MTFs for Canon R 100-500mm, 600 f/11 and 800 f/11

[mandog]

I have always been a strong advocate for removal of low-pass filters, and the 5DSR was my favourite body until I got the R5. To my astonishment, the R5 is as sharp as the 5DSR - I compared carefully my EF lenses on the R5 with an adaptor with them on the 5DSR. The 5DSR still has a low-pass filter, but it is supposedly neutralised, but I think that there still some loss of resolution.

I found the RF 2x on the RF 100-500 gives less degradation than the EF 2xTC III on the EF 100-400mm II, and I use it a lot.

I think the 5DsR still has some slight advantage over the R5 in my comparisons, but not enough to give up the other niceties like a flip-out touch screen in most cases. But if I really need maximum IQ from an EF600/4, then I will still use the 5DSR.

I tried an RF 100-500, but decided that taking the TC off to put the lens away all the time was just too much risk of getting dust on the sensor.

 

[AlanF]

I think the 5DsR still has some slight advantage over the R5 in my comparisons, but not enough to give up the other niceties like a flip-out touch screen in most cases. But if I really need maximum IQ from an EF600/4, then I will still use the 5DSR.

I tried an RF 100-500, but decided that taking the TC off to put the lens away all the time was just too much risk of getting dust on the sensor.

I've been taking lenses on and off the R5 several times daily for 18 months, putting on and taking off TCs and haven't got dust on the sensor. The R had the important innovation from Canon of having the shutter close to protect the senor and the sensor cleaning seems very effective. If you keep the open surfaces pointing down during manipulations, like a bacteriologist does with Petri dishes, gravity is your friend.

 

[AlanF]

I think the 5DsR still has some slight advantage over the R5 in my comparisons, but not enough to give up the other niceties like a flip-out touch screen in most cases. But if I really need maximum IQ from an EF600/4, then I will still use the 5DSR.

I tried an RF 100-500, but decided that taking the TC off to put the lens away all the time was just too much risk of getting dust on the sensor.

Canon, I think it was Chuck, said in an interview that the R5 would outresolve the 5DSR, and I thought it was hype. But, optyczne.pl measures sensor resolution and it does seem true. Canon claimed that their innovation with the 1DXIII of splitting the beam into 16 components instead of 6 by the filter would reduce loss of resolution. They appear to be telling the truth, the 45 Mpx R5 sensor being very similar to the Z7's 45 Mpx without the filter.

 

[mandog]

I've been taking lenses on and off the R5 several times daily for 18 months, putting on and taking off TCs and haven't got dust on the sensor. The R had the important innovation from Canon of having the shutter close to protect the senor and the sensor cleaning seems very effective. If you keep the open surfaces pointing down during manipulations, like a bacteriologist does with Petri dishes, gravity is your friend.

IDK, I got dirt on an R5 swapping TCs outdoors, and it wouldn't come off. Fortunately, it was a Canon CPS loaner, so it wasn't my problem in the end....

 

[mandog]

Canon, I think it was Chuck, said in an interview that the R5 would outresolve the 5DSR, and I thought it was hype. But, optyczne.pl measures sensor resolution and it does seem true. Canon claimed that their innovation with the 1DXIII of splitting the beam into 16 components instead of 6 by the filter would reduce loss of resolution. They appear to be telling the truth, the 45 Mpx R5 sensor being very similar to the Z7's 45 Mpx without the filter.

Alan, thanks for pointing me to optyczne.pl--interesting stuff there. I'd be careful comparing the two graphs you posted though, because they used different lenses.

If you look at Bryan Carnathan's cross-test, he used the same lens: Canon EF 200mm f/2L IS USM Lens Image Quality

Another possibility would be to get a 90D instead of the RF800 (a refurb is cheaper now). Then I could shoot with just the 1.4x, which would give 2x shutter speed at ISO100 with about the same MP. But from their test, it looks like the AA filter would eat up most of the extra resolution, even though it only works in one direction.

 

[AlanF]

Alan, thanks for pointing me to optyczne.pl--interesting stuff there. I'd be careful comparing the two graphs you posted though, because they used different lenses.

If you look at Bryan Carnathan's cross-test, he used the same lens: Canon EF 200mm f/2L IS USM Lens Image Quality

Another possibility would be to get a 90D instead of the RF800 (a refurb is cheaper now). Then I could shoot with just the 1.4x, which would give 2x shutter speed at ISO100 with about the same MP. But from their test, it looks like the AA filter would eat up most of the extra resolution, even though it only works in one direction.

The data at f/8 and greater should be independent of lens as most good lenses have negligible aberration there and it's diffraction that is limiting. The R5 is still outresolving the 5DSR at these narrower apertures. I don't put much store by the TDP's charts, they are the weakest point of one of the best websites, if not the very best.

The 90D does not realise the full potential of its 32 Mpx sensor. And, you would have to use low isos. I found at the high isos I shoot at, usually 640 to much more, I wasn't getting better resolution than a 20 Mpx crop without an AA-filter (D500). I just checked on https://www.optyczne.pl/433.4-Test_aparatu-Canon_EOS_90D_Rozdzielczość.html and indeed they find the D7500, another 20 Mpx sensor without the AA-filter, is just as sharp! optyczne.pl is a hidden gem of a website. I read it using chrome browser to translate automatically.