EOS M50 - Sensor or CR3 Data issue? Vertical Banding & Color Loss

[AiDon]

Hi,

I have attached a couple of images that show a distinct problem with RAW images from the EOS M50, unfortunatly it is 3 years old so there will be very little will to look at these and I belive expensive to correct.

The first issue is the inability to recover shadow detail from an image shot at -2EV either in DPP or in other RAW processors and the vertical banding can be seen distinctly.. The general consensus amongst the commercial RAW image processors is that there is simply not enough data available in the RAW image to recover the detaill. Note that the attached screenshot showing the banding is from DPP.

This second image shows skintone missing in the areas noted. It is also missing from the RH side of the actual image and shows up as a grey blotch on the skin. Again it cannot be recovered because there is no color detail.

These are just examples of some problems with the images coming out of the EOS M50 and seems to be mainly associated with backlit images. Note I also have other camera brands that don't exhibit the same symptons in similar situations, and Canon mirrorless or DSLR writing CR2 images don't have the same issues.

Note it is easily repeatable. Is there any way that I can correct this issue as it doesn't seem limited to the M50 as I have also seen similar issues with the EOS R & RP reported here and on other forums.

Is there any way the engineers can explain this issue? I am also willing to send the images so they can examine.

Regards, Don

[Peter]

I am watching your pictures from my phone, but take a look if you think it is this https://community.usa.canon.com/t5/EOS/Eos-R-Strange-vertical-banding-help/m-p/316635

Send me some raw files (Google drive or Dropbox) if you want me to take a look at them and perhaps even try to correct them if the banding is dynamic row noise.

[ebiggs1]

"The general consensus amongst the commercial RAW image processors is that there is simply not enough data available in the RAW image to recover the detaill."

I agree with these folks. There is just so much you can do. Everything has its limits. DPP4 is a capable program but I prefer Photoshop.  I think you will have much better luck recovering any possible detail with PS. The skin tone thing, I think you  might be too critical and that is likely just how the camera saw the portrait.

The banding issue will be worse in extreme conditions. Don't go crazy with the editing sliders especially clarity.

[AiDon]

---

@Peter wrote:

I am watching your pictures from my phone, but take a look if you think it is this https://community.usa.canon.com/t5/EOS/Eos-R-Strange-vertical-banding-help/m-p/316635

Send me some raw files (Google drive or Dropbox) if you want me to take a look at them and perhaps even try to correct them if the banding is dynamic row noise.

---

I think it is similar but the issue here is that there is simply not enough data to recover anything more. And as it is only -2EV it is quite strange, but it is a color that is a single color. The issue is that I need to be able to recover the details but there is no way to do it using DPP, PS, Capture One etc.

I really think the issue is a sensor issue.

[AiDon]

---

@ebiggs1 wrote:

"The general consensus amongst the commercial RAW image processors is that there is simply not enough data available in the RAW image to recover the detaill."

I agree with these folks. There is just so much you can do. Everything has its limits. DPP4 is a capable program but I prefer Photoshop.  I think you will have much better luck recovering any possible detail with PS. The skin tone thing, I think you  might be too critical and that is likely just how the camera saw the portrait.

The banding issue will be worse in extreme conditions. Don't go crazy with the editing sliders especially clarity.

---

Yes, obviouslty the skin tone issue is a dead loss and Clarity is not an issue with the other image (actually I have lots with the vertical banding) as it is not used. Simply using Shadow recovery accentuates the banding. And adjusting blacks doesn't help.

These issues are not there in images from 5DsR, 5D Mk3, 7D or my wife's M100, but they all use CR2 and different sensors except for the M100. Note not there in my Sony work cameras either, but that is neither here or there.

[Peter]

---

@AiDon wrote:

---

@Peter wrote:

I am watching your pictures from my phone, but take a look if you think it is this https://community.usa.canon.com/t5/EOS/Eos-R-Strange-vertical-banding-help/m-p/316635

Send me some raw files (Google drive or Dropbox) if you want me to take a look at them and perhaps even try to correct them if the banding is dynamic row noise.

---

I think it is similar but the issue here is that there is simply not enough data to recover anything more. And as it is only -2EV it is quite strange, but it is a color that is a single color. The issue is that I need to be able to recover the details but there is no way to do it using DPP, PS, Capture One etc.

I really think the issue is a sensor issue.

---

If it is dynamic row noise it may be possible to correct it, but not with PS, DPP or Capture One.

[ebiggs1]

"These issues are not there in images from 5DsR, 5D Mk3, 7D or my wife's M100,..."

You are correct because these are all different sensors and have different characteristics. They are going to have differing IQ. The M50 was/is not a top of the mark camera like a 5 series.

[ebiggs1]

"If it is dynamic row noise it may be possible to correct it, but not with PS, DPP or Capture One."

I totally disagree with this about Photoshop anyway, I have no idea about Capture One, never used it, and I don't use DPP4 enough to be an expert on it. But if PS can't do something no photo software can either. Just because you may not know how to make PS correct something or you have an automated software for correction is no indication that PS can't do it.

[AiDon]

It seems that the issue is mainly concentrated in the Red channel with a little in the Blue channel.

RED Channel

BLUE Channel

[Peter]

---

@ebiggs1 wrote:

But if PS can't do something no photo software can either.

---

Seems like a great topic here https://community.usa.canon.com/t5/Software/bd-p/software

Dynamic row noise, demosaicing methods and hot/dead pixel filters could be interesting to compare.

---

@ebiggs1 wrote:

Just because you may not know how to make PS correct something or you have an automated software for correction is no indication that PS can't do it.

---

Glad you want to help us out here. Here you have some raw samples to play around with  https://www.dpreview.com/reviews/image-comparison/fullscreen?attr144_0=canon_eos77d&attr144_1=canon_...