Effects of diffraction and R5/R6 sensor on resolution of f/5.6, f/7.1 and f/11 lenses and TCs

[LogicExtremist]

There aren't much data out there, and I have never handled an RF 600mm f/11. So, here is my indirect opinion. Canon's own MTF values and the-digital-picture show clearly that the RF 100-500mm is significantly sharper and with better contrast at 500mm than the RF 600mm. My 100-400mm II + 1.4x TC at 560mm is pretty close to the Rf 100-500mm in sharpness and contrast. My very good copy of the Sigma 150-600mm C at 600mm had similar resolution but poorer contrast than my 100-400mm II at 560mm, and was definitely inferior to my RF 100-500mm - the RF cropped from 500 was better than the 150-600mm C at 600mm. I would guess that the weaker performance due to diffraction at f/11 of the RF 600mm is not far from the performance of the 150-600mm at f/6.3, which is n ot optimised for optical quality at the longest end. I found the AF and the IS of the Sigma much below those of the Canon lenses on the R5.

Thanks, that's quite helpful. I've been getting more interested in trying out bird photography, and I've put my EF 70-200 f/2.8 III or 55-250 STM on my 80D for more reach, but it's not really long enough.

I've been considering purchasing a longer telephoto but I can't decide whether to go with the RF 600 f/11, Rf 800 f/11 or wait for the RF 100-400 f/8 when that comes out, to use on my RF mount camera, or to go with something like the Sigma 100-400 or 150-600 on the crop body. The images I've seen from the RF f/11 lenses always look dark to me, like they're shot through a ND filter, guessing that's due to the narrow aperture. I've only ever shot that aperture for macro from an inch or two away!

The prices of the Canon EF 100-400 ($3500) or RF 100-500 ($5000) in my part of the world are hard to justify, even though the exchange rate isn't double, Canon likes doubling the prices here, an R5 with an RF 100-500 is a $10,500 setup!

 

[AlanF]

Thanks, that's quite helpful. I've been getting more interested in trying out bird photography, and I've put my EF 70-200 f/2.8 III or 55-250 STM on my 80D for more reach, but it's not really long enough.

I've been considering purchasing a longer telephoto but I can't decide whether to go with the RF 600 f/11, Rf 800 f/11 or wait for the RF 100-400 f/8 when that comes out, to use on my RF mount camera, or to go with something like the Sigma 100-400 or 150-600 on the crop body. The images I've seen from the RF f/11 lenses always look dark to me, like they're shot through a ND filter, guessing that's due to the narrow aperture. I've only ever shot that aperture for macro from an inch or two away!

The prices of the Canon EF 100-400 ($3500) or RF 100-500 ($5000) in my part of the world are hard to justify, even though the exchange rate isn't double, Canon likes doubling the prices here, an R5 with an RF 100-500 is a $10,500 setup!

The brightness of images depends on whether you expose or underexpose or on your post processing settings. The RF f/11 lenses are no exception to this, as are any images from any other lens at f/11. (Strictly speaking T-stops are more accurate for absolute transmittance, but the difference between T and f is corrected by your exposure metering or you.)

 

[LogicExtremist]

The brightness of images depends on whether you expose or underexpose or on your post processing settings. The RF f/11 lenses are no exception to this, as are any images from any other lens at f/11. (Strictly speaking T-stops are more accurate for absolute transmittance, but the difference between T and f is corrected by your exposure metering or you.)

That's what I would have thought, you'd just just adjust the exposure meter to the centre, and brighten the images in post if necessary. I've seen this 'image darkness' for lack of a better expression in most 600 and 800 f/11 image I've seen posted on the internet, and that put me off. I can't imagine they're all shooting jpegs SOOC. I'm at loss to explain it, though you might have some ideas of why this might be the case.

Not sure if I'm allowed to link to the dpreview image gallery:
https://www.dpreview.com/sample-gal...f-800mm-f-11-is-stm-sample-gallery/2421245585

To me many of these images seem to have a grey cast, does it seem that way to you? Thanks

 

[koenkooi]

That's what I would have thought, you'd just just adjust the exposure meter to the centre, and brighten the images in post if necessary. I've seen this 'image darkness' for lack of a better expression in most 600 and 800 f/11 image I've seen posted on the internet, and that put me off. I can't imagine they're all shooting jpegs SOOC. I'm at loss to explain it, though you might have some ideas of why this might be the case.

Not sure if I'm allowed to link to the dpreview image gallery:
https://www.dpreview.com/sample-gal...f-800mm-f-11-is-stm-sample-gallery/2421245585

To me many of these images seem to have a grey cast, does it seem that way to you? Thanks

You can download the RAW files and see if it's due to how DPReview processed them or present in the picture itself, the gallery states "Edited in Adobe Camera Raw 12.2 with Adobe Standard profile and sharpening set to 25. Adjustments limited to exposure, shadows, highlights and white balance."

DPReview complained about the R pictures having a green cast and took a long time to admit that they manually set the camera to green+1, which was recorded in the EXIF. Since then I download the RAWs they provide and check how they look in both LR and DPP4 instead of relying on the JPEGs in their gallery.

Adobe produces horrible colours of R series out of the box, so check the RAW in DPP4 to see what Canon thinks the picture should look like. The RF24-240 RAWs DPReview provided showed how much work DPP4 did for corrections and how well it did.

 

[LogicExtremist]

You can download the RAW files and see if it's due to how DPReview processed them or present in the picture itself, the gallery states "Edited in Adobe Camera Raw 12.2 with Adobe Standard profile and sharpening set to 25. Adjustments limited to exposure, shadows, highlights and white balance."

DPReview complained about the R pictures having a green cast and took a long time to admit that they manually set the camera to green+1, which was recorded in the EXIF. Since them I download the RAWs they provide and check how they look in both LR and DPP4 instead of relying on the JPEGs in their gallery.

Adobe produces horrible colours of R series out of the box, so check the RAW in DPP4 to see what Canon thinks the picture should look like. The RF24-240 RAWs DPReview provided showed how much work DPP4 did for corrections and how well it did.

Thanks, that great advice,, I'll download the raw files and post process them to see if I can get them to look any better. Much appreciated!

 

[AlanF]

That's what I would have thought, you'd just just adjust the exposure meter to the centre, and brighten the images in post if necessary. I've seen this 'image darkness' for lack of a better expression in most 600 and 800 f/11 image I've seen posted on the internet, and that put me off. I can't imagine they're all shooting jpegs SOOC. I'm at loss to explain it, though you might have some ideas of why this might be the case.

Not sure if I'm allowed to link to the dpreview image gallery:
https://www.dpreview.com/sample-gal...f-800mm-f-11-is-stm-sample-gallery/2421245585

To me many of these images seem to have a grey cast, does it seem that way to you? Thanks

I occasionally use the RF 800 f/11 and find no problems with it regarding image exposure. There are others like @usern4cr who have posted nice bright images with it in the birds thread.

 

[HenryL]

That's what I would have thought, you'd just just adjust the exposure meter to the centre, and brighten the images in post if necessary. I've seen this 'image darkness' for lack of a better expression in most 600 and 800 f/11 image I've seen posted on the internet, and that put me off. I can't imagine they're all shooting jpegs SOOC. I'm at loss to explain it, though you might have some ideas of why this might be the case.

Not sure if I'm allowed to link to the dpreview image gallery:
https://www.dpreview.com/sample-gal...f-800mm-f-11-is-stm-sample-gallery/2421245585

To me many of these images seem to have a grey cast, does it seem that way to you? Thanks

Those images are just underexposed. The worst offenders seem to be examples where the intent might have been to protect the highlights (ie the kingfishers, geese and osprey, even Space Needle). Clearly still underexposed, bumping +1 on some of them and still no blown highlights. I'd say it's the photographer not the lens in those examples.

 

[LogicExtremist]

I occasionally use the RF 800 f/11 and find no problems with it regarding image exposure. There are others like @usern4cr who have posted nice bright images with it in the birds thread.

Thanks, I'll check those out too! Great to know you have this lens, I've been considering it.

How would you rate the 800mm f/11 against the EF 100-400 without any extender. Does the 800mm focal length resolve way more than the bare zoom at 400? How does the contrast compare in your opinion?

From your graphs, the theoretical performance at the lens centre (if I'm reading the graphs correctly) is as follows:

• 400mm @ f/5.6 resolves around 225 lp/mm (R5) and 150 lp/mm (R6)
• 600mm @ f/11 resolves around 275 lp/mm (R5) and 225 lp/mm (R6)
• 800mm @ f/11 resolves around 370 lp/mm (R5) and 300 lp/mm (R6)

This shows the 800mm f/11 resolving twice the detail of the bare 100-400 @ 400mm on the R6 (150 vs 300 lp/mm). Does this really reflect real world experience?

 

[AlanF]

Thanks, I'll check those out too! Great to know you have this lens, I've been considering it.

How would you rate the 800mm f/11 against the EF 100-400 without any extender. Does the 800mm focal length resolve way more than the bare zoom at 400? How does the contrast compare in your opinion?

From your graphs, the theoretical performance at the lens centre (if I'm reading the graphs correctly) is as follows:

• 400mm @ f/5.6 resolves around 225 lp/mm (R5) and 150 lp/mm (R6)
• 600mm @ f/11 resolves around 275 lp/mm (R5) and 225 lp/mm (R6)
• 800mm @ f/11 resolves around 370 lp/mm (R5) and 300 lp/mm (R6)

This shows the 800mm f/11 resolving twice the detail of the bare 100-400 @ 400mm on the R6 (150 vs 300 lp/mm). Does this really reflect real world experience?

On the R5, the RF 800mm f/11 resolves about 1.6x the EF 100-400mm II at 400mm (my copies of the lenses). The factor will be higher on the R6. The contrast of both lenses is good. If you want resolution, the 800mm f/11 will significantly outperform the 600mm f/11. They have similar MTF values but the 33% extra length of the 800 will increase its resolution by 1.33x. An 800mm f/11 although having the same f-number as the 600mm f/11, has effectively nearly an extra stop of iso (~0.8) in its favour as it projects 77% more photons per duck.

 

[stevelee]

“photons per duck” should be an international standard.

 

[PhotonShark]

Alan, this is very interesting, thank you very much. Your last finding "the f/11 lenses are more suited to the lower resolution R6, and this makes it an attractive lower priced alternative" didn't surprise me. That's basically what I try to tell the high MP sensor fans since many years.

I know this thread is old, but was thinking the same thing with the release of the R7. It’s surprising that Canon went with such a high resolution sensor. A 20 megapixel low light APS-C mirrorless would have been great with those f/11 lenses. Or, even the new 100-400.

 

[Czardoom]

The new R7 and R10 have we wondering about the sensor and pixel size as well. AlanF, I hope you revisit this thread and just give your educated opinion on how you think these two new crop cameras will perform with the RF 100-400 for example. I am leaning towards getting the R10 because I really don't think I will see any resolution gain from the R7's 32 MP sensor compared to the R10's 24 MP sensor. I know it is not a great sample size, but looking at DPReview's sample photos shot with the RF 100-500, they look quite soft to me. I realize that DPReview's sample photos are usually not very high quality, but that's another story...

Any thoughts on the 32 vs 24 MP sensor and the ability of the Canon telephoto lenses to take advantage of all those MPs before diffraction and the ability to hand hold come into play will be welcome.

 

[AlanF]

@Czardoom and @PhotonShark - Phoblographer has done one of the better pre-reviews of the R7 and has recommended some lenses, which he has actually tested, and recommends the RF 100-500mm https://www.thephoblographer.com/2022/05/24/the-best-lenses-for-the-canon-eos-r7/
My own thoughts are that the 100-500mm will perform very well. It is a really sharp lens, and the older EF 100-400mm did perform well on the 90D. Although the 500mm is f/7.1 vs f/5.6 on the 400mm II, the extra 100mm compensates for the worse diffraction. A good f/4 lens will be better. What is key is to have good RAW processing software like DxO PL5 DeepPrime or its plugin PureRaw for PS to deal with the noise at higher isos. Unfortunately, we will have to wait for the new modules for the RAW converters but Topaz Denoise does a pretty good job on jpegs from DPP.

The RF 100-400mm punches well above its weight in all senses. The Phoblog wrote it didn't focus as fast though, and I will test that when mine arrives. I think that it will give good images on the R7, but as you speculate not much more detail than from the R10 at the isos I work at. The RF 100-400 does give excellent images on the 45 Mpx R5, and the R10 isn't that much more pixel dense, equivalent to about 62 Mpx full frame, and so I think the RF100-400mm on the R10 will make many birders and naturers very happy. I wouldn't use an extender with the 100-400mm.

I have had no trouble hand holding the R5 with both RF zooms at speeds down to 1/200s with little shake - I crop greatly and look for fine detail so I am much more critical about shake than the normal testers. There has always been mirror slap/shutter shock with all my telephotos on all of my DSLRs, and it is simply so much better using ES on the R5.

 

[Czardoom]

Thanks, Alan. I look forward to what you find out when you get the new camera!

 

[PhotonShark]

The RF 100-400mm punches well above its weight in all senses. The Phoblog wrote it didn't focus as fast though, and I will test that when mine arrives. I think that it will give good images on the R7, but as you speculate not much more detail than from the R10 at the isos I work at. The RF 100-400 does give excellent images on the 45 Mpx R5, and the R10 isn't that much more pixel dense, equivalent to about 62 Mpx full frame, and so I think the RF100-400mm on the R10 will make many birders and naturers very happy. I wouldn't use an extender with the 100-400mm.

Thanks for the reply. I look forward to your investigations.

 

[stevelee]

Diffraction is just one of many things you need to take into account in the compromises you have to make. If you stop down enough, everything can be in focus, and wind up equally fuzzy, for example. Most of you guys know better, but it is too easy to see diffraction more like a wall than a gradual slope. Up to some point it is less of an issue than other things. Lots of factors go into determining where that point is.

In some of my messing around, not anything too serious, I have found that the kind of softness that can set it at f/32 with FF or smaller sensors can have a pleasant look. So far I have never used that effect purposely, but maybe some time when I am wanting a more romantic or wistful look, it might be worth trying.

 

[justaCanonuser]

I know this thread is old, but was thinking the same thing with the release of the R7. It’s surprising that Canon went with such a high resolution sensor. A 20 megapixel low light APS-C mirrorless would have been great with those f/11 lenses. Or, even the new 100-400.

I am with you - Bryan Carnathan has already calculated that the diffraction limited aperture (DLA) is f = 5.2 only. That's the price you pay for small pixels. So, if you choose e.g. f/8 to get a nice sharp tele shot, if the light is good, you already left the range of maximum resolution. Well, at least, if you downsize the image to the level of a 20 MP sensor, you'll have no losses but the image quality should beat the old 20 MP 7D. Maybe I'd reside more frequently to MRAW images than with my 7D2, when I upgrade.

 

[AlanF]

I am with you - Bryan Carnathan has already calculated that the diffraction limited aperture (DLA) is f = 5.2 only. That's the price you pay for small pixels. So, if you choose e.g. f/8 to get a nice sharp tele shot, if the light is good, you already left the range of maximum resolution. Well, at least, if you downsize the image to the level of a 20 MP sensor, you'll have no losses but the image quality should beat the old 20 MP 7D. Maybe I'd reside more frequently to MRAW images than with my 7D2, when I upgrade.

For people like me who use mainly telephotos with apertures narrower than f/7.1, the the resolution of the R10 is good enough. But, if you are using f/2.8 or f/4, lenses or wider then diffraction isn’t much of a problem.