EOS R5 Mark II issues with colors and corrupted images

[jasonleecs]

Hi,

Any R5M2 owner encountered a "layer" of additional/missing colour platter in your photo taken?

Initially, I encounter image preview issue where the camara states that the "current image cannot be previewed" on the camera but didn't thought much about as it was already night and I'm shooting the cuscus near the homestay. But things took a turn the next day when I'm shooting the bird of paradise. A quick preview on some of the images show symptom the night before and worst, photos with additional/missing colour platter.

As I continue to review the images, I noticed some of the images either have a tinge of colour, e.g. green, yellow, pink etc., either in in full or partial on the images.

Initially, I thought it could be due to the slower SanDisk UHS 2 SD card wasn't able to catch-up with the write speed that causes the problem. So, I removed the SD card and let the camera write to the faster Prograde CFExpress card with the sustainable write speed of 1300MB/s.

However, the issue persisted and even more ridiculous! I took the camera for a sunrise shoot one morning, instantly, the 1st shot encountered the problem describe above on the CFExpress card. Therefore, the theory of incapable card or heating is ruled out. 

Would like to hear from follow R5M2 owner if you've encountered such issues.

The camera config is as followed

Firmware v1.01 // Highlight tone priority D+2 on shooting the bird, and was disabled when shooting sunrise // Elec shutter for BOP @ H+ and Mech shutter for sunrise at single // SanDisk 256GB Extreme PRO SDXC UHS-II Memory Card with R=280MB, W=150MB // ProGrade Digital 1TB CFexpress Type B Memory Card (GOLD) with R=1700MB, R=1300MB

Thanks.

[Peter · Accepted solution]

The masked pixels are good to use to set the right black level or to reduce row noise. Black level 0 will give you the pale colour cast. Canon normally uses something around 512 at ISO 100 and something around 2048 at higher ISO. If you don't want to use the masked pixels to set the black level you can use the information from the Makernotes. It is often correct even if there are exceptions. DPP, Lightroom and DxO read the black level from the Makernotes. Here I have changed the Makernotes from 512, I think, to 0.

 

Using HTP 200 means that you are using ISO 100 under exposed 1 stop. The raw file will therefore be 1 stop darker. DPP then adds a tonecurve to make it brighter.

[Peter · Accepted solution]

Yes, corrupted. One of three embedded JPEG files corrupted too. This one is still okey but only low resolution. Extracted with Exiftool.

The blue pixels in the raw file are in the bad area hitting 16383. It means that they clip the highlights and do that even in the masked pixels. The value there should be 512.

[Waddizzle]

It seems like there could be an issue with the camera.  

Let’s dial down your settings and test the camera.  Set the camera to One Shot AF and Single Shot Drive mode.  Does the issue still appear?  

[jasonleecs]

Well, it does. It happened on the sunrise shot, in fact, that shot was shoot with single drive and manual focus.

[Peter]

Your second print screen shows the full size of the raw files. That is how a raw file from Canon looks like when it is not cropped and without black level 512/2048. Why people don't know that characteristic look is because they use software that hides those steps.

[jasonleecs]

No, it should not. It is clearly a defect as the pic looks like a layer of misty white was applied on the centre pic, while the one on the right is alright.

[Peter]

The masked pixels are visible on all the images there in the second screenshot.

Here from my 6D

 Your D+2 is under exposed, and that is normal too.

[jasonleecs]

Thanks Peter.

I agree with you after closer examination. But I don't understand how the mask was applied to the RAW files. I used FastStone for previewing images prior to selecting it for edit/delete, and that as you mentioned, it was mask. In fact, only file #from *049 ~ 061 have the issue. Those before and after does not have the mask applied.

Viewing the file thumbnail in Windows Explorer also show the same. When I opened it on DPP 4.18, DPP can't even load the few files that was masked.

Upgrading it to 4.19 seems to solve the issue and shows the image correctly.

DPP 4.18FastStone

[p4pictures]

So this is not a camera issue at all, but an issue with FastStone and Windows Explorer. The images are 100% as expected in the DPP version that supports the EOS R5 Mark II.

[jasonleecs]

The key issue here is about image corruption with the R5M2 writing to the V60 SDXC cards. Canon in my region have acknowledge the issue and have elevate the issue to Canon Inc and they are still pending the 

[Peter · Accepted solution]

The masked pixels are good to use to set the right black level or to reduce row noise. Black level 0 will give you the pale colour cast. Canon normally uses something around 512 at ISO 100 and something around 2048 at higher ISO. If you don't want to use the masked pixels to set the black level you can use the information from the Makernotes. It is often correct even if there are exceptions. DPP, Lightroom and DxO read the black level from the Makernotes. Here I have changed the Makernotes from 512, I think, to 0.

 

Using HTP 200 means that you are using ISO 100 under exposed 1 stop. The raw file will therefore be 1 stop darker. DPP then adds a tonecurve to make it brighter.