All new 24mp sensor coming to the next Canon ILC’s [CR1]

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,355
22,528
My favourite article on MTF is by Norman Koren, and I quote from it.

“Perceived image sharpness (as distinguished from traditional lp/mm resolution) is closely related to the spatial frequency where MTF is 50% (0.5)— where contrast has dropped by half.”

“MTF is the spatial frequency response of an imaging system or a component; it is the contrast at a given spatial frequency relative to low frequencies.

Spatial frequency is typically measured in cycles or line pairs per millimeter (lp/mm), which is analogous to cycles per second (Hertz) in audio systems. Lp/mm is most appropriate for film cameras, where formats are relatively fixed (i.e., 35mm full frame = 24x36mm), but cycles/pixel (c/p) or line widths per picture height (LW/PH) may be more appropriate for digital cameras, which have a wide variety of sensor sizes.

High spatial frequencies correspond to fine image detail. The more extended the response, the finer the detail— the sharper the image.”

Those MTF values by Lensrentals and Optyczne show by proper measurement that the 5DSR has significantly higher MTF values than the 5DS and hence higher perceived sharpness. Eyeballing photos of playing cards that lack fine detail that needs to be resolved gives little information on resolution and is certainly no substitute for quantitative measurements.
 
  • Like
Reactions: 1 user
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
High spatial frequencies correspond to fine image detail. The more extended the response, the finer the detail— the sharper the image.”

This last part is false, and that's surprising considering the source. Image sharpness and resolved detail are not only different things, lens designers sometimes find they can trade off one against the other. (Not sure how often that comes up with CAD and modern manufacturing, but it came up in the past.)

Edit: Norman's other writings are consistent with what I'm saying. I think the problem here is a poorly worded 2nd sentence.

Those MTF values by Lensrentals and Optyczne show by proper measurement that the 5DSR has significantly higher MTF values than the 5DS and hence higher perceived sharpness.

Such confidence considering you can't see the actual plot, just one number at one contrast point.

Eyeballing photos of playing cards that lack fine detail that needs to be resolved gives little information on resolution and is certainly no substitute for quantitative measurements.

You don't have a quantitative measurement of the extinction resolution for either camera. I would challenge you to find it because I looked for it and the problem is that the couple sites that report extinction resolution all found that both cameras out resolved their targets.
 
Last edited:
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
Reminder: for me on a 4k monitor (at least) images display larger than 100% on this forum. I find this odd but if you download them they should "pixel peep" normally.

On a 4k monitor 100% should be roughly equivalent to a 40" print. On a traditional FHD monitor it will be roughly equivalent to a 90" print. Give or take based on your specific monitor's dpi (mine is 216). I'll refer to print sizes since 100%/200% is relative to your monitor. For me those values = 40" / 80" print sizes.

With that out of the way...

I'm going to post a series of comparisons similar to (some) of what I did when I was deciding between the two. All crops are from Imaging Resource test RAWs. In all cases I used ACR, Adobe Standard v2, all sharpening and LNR off, and all lens corrections off. Order (left to right or top to bottom) is 5Ds / 5Dsr / 5Ds sharpened (if present).

First sample makes one wonder if there's ANY reason to choose one over the other. Moire shows in the verticals and sharpness is indistinguishable at 40". The verticals are the smallest bit less "rough" on the 5Ds and the moire on the bottom left vertical is ever so slightly less.

If I had gone off this comparison alone I would have bought the 5Dsr simply because I think its badge colors look nicer.

183954
 
Last edited:
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
Here's a section including the horizontal test lines. At 40" there are slight sharpness differences with the far right sharpened 5Ds sample being the sharpest, indicating a difference of less than 20%/1px between the cameras. Horizontal test lines are fully resolved on both. (As I said, the 5Ds bodies out resolve the target.)

183956
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
Now the fabric section from their studio scene. In this case I sharpened 35%/1px based on the red sections. I didn't feel like 20%/1px cut it. (Expected with lower contrast detail and the color red.) A bigger difference in favor of the 5Dsr! But still the same detail resolution, and still no difference that would survive the full image chain to print at any size.


183957
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
And last one, vertical lines from the resolution chart. In this case I bumped the sharpening to 40%/1px. At 40" I can see the small difference between them. The 5Dsr (middle) is sharper but is also aliasing, and the sharpened 5Ds (right) is the best. If we're splitting hairs one could point out that near extinction resolution aliasing will give you more problems in post than a bit less detail contrast (sharpening).

183960
 
Last edited:
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
Now that I've posted those comparisons...

Before buying I obsessed more than I should have and I reviewed many more samples than just Imaging Resource and DPReview, and that included some landscape (distant foliage) samples. I did discount two examples where I thought user error was at play. (Not gross user error, but unaccounted for misfocus or motion blur.)

I consistently found the 5Dsr was a bit sharper viewed at 80" but also exhibited a bit more aliasing. I never found a difference in resolved detail, only a difference in detail contrast. In no case did I think the sharpness/aliasing differences would actually matter even in unprocessed prints. More important to me: I found no difference which would survive post processing. I was primarily interested in whether or not the sharpness difference would force me to over sharpen to compensate, thereby introducing processing artifacts. (Something I sometimes struggled with on images from Canon crop sensors.) Not even close. The sharpness differences are far below the sharpening I would normally perform on any image to achieve what I feel is optimal for a given print size.

That's why I went with a slightly lower chance of full on moire, because when moire breaks out it's visible in an 8x10 and on the web and it has to be cleaned up.

These two bodies (and the D800 vs D800E) are basically interchangeable. In real life the advice I gave to another photographer was: if you're OCD about being able to claim the mostest sharpest ever! buy the 5Dsr. If you're OCD about moire buy the 5Ds. If you're not OCD choose gold or silver+red for your badge color.

And no, I don't care about one number from a test that's supposed to produce a graph. The full graph can give you some idea about real world performance, but the one number is nearly useless. At the end of the day I care about real world prints of real world subjects.
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,355
22,528
I suggest we let the discussion end at this point with a mutual agreement to disagree. It's not like it's for or against the phenomenon of global warning that affects the whole planet and its individual inhabitants but instead merely that you go on happily with the view that AA-filters do not slightly blur images and I take the opposite view that they do slightly blur as they are designed to lower MTFs.
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
I suggest we let the discussion end at this point with a mutual agreement to disagree...but instead merely that you go on happily with the view that AA-filters do not slightly blur images and I take the opposite view that they do slightly blur

You could have said 'let's agree to disagree' without intentionally misrepresenting what I've said. Sad...
 
Upvote 0