Help with r7

[Greyswan]

I bought a new r7 this last February, thinking it would be a great complement to my 30 mpx eos R. I wanted the 32mpx crop factor and the highly-touted AF with eye tracking.

I am SO disappointed! I have tried all the different settings, kept my ISO below 1000 and shutter speed up for BIF. The images I get even at loww ISOs like 400, are muddy, not sharp, and just, IMO, horrible. The images almost look like a watercolour painting with the paper grain showing though, even with correct exposure and no editing in the Raws. Well below my standards for images. I have shot the same subjects on the same day with my Eos R, and the images are sharp and clean. I switch to the r7 and it turns into mud! Blacks are dull, without detail, and just blocky. Ugly.

I am wondering, at this point, if I got a bad sensor - I have heard that there was a bit of an issue with QC on the r7 and some r7's are just not good - I've heard of people sending it in to Canon with Canon claiming there's nothing wrong with the camera, and getting it returned with much better images - apparently Canon does not want to admit they have some bad cameras selling out there. Hearsay of course, but how to know?

Is the 32mpx sensor in a crop adding to this blocky, muddy effect? Feather detail seems to be lacking as well. I should also add that I'm getting very few keepers with the r7, well below what I get with the R, and yes, I'm taking into account the that the R is full-frame with larger pixels, therefore more smoothness and easier to focus. The R7 seems to miss focus often even when the confirmation box is green.

So, I decided to compare the same subject shot under the same conditions, and similar distance from my old 20 mpx 5dmkii in 2020, and the r7 file. The difference is astounding! The image attached shows the sandhills, SOOC from the 5dii, superimposed on the r7 file. Neither file has had any Raw editing, both zeroed out. Sizes have not been changed from the original Raws, the subjects almost match in actual size in the images. The 5dii file is SO much cleaner, colour differences aside. The r7 file is slightly underexposed, and it seems to do that all the time as well. Both images were shot in almost idendical weather conditions, a bright sunny day.

Is this normal? I mean, the 5dmkii is 20mpx while the r7 is 32mpx. the r7 should be miles better IMO. In the full size image the neck detail is much cleaner from the 5D.
I'd appreciate opinions on this. Is this normal for the r7? I'm very frustrated at this point.

Cheers. All opinions welcome!

Here's a detail comparison, and the full image below. Look at the difference in the eye on the 5Dii, even with the quality loss from uploading. Wing edges are sharp and clean.

 

[paddler4]

Same lens?

It looks out of focus to me, but it's very hard to tell at this resolution. I suggest taking test shots with something detailed and static. A brick wall works fine. If that's sharp, the problem was with focusing. If that isn't sharp either, it suggests a problem with the body or lens.

The additional megapixels may not be relevant. Unless you are cropping a lot or displaying at more than 20 MPX, they won't give you more detail. 20 should be ample, if you aren't cropping much. For example, here's a shot taken with my OM-1 II, which is only 20 MPX.

 

[Archibald]

Hi, Chris.

I would not test a camera on birds in flight. As Paddler says, test it on a fixed, detailed object. I would pick something like a traffic sign or the facade of a building, in good light. Give the camera a bit of time to focus.

That being said, a lot of people have complained about the AF of the R7. I too have had issues, with the camera seemingly finding the eye of a subject, yet not focusing on it. It seems to perform worse for distant subjects.

Although the AF can be finicky, the R7 takes great photos when it succeeds. I have taken more good photos with the R7 than any camera body.

 

[Bobtographer]

So the image is 100% crop from both cameras?

 

[Greyswan]

Same lens on both cameras, yes. 100-400 mkii , with and without the 1.4 tc. No filters. That lens is tack sharp on the R, on my 6d, never a problem. My image of the crane is about as good as the focus gets, on all subjects, even close up. Your heron looks perfect, the detail on the feathers are there.

I will try some static tests. I would add, that even the images that are properly focused, still look muddy to me.

I have tried it with subject tracking off & on, eye detect off and on, different cases for stickiness, every suggestion I could find on the net, lol. My big beef is just with the image quality.

Anyway, I'll try the static tests and see what happens.

 

[paddler4]

PS: the reason I suggest a brick wall is that it is full of sharp lines that extend edge to edge, with textured surfaces in between. (Rough brick is the best.) Short of some detailed focus charts, it's hard to get that level of detail in test shots. Then start by checking the lines between the mortar and brick, starting at the center and then checking edges and corners.

 

[Greyswan]

So the image is 100% crop from both cameras?

Yes, 100% crop from both. The birds are almost exactly the same size in the Raw files, so not a distance differential or upsized to match, nothing like that. The edges on the bird from the r7 look fuzzy as you see on the beak. All the edges look similar on all the r7 images, not as defined as the 5d. The images look like watercolours.

 

[Greyswan]

PS: the reason I suggest a brick wall is that it is full of sharp lines that extend edge to edge, with textured surfaces in between. (Rough brick is the best.) Short of some detailed focus charts, it's hard to get that level of detail in test shots. Then start by checking the lines between the mortar and brick, starting at the center and then checking edges and corners.

Indeed, I should have done that right away, but I wanted to get out and shoot the spring birds, lol. I will do those tests though, and see what happens.

 

[paddler4]

Sorry, I missed that these were 100% crops. That raises other issues. I think the more meaningful comparison would be cropping them to the same image size, not cropping them to 100%.

 

[paddler4]

I'm not an expert in optics, but I have a little more time now and can try to explain a bit what I meant in my last post.

All other things being equal, a 100% crop from an R7 will look softer than those from many other cameras because of the high pixel density and hence small photosites. 32.5 MPX is a lot for an APS-C sensor, leading to a pixel pitch of only 3.2 microns. Even for a superb lens like the 100-400, that's pushing it, and with a TC, the resolution of the lens drops.

The result of this is that if all other things are equal, 100% crops will often look softer than those from lower-density sensors. I haven't done A/B tests to be more precise, but I notice this in comparing my two cameras, which have almost the same count of pixels. One is a full frame R6 II, while the other is a micro-four-thirds OM-1 Mark II. When I first got the OM-1, I thought something was wrong with it because when blown up to 100%, the images appear softer than those from the R6 Mark II. However, I think what I am seeing is just the very large difference in pixel pitch: about 6 microns for the R6 II and about 3.3 microns for the OM-1.

An additional factor that probably doesn't matter too much in your examples is that the smaller the pixels, the wider the aperture at which you start to get softening from diffraction. I read somewhere that the diffraction limit for the R7 is about f/5.2, which is not all that far from the limit for my OM-1. I doubt this had much impact in your comparison of images at normal size because the effects of diffraction at normal magnifications is very small until you get quite a number of stops narrower than the limit, but it may be contributing a bit to what you see at 100%.

The bottom line: comparisons are more meaningful at the same image size, not at the same percentage magnification.

 

[Greyswan]

I'm not an expert in optics, but I have a little more time now and can try to explain a bit what I meant in my last post.

All other things being equal, a 100% crop from an R7 will look softer than those from many other cameras because of the high pixel density and hence small photosites. 32.5 MPX is a lot for an APS-C sensor, leading to a pixel pitch of only 3.2 microns. Even for a superb lens like the 100-400, that's pushing it, and with a TC, the resolution of the lens drops.

The result of this is that if all other things are equal, 100% crops will often look softer than those from lower-density sensors. I haven't done A/B tests to be more precise, but I notice this in comparing my two cameras, which have almost the same count of pixels. One is a full frame R6 II, while the other is a micro-four-thirds OM-1 Mark II. When I first got the OM-1, I thought something was wrong with it because when blown up to 100%, the images appear softer than those from the R6 Mark II. However, I think what I am seeing is just the very large difference in pixel pitch: about 6 microns for the R6 II and about 3.3 microns for the OM-1.

An additional factor that probably doesn't matter too much in your examples is that the smaller the pixels, the wider the aperture at which you start to get softening from diffraction. I read somewhere that the diffraction limit for the R7 is about f/5.2, which is not all that far from the limit for my OM-1. I doubt this had much impact in your comparison of images at normal size because the effects of diffraction at normal magnifications is very small until you get quite a number of stops narrower than the limit, but it may be contributing a bit to what you see at 100%.

The bottom line: comparisons are more meaningful at the same image size, not at the same percentage magnification.

I understand what you're saying, but what I am seeing is quite different than just softness, which I could deal with.

I also have the OMD EM1 system and also have had many other cameras, so I feel I have the experience to see whether the problem is on my end as user error or if it's the camera. I used to make my living as a product photographer, so consider myself reasonably knowledgeable about image quality. I don't pixel peep, and am not rabid about perfect images either.

I'm seeing images from the r7 that are exposed properly, that do not have motion blur, and have a high enough shutter that it should easily pick up detail. I shot a hawk overhead yesterday that almost filled the frame, would have been sharp as a tack on my R, yet the r7 shows no eye detail at all, just a grainy image. I have to reiterate the image is grainy/gritty, not noisy, there's no colour noise. The ISO was 800, which should be low enough even for a crop camera, when the subject almost fills half the frame. I've seen people post images on the r7 shot at 6400 and much more, with good results on editing. The images I get are not editable, you can't draw the detail out.

I am taking into account the r7 is a crop camera and not as good with noise as a full frame, and smaller pixel depth. Never had this problem with m43s.

Here's another example of what I mean: the merganser image is the full size RAW, and the detail shows how crushed the blacks are.
I don't know if it will show with the upload, but if you look at the white parts of the wing on the male on the screenshot, you'll see that grainy, gritty look that I mentioned. Just mushy without decent detail. These are completely unedited. 800 ISO - 1/4000 ss - 560mm. I do feel the birds are large enough in the frame that I should be getting decent detail. I don't think I'd be getting diffraction at f8? I have also shot at f9 with the same result.

I have a feeling the AF is either off, and is misfocusing just enough to crush the detail, or maybe the sensor is defective. I have tried all the different cases for stickiness, taken without the 1.4 tc, all the settings for birds recommended by different experts. I'm not a world-class photographer by any means, NatGeo isn't blowing up my phone , but I think I should be able to do better than these.

I've shot a few thousand images thus far on the r7 and my keeper rate is absolutely abysmal. Maybe eight or nine images I am *reasonably* happy with. All of birds ten feet away and sitting still which is why I'm wondering if the AF is misfocusing at further distances. But even those images were only a few good ones out of bursts.

I've included a second sample image of an osprey, shot under the same conditions, smaller in the frame and shot at 1600 ISO on my R, it has clear detail despite being much smaller in the frame - even the darker bird has lots of detail. Again, I am taking into account the r7 is a cropper.

Cheers, I appreciate your taking the time to peruse my moaning about the camera!

Original size: 800 ISO @ 1/4000 f8


Original size and 1.1 crop. 1600 ISO.

 

[Bobtographer]

Didnt Canon say they could micro adjust your lenses to your R7?

 

[paddler4]

The key to diagnosing this is creating minimal contrasts. I'd amend my earlier suggestion. Start with a brick wall and manual focus. If that's not good, and the lens has functioned well on other bodies, you pretty much know it's the body. If MF works but AF doesn't, you again know what the problem is. Since you are experienced with cameras that have small sensors and high pixel densities, that must not be the issue.

Re Bobtographer's question: microadjustment is for the most part not relevant to mirrorless cameras. The back and front focus it is intended to correct arises because the autofocus sensor and the image sensor are not the same. They are the same in the case of mirrorless cameras.

Good luck.

 

[Archibald]

I understand what you're saying, but what I am seeing is quite different than just softness, which I could deal with.

I also have the OMD EM1 system and also have had many other cameras, so I feel I have the experience to see whether the problem is on my end as user error or if it's the camera. I used to make my living as a product photographer, so consider myself reasonably knowledgeable about image quality. I don't pixel peep, and am not rabid about perfect images either.

I'm seeing images from the r7 that are exposed properly, that do not have motion blur, and have a high enough shutter that it should easily pick up detail. I shot a hawk overhead yesterday that almost filled the frame, would have been sharp as a tack on my R, yet the r7 shows no eye detail at all, just a grainy image. I have to reiterate the image is grainy/gritty, not noisy, there's no colour noise. The ISO was 800, which should be low enough even for a crop camera, when the subject almost fills half the frame. I've seen people post images on the r7 shot at 6400 and much more, with good results on editing. The images I get are not editable, you can't draw the detail out.

I am taking into account the r7 is a crop camera and not as good with noise as a full frame, and smaller pixel depth. Never had this problem with m43s.

Here's another example of what I mean: the merganser image is the full size RAW, and the detail shows how crushed the blacks are.
I don't know if it will show with the upload, but if you look at the white parts of the wing on the male on the screenshot, you'll see that grainy, gritty look that I mentioned. Just mushy without decent detail. These are completely unedited. 800 ISO - 1/4000 ss - 560mm. I do feel the birds are large enough in the frame that I should be getting decent detail. I don't think I'd be getting diffraction at f8? I have also shot at f9 with the same result.

I have a feeling the AF is either off, and is misfocusing just enough to crush the detail, or maybe the sensor is defective. I have tried all the different cases for stickiness, taken without the 1.4 tc, all the settings for birds recommended by different experts. I'm not a world-class photographer by any means, NatGeo isn't blowing up my phone , but I think I should be able to do better than these.

I've shot a few thousand images thus far on the r7 and my keeper rate is absolutely abysmal. Maybe eight or nine images I am *reasonably* happy with. All of birds ten feet away and sitting still which is why I'm wondering if the AF is misfocusing at further distances. But even those images were only a few good ones out of bursts.

I've included a second sample image of an osprey, shot under the same conditions, smaller in the frame and shot at 1600 ISO on my R, it has clear detail despite being much smaller in the frame - even the darker bird has lots of detail. Again, I am taking into account the r7 is a cropper.

Cheers, I appreciate your taking the time to peruse my moaning about the camera!

Original size: 800 ISO @ 1/4000 f8
View attachment 209802View attachment 209812

Original size and 1.1 crop. 1600 ISO.
View attachment 209813View attachment 209814

Your examples look a lot like some of my R7 images. They are taken in difficult shooting situations. There is motion, both of the subject and of the camera, and the subjects are some distance away. The R7 isn't great at acquiring focus in situations like that. (It can struggle with some static subjects too.)

I was out this morning with some photog friends shooting birds. It was a tough session with very few keepers. One of the guys was using a Sony system and judging from what his LCD was showing, he was getting much better focus of a particular subject than I was. This was a distant subject. Yet in other situations my gear can do just as well.

I'm not going to show you the bad pics. But this sparrow was fairly close and allowed me to get a good shot. It's about a 50% crop (of the width).

 

[Lester Wareham]

It is out of focus for sure.

The tracking can struggle to keep up with BIF but you can get tack sharp results. Was the subject range changing rapidly in the example ?

Can I suggest you go back to basics with stationary subjects (the brick wall is a good idea) then check with static birds in optimal conditions.

If it is not sharp with static it will not be sharp with IF.

I have noticed the R bodies can get stuck more on background subject detail than DSLRs but your background in your subject image is clean.

 

[Greyswan]

It is out of focus for sure.

The tracking can struggle to keep up with BIF but you can get tack sharp results. Was the subject range changing rapidly in the example ?

Can I suggest you go back to basics with stationary subjects (the brick wall is a good idea) then check with static birds in optimal conditions.

If it is not sharp with static it will not be sharp with IF.

I have noticed the R bodies can get stuck more on background subject detail than DSLRs but your background in your subject image is clean.

I use the Eos R and get tack sharp results. The R is an old camera, and yet better at following focus even on faraway subjects, as well as close. The mergansers are not, in my opinion, that far that it would be difficult to get a good focus. And there's no motion blur in that image, as you can see from the edges.

Didnt Canon say they could micro adjust your lenses to your R7?

I really don't want to send both the lens and camera in, I need that lens for work. And I don't believe mirrorless needs those adjustments. I have done that myself with my dslrs. at home.

Shooting a brick wall, with no IS and no Ibis, on a tripod was clear enough, but not as sharp and crisp as I think it should be. The merganser is not that far away, and @ 1/4000 ss it should be clear. I shot a burst, and this was the only one that was even slightly clear. Blue sunny skies, open areas, no distracting focus objects, high ISO... what other considerations are there? I'm pretty good at panning as well. They flew in a straight line.

As for static birds, I've been shooting smaller perching birds with the same results, maybe one in twenty is sharp enough (but still not what I consider great!) Misfocusing all the time. The focus seems to have been grabbed behind the subject fairly often as well.

I do realise that a crop with smaller pixels will be more sensitive to micro movements, but I don't believe I am that bad that it would affect 95% of the images.

I see focus confirmation through the viewfinder, on wide angle focus and spot focus, and looking at the images in DPP showing the focus point is correct, it still shows a flat image with no detail.

Archibald: I cannot achieve that kind of detail on the r7 even with birds ten feet away. - I can do that on the R and the OMD. Great pic. But I simply am not getting the kind of detail that I see posted everywhere with this camera.

I will send it in to Canon for a checkup, and I'll see what they say. I'm beginning to hate it, lol. Even my old 6D was far better at focusing quickly. I should mention it's a brand new camera, not used.

Cheers, everyone, thank you all for your input, much appreciated.

 

[Archibald]

You've only shown us pics from tough shooting conditions. You were challenging the camera. If ideal conditions still give poor IQ, then indeed there is a problem. If it's new, return it.

 

[Bobtographer]

What focus settings are you using on the R7?

 

[Capn Jack]

All other things being equal, a 100% crop from an R7 will look softer than those from many other cameras because of the high pixel density and hence small photosites. 32.5 MPX is a lot for an APS-C sensor, leading to a pixel pitch of only 3.2 microns. Even for a superb lens like the 100-400, that's pushing it, and with a TC, the resolution of the lens drops.

Under what conditions? If the pixel size is smaller than the point spread function for a particular lens, the image is merely oversampled, which is a good thing! Looking at an image at the pixel scale will make an oversampled image look blurry compared to one which is less oversampled.

Also the resolution of a lens is the same regardless of the sensor used, referring to the last quoted sentence.

 

[paddler4]

yes, of course, the resolution is the same, but the pixels are smaller relative to the resolving power of the lens and the circle of confusion. There is no oversampling in a 100% crop, AFAIK. It just blows up the image so that one pixel in the image corresponds to 1 pixel on the display. Or is that wrong.