Canon RF 100-400mm f/5.6-8 IS USM Review

[EOS 4 Life]

No. All things being equal the 45 Mpx sensor has to give a sharper image. Some thing else has to be happening.

CF shows the image at 100%.
Extra MP is no advantage to sharpness at that point.
It just renders a larger image.

 

[AlanF]

CF shows the image at 100%.
Extra MP is no advantage to sharpness at that point.
It just renders a larger image.

The MTF of a lens is independent of the sensor. If you are doing it visually, the viewing has to be at the same physical size of the image from sensors of the same size for comparison of sharpness of a lens. That's how it's done for most measurements - DR, Signal/Noise etc etc. Otherwise it means you are magnifying all aberrations.

 

[neuroanatomist]

CF shows the image at 100%.
Extra MP is no advantage to sharpness at that point.
It just renders a larger image.

The MTF of a lens is independent of the sensor. If you are doing it visually, the viewing has to be at the same physical size of the image from sensors of the same size for comparison of sharpness of a lens. That's how it's done for most measurements - DR, Signal/Noise etc etc.

Never bet against physics. Ever.

 

[DhlcadR6]

DLO isn’t active on RAW images, so that isn’t a factor.

but DLO should be applied in post-processing (?

I mean I see many who just use the RAW and then proceed to LR and then complains the optical quality is not good.

Could that be because the RF50STM can't "handle" a 45mp sensor?
That's a theory Christopher Frost has proposed with the 28-70 he re-reviewed. He's saying the images are sharper on the R than they are on the R5.

I still thinks CF got a bad copy for the testing on R5....

 

[koenkooi]

but DLO should be applied in post-processing (?

I mean I see many who just use the RAW and then proceed to LR and then complains the optical quality is not good.[…]

Yes, processing the RAW in DPP4 or through DxO will give much better results than the Adobe RAW converter.

 

[shadowsports]

Yes, processing the RAW in DPP4 or through DxO will give much better results than the Adobe RAW converter.

I've used DxO for the last 6 yrs at home, and DPP when I'm on the road. I know people like Adobe CC, but I'm happy not to have to deal with Camera RAW. I'm sure its not so bad, but any time I can avoid an extra step is just fine with me.

 

[EOS 4 Life]

The MTF of a lens is independent of the sensor. If you are doing it visually, the viewing has to be at the same physical size of the image from sensors of the same size for comparison of sharpness of a lens. That's how it's done for most measurements - DR, Signal/Noise etc etc. Otherwise it means you are magnifying all aberrations.

I am not sure that I understand what you are saying.
To test a sensor then it makes sense to downscale the image.
To test a lens then it would make more sense to test at the highest resolution of the sensor.
The same noise would be consistent across different lenses.

 

[koenkooi]

I've used DxO for the last 6 yrs at home, and DPP when I'm on the road. I know people like Adobe CC, but I'm happy not to have to deal with Camera RAW. I'm sure its not so bad, but any time I can avoid an extra step is just fine with me.

I use Lightroom to organize and edit, so DPP/DxO are an extra step for me But for denoising, lens corrections and extracting that last bit of detail, roundtripping to DxO or DPP4 is worth it.

 

[neuroanatomist]

I am not sure that I understand what you are saying.
To test a sensor then it makes sense to downscale the image.
To test a lens then it would make more sense to test at the highest resolution of the sensor.
The same noise would be consistent across different lenses.

The original comment was:

Strangely enough I get much sharper images out of the RF50STM on my R8 than on the R5, even after Canon fixed the IBIS twitch issue when using primes.

and the hypothesis proposed was that the 50/1.8 'can't handle' the higher resolution of the R5. That's bovine scat. The additional resolution of the sensor will result in additional resolution in the images with the same lens. The magnitude of that increase will not necessarily be the same as the difference in the resolution of the sensors, but it will be an increase.

 

[koenkooi]

The original comment was:

and the hypothesis proposed was that the 50/1.8 'can't handle' the higher resolution of the R5. That's bovine scat. The additional resolution of the sensor will result in additional resolution in the images with the same lens. The magnitude of that increase will not necessarily be the same as the difference in the resolution of the sensors, but it will be an increase.

And I could get nicely sharp images out of the R5+50STM, but most of the time there was something 'wrong' with the images. Not enough to discard the image, but still annoying. Before I had the R8, using the M6II+32mm has been a decent work around when a silent shutter isn't needed.

I'm not the steadiest shooter, so I can't rule out operator error, IBIS can make things worse, as was shown by the twitch bug. I'm happy the R8 makes it behave, since it's the better camera for this time of year, the HF anti flicker setting works very well with the LED lighting we have indoors.

 

[EricN]

And I could get nicely sharp images out of the R5+50STM, but most of the time there was something 'wrong' with the images. Not enough to discard the image, but still annoying. Before I had the R8, using the M6II+32mm has been a decent work around when a silent shutter isn't needed.

I'm not the steadiest shooter, so I can't rule out operator error, IBIS can make things worse, as was shown by the twitch bug. I'm happy the R8 makes it behave, since it's the better camera for this time of year, the HF anti flicker setting works very well with the LED lighting we have indoors.

I'm not steady either (lucky enoughto have inherited a familial tremmor). I started using 1/4 speed and focal length ratio on ff(sorry if I wrote the numbers in the opposite order) and it's pretty safe for anything not moving. Don't know if you're willing to slowdown that much, but it's an option.

 

[koenkooi]

I'm not steady either (lucky enoughto have inherited a familial tremmor). I started using 1/4 speed and focal length ratio on ff(sorry if I wrote the numbers in the opposite order) and it's pretty safe for anything not moving. Don't know if you're willing to slowdown that much, but it's an option.

1/200s wasn’t safe enough, which is what made me suspect it wasn’t purely camera shake.

I have to use 1/125s or faster for my kids anyway

 

[mdmphoto]

R5, RF 100-400: 1/320 @ F/11, ISO 2500 LRC Denoise, 40% crop...

 

[AlanF]

R5, RF 100-400: 1/320 @ F/11, ISO 2500 LRC Denoise, 40% crop...

Nice bird. What is it? Why did you stop down to f/11?

 

[mdmphoto]

Nice bird. What is it? Why did you stop down to f/11?

Brazilian, or Red-Crested Cardinal: have a habit of stopping down once, for sharpness; though with this lens the difference between F/8 and F/11 @ 400mm is negligible...

 

[AlanF]

Brazilian, or Red-Crested Cardinal: have a habit of stopping down once, for sharpness; though with this lens the difference between F/8 and F/11 @ 400mm is negligible...

Thanks for the reply - I loved the bird. I have tested all my telephoto lenses fro IQ vs f-number and unless I need more dof I have never stopped down. You do see reports on some sites that the RF 100-400mm is significantly sharper at f/11. But, my pair are not and neither are those tested by opticallimits, which is slightly sharper at f/8, and ephotozine, which is slightly softer.