Sigma 56 F1.4 on M6II versus RF 85 F2 on R

In another thread there has been some discussion recently of how the Sigma 56 on the M6II compares with the Canon RF 85 F2 on the R (or other R camera). This might be of interest to those deciding whether to go the M or R route, because both lenses are good, sharp, moderately priced (not dirt cheap, but not silly expensive either) primes that are well suited to traditional portrait shooting. Luckily for me, I don't have to decide which route to go, because I have both systems (an R and an M6II and two other M bodies), with several primes of roughly equivalent FOV for each system (22, 32, and 56 on M, and 35, 50, and 85 on R). A bit over a year ago, I did a comparison of several primes on the M6II and R, shooting the same subject from the same distance in the same lighting (within a few minutes), all at base ISO. This was the thread in which I compared the 22 and 56 on M6II with the 35 and 85 on R:

https://www.dpreview.com/forums/post/64718104

And this is the one in which I compared the 32 on M6II with the 50 F1.8 on R:

https://www.dpreview.com/forums/post/64746137

I thought it would be interesting to go back and reprocess some of the shots with DXO PL5, to see whether that makes a difference. I only did it for the 56 v. 85 comparison, because I have limited time, and that was the specific comparison that came up in the other thread. I processed in both PL5 and LR, both a web-sized version (2000 pixels on long side) and a 100% crop of the subject (Paddington Bear). I only did wide open shots (F1.4 on the 56, F2 on the 85), and did both a corner/edge shot and a center shot for each. Here are the results, first with Paddington in the center:

Sigma 56 F1.4,Lightroom

Sigma 56 F1.4, DXO PL5

And 100% crops:

Sigma 56 F1.4, Lightroom

Sigma 56, DXO PL5

Canon RF 85 F2, Lightroom

Canon RF 85 F2, DXO PL5

and 100% crops:

Canon RF 85 F2, Lightroom

Canon RF 85 F2, DXO PL5

And now the shots with Paddington in the upper left:

Sigma 56 F1.4, Lightroom

Sigma 56 F1.4, DXO PL5

And 100% crops:

Sigma 56 F1.4, Lightroom

Sigma 56 F1.4, DXO PL5

Canon RF F2, Lightroom

Canon RF 85 F2, DXO PL5

And 100% crops:

Canon RF 85 F2, Lightroom

Canon RF 85 F2, DXO PL5

After peering at the results for several minutes, here are my preliminary observations: (i) both the Sigma 56 on the M6II and the Canon 85 on the R give excellent results. I'm sure different people will find different differences between them, but let's be honest, if you can't take a good shot with either of these combinations, it's on you, not the gear. Does one combination have better sharpness and/or bokeh than the other? Darned if I know. My choice between them is based on which combination I feel like using on the day, and has more to do with size, weight, ergonomics, etc. than with any possible differences in the output. (ii) the difference between Lightroom and DXO PL5 is much less apparent at base ISO than at higher ISOs, where DXO clearly does a better job at controlling noise and preserving details. At least that's seems true for these two lenses. Perhaps the difference in lens corrections would be more apparent for different lenses. I think the DXO versions here have a bit more detail, but nothing like the striking difference you see at high ISO.

Edit: After going back and looking closely, I can definitely see more detail in the DXO versions. But the bokeh isn't as nice in them. That's probably because the settings are different. It's really hard to get similar settings between Lightroom and PL5, because some of the adjustments don't exactly correspond to each other.

The thing that stood out to me most was the difference in bokeh. I don't like what PL5 has done with that. Did you dial up the microcontrast?

Process them the way God intended.  In DPP4. 😁

I would be interested to know if Canon has lens profiles for Sigma lenses.

I have the 85.  For portraits I prefer it because my RP picks up the eye so much further away than my wife’s  M6II. So we happily keep the EF-M 32 for a nice general candid focal length (the RF 50 f1.8 won’t top that), and the 85 on my RP.

dan the man p wrote:

The thing that stood out to me most was the difference in bokeh. I don't like what PL5 has done with that.

I agree with you.  Nonetheless things get very interesting here.

I compared the Sigma + Lightroom in the corner to the RF + Lightroom in the corner.

Conclusion: it's hard to compare with fields of view not being the same and - as I assume - the camera at the same distance. The contrast of the RF 85mm is better.

The difference in processing show it's possible Canon's in camera jpg engine is doing tricks with budget RF lenses to make those appear to be sharper like the DXO PL5 does, giving better contrast at the cost of bokeh quality. At the same time the RF lens has better contrast with the same processing in Lightroom, so comparing to the sigma there's an optical difference too for contrast.

dan the man p wrote:

The thing that stood out to me most was the difference in bokeh. I don't like what PL5 has done with that. Did you dial up the microcontrast?

Good points. I had the microcontrast set to 15, which is my default, and what I thought roughly matches the clarity setting of 25 on LR that I use as a default. But it's entirely possible that the other lens corrections on PL5 mean I just don't need as high a microcontrast setting. I went back and redid one of them (the 56 center shot), lowering the microcontrast to 0, the midtones contrast to -16, and the ClearView Plus setting to 10 (from 20, which is my default, and what I thought roughly corresponds to the dehaze setting of 25 that I use on LR).

So here is the 56 F1.4 center shot again, first LR then DXO:

Now here's the reprocessed DXO version:

And here are the 100% crops from the two DXO versions, to see whether the adjustments to help the bokeh had an adverse effect on detail:

This just goes to show how difficult it is to match settings between processors. It's entirely possible, of course, that the extra detail you get with the DXO lens corrections comes at a price of less pleasant-looking bokeh. But it does seem like there are some things you can do to compensate for that.

thunder storm wrote:

I compared the Sigma + Lightroom in the corner to the RF + Lightroom in the corner.

Conclusion: it's hard to compare with fields of view not being the same and - as I assume - the camera at the same distance. The contrast of the RF 85mm is better.

Well, to be fair, the fields of view are pretty darn close. The magnification of the 56, giving a 90 FF equivalent FOV, is only slightly more than the 85. The framing of one shot starts ever so slightly more to the left than the other. The subjects in each case are pretty close in size. From the top of his head to the bottom of his feet, Paddington is 1320 pixels on the Sigma shot and 1260 pixels on the Canon RF shot.

The difference in processing show it's possible Canon's in camera jpg engine is doing tricks with budget RF lenses to make those appear to be sharper like the DXO PL5 does, giving better contrast at the cost of bokeh quality. At the same time the RF lens has better contrast with the same processing in Lightroom, so comparing to the sigma there's an optical difference too for contrast.

You might be correct, though all that matters for me is the end result, not how we get there. But still, the differences are so small that you'd have to be an obsessive pixel peeper (like most people here are!) to base a choice of system on them, rather than differences in usability, feel, portability, etc, which are far more important to a lot of people (me, for instance).

Wow !

I put them into Photoshop and pixel peeped and am surprised how well the M6II with the 56MM does.

They are so close in sharpness.

I like the background in the 56MM LR photo better.

M6II can deliver the colors too.

Yeah, it looks like the microcontrast setting was to blame. One nice thing about PL5 is that you can apply many of the adjustments, including microcontrast, locally rather than globally. That's what I would do in a situation like this if I wanted the extra detail on the subject while retaining smooth bokeh. Sometimes I even reduce microcontrast to < 0 locally if I want a softer look.

m100 wrote:

Wow !

I put them into Photoshop and pixel peeped and am surprised how well the M6II with the 56MM does.

They are so close in sharpness.

I like the background in the 56MM LR photo better.

M6II can deliver the colors too.

The nose looks better in the DXO processed photo though ?

My wife says that is a cute bear. 

m100 wrote:

My wife says that is a cute bear.

It's better to use those teddy bears for our pixel peeping rather than beautiful young ladies.

Alastair is such a wise forum member.

Alastair Norcross wrote:

thunder storm wrote:

I compared the Sigma + Lightroom in the corner to the RF + Lightroom in the corner.

Conclusion: it's hard to compare with fields of view not being the same and - as I assume - the camera at the same distance. The contrast of the RF 85mm is better.

Well, to be fair, the fields of view are pretty darn close. The magnification of the 56, giving a 90 FF equivalent FOV, is only slightly more than the 85. The framing of one shot starts ever so slightly more to the left than the other. The subjects in each case are pretty close in size. From the top of his head to the bottom of his feet, Paddington is 1320 pixels on the Sigma shot and 1260 pixels on the Canon RF shot.

I'm comparing these.

https://2.img-dpreview.com/files/p/E~forums/65954532/ec257b13b110466ca1fd81223962516c

https://3.img-dpreview.com/files/p/E~forums/65954532/ef60b5c9d37149e8943bc3466bca336e

For me the size difference makes it difficult to compare in my browser.  I've downloaded the files to view in another program being a bit more flexible for resizing, and I have to admit the Sigma seams to be a tiny bit sharper in your Lightroom processed samples, like nowak mentioned in the other topic.

This contradicts with my own experience with the lenses, so maybe there is some copy to copy variation playing a role.

The difference in processing show it's possible Canon's in camera jpg engine is doing tricks with budget RF lenses to make those appear to be sharper like the DXO PL5 does, giving better contrast at the cost of bokeh quality. At the same time the RF lens has better contrast with the same processing in Lightroom, so comparing to the sigma there's an optical difference too for contrast.

You might be correct, though all that matters for me is the end result, not how we get there. But still, the differences are so small that you'd have to be an obsessive pixel peeper (like most people here are!) to base a choice of system on them, rather than differences in usability, feel, portability, etc, which are far more important to a lot of people (me, for instance).

For me both lenses aren't satisfying for AF, although in different ways. The 56mm has the size & weight advantage. The 85mm has the half way macro feature, the IS, and the ability to handle back lit situations very well.  I don't need half way macro so much, but I do  love the last two features.  If only the AF would be a little bit better. I'm hoping for the firmware update of the end of march for the R5 to improve the AF of this RF 85mm f/2.0 IS stm.  I don't need it to be faster so much. I would like it to be faster, but I can live with the speed as it is now. Just a little less jumping in low light, and killing that short "jumping for no reason" even in good light, that would be good enough for me.

thunder storm wrote:

Alastair Norcross wrote:

thunder storm wrote:

I compared the Sigma + Lightroom in the corner to the RF + Lightroom in the corner.

Conclusion: it's hard to compare with fields of view not being the same and - as I assume - the camera at the same distance. The contrast of the RF 85mm is better.

Well, to be fair, the fields of view are pretty darn close. The magnification of the 56, giving a 90 FF equivalent FOV, is only slightly more than the 85. The framing of one shot starts ever so slightly more to the left than the other. The subjects in each case are pretty close in size. From the top of his head to the bottom of his feet, Paddington is 1320 pixels on the Sigma shot and 1260 pixels on the Canon RF shot.

I'm comparing these.

https://2.img-dpreview.com/files/p/E~forums/65954532/ec257b13b110466ca1fd81223962516c

https://3.img-dpreview.com/files/p/E~forums/65954532/ef60b5c9d37149e8943bc3466bca336e

For me the size difference makes it difficult to compare in my browser.

Interesting. When I view them in Firefox on my 24" iMac, the difference is pretty small. I just used a ruler on my screen to measure them. In the slightly larger, the distance from the top of Paddington's head to the bottom of his feet is 7.55 inches, and in the smaller one, it's 7.25 inches. That's only a difference of 4.1%. Are you seeing a bigger difference than that in your browser?

I've downloaded the files to view in another program being a bit more flexible for resizing, and I have to admit the Sigma seams to be a tiny bit sharper in your Lightroom processed samples, like nowak mentioned in the other topic.

This contradicts with my own experience with the lenses, so maybe there is some copy to copy variation playing a role.

Or maybe it's just slightly different processing settings.

The difference in processing show it's possible Canon's in camera jpg engine is doing tricks with budget RF lenses to make those appear to be sharper like the DXO PL5 does, giving better contrast at the cost of bokeh quality. At the same time the RF lens has better contrast with the same processing in Lightroom, so comparing to the sigma there's an optical difference too for contrast.

You might be correct, though all that matters for me is the end result, not how we get there. But still, the differences are so small that you'd have to be an obsessive pixel peeper (like most people here are!) to base a choice of system on them, rather than differences in usability, feel, portability, etc, which are far more important to a lot of people (me, for instance).

For me both lenses aren't satisfying for AF, although in different ways. The 56mm has the size & weight advantage. The 85mm has the half way macro feature, the IS, and the ability to handle back lit situations very well. I don't need half way macro so much, but I do love the last two features. If only the AF would be a little bit better. I'm hoping for the firmware update of the end of march for the R5 to improve the AF of this RF 85mm f/2.0 IS stm. I don't need it to be faster so much. I would like it to be faster, but I can live with the speed as it is now. Just a little less jumping in low light, and killing that short "jumping for no reason" even in good light, that would be good enough for me.

I've never experienced anything like that with the RF 85 on my R. Do you think it's peculiar to the R5?

thunder storm wrote:

m100 wrote:

My wife says that is a cute bear.

It's better to use those teddy bears for our pixel peeping rather than beautiful young ladies.

Alastair is such a wise forum member.

I'll not argue with that.

dan the man p wrote:

Yeah, it looks like the microcontrast setting was to blame. One nice thing about PL5 is that you can apply many of the adjustments, including microcontrast, locally rather than globally. That's what I would do in a situation like this if I wanted the extra detail on the subject while retaining smooth bokeh. Sometimes I even reduce microcontrast to < 0 locally if I want a softer look.

Thanks for the reminder about local adjustments. I have only been using PL5 since last Thanksgiving (bought it in the Black Friday sale), and I'm not yet used to using control points a lot. I have used them a bit, but I should probably get in the habit of using them more, because, as you say, it's pretty easy to apply local corrections that way. You can do it in LR too, but PL5's control points makes it easier.

Alastair Norcross wrote:

I've never experienced anything like that with the RF 85 on my R. Do you think it's peculiar to the R5?

It's mainly with AF face/eye tracking at speeds and/or in situations needing "stickyness" the R can't handle anyway. Now that might seem challenging, but my Sigma lenses don't have these kind of problems on the R5, while these are third party lenses in stead of OEM brand native RF. It seems somehow AF speed is prioritized over stickyness to the subject. Now there isn't a whole lot of speed to trade here, however, I do prefer stickyness to be main priority, and speed the secondary. Or, even better, gimme a setting to tweak these things.

The low light jumping though was the same on the R.  With low light the AF can jump even with moderate movements.

When using the IS at full extent there's not a whole lot of light as the IS is very good, and then it even doesn't take any movement to get hunting AF.  Still, my EF 24-70mm f/2.8 mkII doesn't hunt in the exact same situations (on a tripod), while it is actually a stop slower and isn't a native RF lens.

thunder storm wrote:

Alastair Norcross wrote:

I've never experienced anything like that with the RF 85 on my R. Do you think it's peculiar to the R5?

It's mainly with AF face/eye tracking at speeds and/or in situations needing "stickyness" the R can't handle anyway. Now that might seem challenging, but my Sigma lenses don't have these kind of problems on the R5, while these are third party lenses in stead of OEM brand native RF. It seems somehow AF speed is prioritized over stickyness to the subject. Now there isn't a whole lot of speed to trade here, however, I do prefer stickyness to be main priority, and speed the secondary. Or, even better, gimme a setting to tweak these things.

Can't you tweak them in camera?

The low light jumping though was the same on the R. With low light the AF can jump even with moderate movements.

When using the IS at full extent there's not a whole lot of light as the IS is very good, and then it even doesn't take any movement to get hunting AF. Still, my EF 24-70mm f/2.8 mkII doesn't hunt in the exact same situations (on a tripod), while it is actually a stop slower and isn't a native RF lens.

It's possible I haven't used it in those situations. Probably the lowest light I used it in for moving subjects was stage lighting, taking rehearsal shots. It had no problem sticking to the eyes of moving actors, in lighting that varied quite a bit (it was a rehearsal, so the lights were all over the place). Otherwise, I've used it for fast runners in good light, with no problems.

Alastair Norcross wrote:

dan the man p wrote:

The thing that stood out to me most was the difference in bokeh. I don't like what PL5 has done with that. Did you dial up the microcontrast?

Good points. I had the microcontrast set to 15, which is my default, and what I thought roughly matches the clarity setting of 25 on LR that I use as a default. But it's entirely possible that the other lens corrections on PL5 mean I just don't need as high a microcontrast setting. I went back and redid one of them (the 56 center shot), lowering the microcontrast to 0, the midtones contrast to -16, and the ClearView Plus setting to 10 (from 20, which is my default, and what I thought roughly corresponds to the dehaze setting of 25 that I use on LR).

So here is the 56 F1.4 center shot again, first LR then DXO:

Now here's the reprocessed DXO version:

And here are the 100% crops from the two DXO versions, to see whether the adjustments to help the bokeh had an adverse effect on detail:

This just goes to show how difficult it is to match settings between processors. It's entirely possible, of course, that the extra detail you get with the DXO lens corrections comes at a price of less pleasant-looking bokeh. But it does seem like there are some things you can do to compensate for that.

Thanks for the samples, Alastair!

I'd agree with your conclusions.

I got the RF 85 F2 IS for the IS so I can shoot handheld at slower shutter speeds

This redo also shows how process settings can make a large difference in how bokeh is rendered. Those first dxo samples looked like bad bokeh but you've been able to re-do the dxo setting to make the bokeh pleasant again like the lightroom rendering.

first time in 20 years I am able to see this impact of processing so thank you again for the work!

I appreciate your comparison as I was wondering how DXO PL5 works. I couldn't think that DXO's sharpening may spoil bokeh until seeing your comparison. In terms of sharpness, Sigma 56 with M6mk2 is good enough for my taste. When it comes to bokeh, I prefer Sigma to RF 85 according to your test shots. So I do not need DXO nor R with RF 85 in these cases. To be honest, the shot with 56 with LR looks the best to my eyes. One thing still remains. It is a comparison about noise performance under high ISO.

Thank you a lot for your time and effort.