we treat the pixels as a box blur of an incoming analog signal, they also degrade MTF relative to infinitely fine sampling.

Hi Emil,
Certainly pixel pitch is an effect, but if the blur filter were not there, the MTF lines would shift right. The blur filter is having a greater effect than the pixels. But you are correct, the pixels have a contribution too.

Again this logic leads to an MTF curve that goes like 1/x, so I am puzzled where it is coming from.

I don't understand where my logic dictates a shape to the curve. I simply stated what is observed at a given MTF level (MTF50). Maybe you took my statement out of context. I was talking about MTF50

I would have said that a perfect 1.4x TC would slide you along the MTF curve by a factor 1.4 by simply magnifying everything in the image by that factor on the sensor.

I'm not sure what you are saying. We clearly see the curves moving to the left with added TCs.

The test of whether the images are diffraction limited is to stop down a stop and see that the resolution has dropped by 1.4x, not to change the magnification by 1.4x and stop down a stop and see that the resolution has dropped by 1.4x. But that is a different test than the one that was performed.

On can compute what the MTF50 diffraction limit would be (or any other MTF level). I did that for the 2X TC, and it is falling right on the value expected by diffraction. There is no need to spend time doing more measurements.

That may be because the lens is wide open. Most people argue that results are better with the 2X TC when stopping down some. Again it would be interesting to see the above graph repeated at one stop and two stops down from wide open to see what diffraction is really doing, vs say the effects of lens/TC aberrations.

I see a slight increase in quality with my 500 f/4 and a 1.4x TC: stopping down to f/6.3 is a slight improvement. I haven't seen that on my 300.

My take is the following: MTF50 goes from about 60 lpm bare to about 55 with the 1.4x and 40 with the 2x. That means that with the 1.4x TC a given object is magnified by 1.4x with a 1.2x loss in resolution at the sensor, meaning that the object is resolved 1.2x better. With the 2x TC, the gain is 2x magnification vs 1.5x loss in resolution on the sensor plain, for again of 1.33x in resolution of the subject. I'd be interested to see if those numbers improve if one stops down, as many people seem to say.

Going from 60 to 55 is only a 9% drop, not 20%, so with a 1.4x TC, that would be 1.3x improved detail on the subject. I dreived 12% loss, not 9, but 9 versus 12 are close and indicate a small loss. I agree with your 2x TC numbers.

Roger

[Dan Avelon]

Roger,

My choice of word was perhaps poor, what I meant was loss of contrast not light, you are right MTF is independent of exposure. Maybe I overestimated the amount of contrast degradation with the TC which I had read was proportional to light loss in an old optics book. I only have 1.4X TC, I tried the 2X but images were hazy and useless IMO.

I agree that you can extract detail beyond MTF50, let's take your data, 7D Nyquist is at 116 lpm, 5DMKII is 78lpm and the MTF0 is at 80lpm for lens + TC. There isn't really much there to sharpen, also what about noise? In practice you need to apply some sort of noise reduction to compensate for the low SNR of the small pixel camera, and noise reduction also destroys fine detail.

I seriously doubt that there is any real world advantage here based on your MTF plot, have you used the TC with your 7D in the field where noise and focus tolerance were also variables? Did you observe any meaningful advantage over your 5D?

Cheers,
Dan

Dan,
This baffles me why you would expect the equation for MTF to be dependent on light level. It matters not that you lose a stop in light. One compensates by a longer exposure so you record the same light level. Even so, the equations for MTF are relative and do not depend on exposure or light level (except in cases of excessive noise will mask low MTF).

Second, if the lpm dropped 1.4x when adding a TC, that would mean zero additional detail on the subject. If that were the case, people would not see more detail on their subjects when using a TC. Clearly that is not the case. I use a TC a lot because I DO get more detail on the subject. And that is exactly what the MTF curves I derived for my camera setup show. I see little degradation in IQ with a 1.4x TC on super telephotos, and a little with a 2X but still well worth it. That again is what the MTF curves say I should observe.




Yes, I would agree that between MTF10 and MTF5 is negligible difference. That was not my point. You said MTF 50 was the metric and there is little information at higher lpm. I countered and said that is where the fine detail is. One could block average 2x2 pixels and downsize all images with a 300 f/2.8 and 2X TC on a 5D2. That would make a hard cut MTF50 filter. Would the resulting image show the same fine detail? No. I would consider such images degraded. Try the same experiment with ISO 100 film and the same lens+TC and scan the film on a drum scanner at 21 megapixels, and you won't see much difference.

The idea of MTF50 being the limit for optical quality dates back to the film area. But digital image quality has moved way beyond film. It's time for the old concept to be modified. Digital also has much higher signal-to-noise ratios than film, so the use of unsharp mask and more sophisticated methods, like Richardson-Lucy, boosts the low MTF at the highest frequencies making them more important. One couldn't do that with film because the S/N was not there.

Regarding my method. It is a pretty standard bar chart, measure the minimum and maximum at multiple points and compute the MTF. The zero and 100% levels were determined by large black and white areas nearby the test target.

I have heard people comment the 7D blur filter is stronger. I would predict that use of a TC on a 7D will help resolve more detail on a subject with a good lens, but it will be a small gain (smaller than on cameras like the 5D2). The use of a 2x TC will not likely improve detail on a subject over a 1.4x TC (maybe very very slightly). I think that is what people are observing.

Roger

[Roger Clark]

Roger,

My choice of word was perhaps poor, what I meant was loss of contrast not light, you are right MTF is independent of exposure. Maybe I overestimated the amount of contrast degradation with the TC which I had read was proportional to light loss in an old optics book. I only have 1.4X TC, I tried the 2X but images were hazy and useless IMO.

I agree that you can extract detail beyond MTF50, let's take your data, 7D Nyquist is at 116 lpm, 5DMKII is 78lpm and the MTF0 is at 80lpm for lens + TC. There isn't really much there to sharpen, also what about noise? In practice you need to apply some sort of noise reduction to compensate for the low SNR of the small pixel camera, and noise reduction also destroys fine detail.

I seriously doubt that there is any real world advantage here based on your MTF plot, have you used the TC with your 7D in the field where noise and focus tolerance were also variables? Did you observe any meaningful advantage over your 5D?

Cheers,
Dan

Dan,
There have been many examples of images people have posted to BPN where they compared various images, including the 7D with TCs. Recently there was a thread on the 7D + 1.4x TCs, e.g.:
http://www.birdphotographers.net/for...ad.php?t=58529

In that thread, Arash stated "1.4 does work well with 500 on 7D though, there is a little bit loss of quality but not very noticeable."

Roger

[Manas Khan]

I am not regular here... was just going through this thread now....

First, among the two 100% crop samples provided by Arash, (to my eye) the 7D sample seems to have a bit more details (resolved lines) - at the shades above the left eye and the curves at the upper right corner.

Secondly, the sample "With 1.4X TC II, 1/80 sec f/5.6 ISO 800." as provided by OP in the thread http://www.birdphotographers.net/for...ad.php?t=55413 (#12) seems to be better than the one shown here. I know that one was a new dollar bill whether the one posted in this thread is an old one. But when the OP wanted to compare the resolving power of two cameras with 500mm f4 + 1.4X TC wide open, why did he use a "low contrast target" ? If one is interested in the real world image quality of low contrast subjects (or subjects without finer details) then it's completely a different thing... but in the case of a experiment to show the power of an optical system (camera + lens system at a defined aperture value) to resolve lines per mm and to support that result with theoretical data one should probably choose a better (more ideal), high contrast target - if not a proper test chart then at least a new dollar bill. In that case it would have been easier to see if any of the samples (among the 5D II and the 7D) shows better details...

It would be interesting to see the results if someone does the same test with a better target.
Reading articles / posts in different forums on the net and seeing some samples it seems many are happy with 500mm + 1.4X TC (even wide open condition) on 7D.... even the OP has probably expressed the same feeling in the other thread.

regards,
Manas.

[Dan Avelon]

Dan,
There have been many examples of images people have posted to BPN where they compared various images, including the 7D with TCs. Recently there was a thread on the 7D + 1.4x TCs, e.g.:
http://www.birdphotographers.net/for...ad.php?t=58529

In that thread, Arash stated "1.4 does work well with 500 on 7D though, there is a little bit loss of quality but not very noticeable."

Roger

Thanks Roger, but that thread does not compare 7D samples with a large pixel camera like the 5DMKII and 1D series side by side which is what I am interested to see. In fact the thread starts with another poster complaining that his TC is no longer as sharp as when he was using it with larger pixel cameras.

Dan

[Emil Martinec]

Hi Emil,
I'm not sure what you are saying. We clearly see the curves moving to the left with added TCs.

I was saying the same thing from a different perspective. An ideal TC would simply magnify the image by 1.4x or 2x; therefore pick a resolution R without the TC and find the MTF; the prediction for the MTF with an ideal TC added would be obtained by looking at the same MTF curve at resolution 1.4*R or 2*R, which is to the right on the curve. To know where this lands you, you need to know the slope of the MTF curve (which seems to be rather straight near the resolution cutoff in your plot).

On can compute what the MTF50 diffraction limit would be (or any other MTF level). I did that for the 2X TC, and it is falling right on the value expected by diffraction. There is no need to spend time doing more measurements.

If the MTF curve is diffraction limited, then extinction resolution should be about when the first minimum of one diffraction spot is at the maximum of a neighboring spot



Here the continuous curve is the diffracted intensity, the dots are a box blur discrete samples representing the values of ideal pixels spaced by the radius of the diffraction spots; seems to me that this is essentially the extinction point. The radius of the Airy disk for f4 is about 2.7µ or 185 lpm. The fact that you are only getting about half of that seems to me an indication that the pixel size of 6.4µ is having a big impact (I would also think demosaic limits the resolution as well). But perhaps there are other effects not accounted for in this idealized approximation.

[Roger Clark]

I was saying the same thing from a different perspective. An ideal TC would simply magnify the image by 1.4x or 2x; therefore pick a resolution R without the TC and find the MTF; the prediction for the MTF with an ideal TC added would be obtained by looking at the same MTF curve at resolution 1.4*R or 2*R, which is to the right on the curve. To know where this lands you, you need to know the slope of the MTF curve (which seems to be rather straight near the resolution cutoff in your plot).

Emil,
I disagree that the ideal TC added would shift the MTF curve 1.4x or 2x right for 1.4x and 2x TC, respectively. You can't really shift the curve to the right because pixel spacing and the blur filter limits it. The real world example I showed was the 1.4x TC on the 5DII which shifted the MTF curve only 12%. If the MTF curve stayed the same (in lpm) the resolution improvement (detail on subject) would scale linearly with TC magnification. That's the ideal system. Unfortunately (and I know you understand this--I'm writing also for other readers) the real world has diffraction. In a diffraction limited system, then there is no benefit to adding a TC and then the MTF curve will move left by X for X magnification TC. And since expensive super telephoto lenses are close to diffraction limited, we will see a shift to the left (lower MTF) as we magnify the image with a TC.

If the MTF curve is diffraction limited, then extinction resolution should be about when the first minimum of one diffraction spot is at the maximum of a neighboring spot

Here the continuous curve is the diffracted intensity, the dots are a box blur discrete samples representing the values of ideal pixels spaced by the radius of the diffraction spots; seems to me that this is essentially the extinction point. The radius of the Airy disk for f4 is about 2.7µ or 185 lpm. The fact that you are only getting about half of that seems to me an indication that the pixel size of 6.4µ is having a big impact (I would also think demosaic limits the resolution as well). But perhaps there are other effects not accounted for in this idealized approximation.

It's really the pixel spacing. You would get the same results if the spacing were the same and the pixels were extremely small. But I agree pixel size and pixel spacing are pretty much the same on most cameras. The blur filter is a major effect too, then the lens aberrations and diffraction.

Roger