5D4 Sensor Defect Discovered

[tron]

Where's Dilbert on this subject? Seems strange without him!

Jack

Jack pleaaaaase! It is not funny! To day UPS picked up my 5D4 to return it to dealer. Kudos to him for arranging everything! He even agreed to perform some tests. I included a usb stick with 5D4 cr2 files and sinilar 5D3 ones to make comparisons. I even included latest DPP and DNG converter software to make it easier.

Since it's UPS standard it will take a week for the return, then a few days for selection/test of a new 5D4 and then another week for delivery...

 

[tron]

We have met the enemy and he is us. Having to boost a photo 5 stops is not a sensor defect, something else is defective.

I just returned my camera for a similar reason. The defect showed in black only with -3 EV Exposure Compensation during exposure and boost in post. A similar 5D3 photo was MUCH better (with the known shadow problems in some very small completely dark (even after exposure) parts. It is not as if I use this way of shooting daily. I don't so I didn't discover it when I shot a low light indoor event but I need a camera that is a superset of my 5D3 (in blacks/DR etc) a few days per year when shooting moonrise behind a temple. If 5D3 does not have this issue I cannot see why 5D4 must have it. I would like to know your opinion when/if you get 5D4, check to see if it behaves similarly and keep it at the same time. By the way I didn't boost shadows. The problem was visible anyway!

 

[Jack Douglas]

We have met the enemy and he is us. Having to boost a photo 5 stops is not a sensor defect, something else is defective.

I just returned my camera for a similar reason. The defect showed in black only with -3 EV Exposure Compensation during exposure and boost in post. A similar 5D3 photo was MUCH better (with the known shadow problems in some very small completely dark (even after exposure) parts. It is not as if I use this way of shooting daily. I don't so I didn't discover it when I shot a low light indoor event but I need a camera that is a superset of my 5D3 (in blacks/DR etc) a few days per year when shooting moonrise behind a temple. If 5D3 does not have this issue I cannot see why 5D4 must have it. I would like to know your opinion when/if you get 5D4, check to see if it behaves similarly and keep it at the same time. By the way I didn't boost shadows. The problem was visible anyway!

There is no problem with realistic situations that show a malfunction and it certainly is valuable to have any malfunction exposed. However, the comment that you can test anything in such a way that it is caused to not perform acceptably, still applies.

Another comment about testing to destruction not being useful is probably not completely accurate either since I'm sure we could all come up with examples where destruction leads to deep soul searching and the discovery of things that would otherwise never be investigated.

Since nothing is perfect one might question whether one really wants to push their personal equipment to the extreme limit to expose whatever shortcoming it has. If everyone does this and if there is actually no practical reason why the limit needs to be pushed, the net effect will be the company having to replace units at a higher rate, and costs ultimately being passed on to customers.

So, really what it comes down to is whether a person is being "picky" and where to draw the line. We all have different standards in this regard. It would be interesting to compare this scenario to the one of people picking up their new car and looking for paint defects or whatever. The pickiest person will always find defects in every car and yet sometimes those same folk have chips and scratches after a few weeks that would exceed any factory defects.

I'm not making any judgment on this particular camera issue relative to the OP but some comments in the thread have been over the top, which has resulted in responses that are not tolerant of the lack of clear judgment.

One thing for sure, when a response comes from someone, with the preface of having great expertise accumulated over many years, and that is used to put down someone who has demonstrated expertise over many years down, I know who I'll tend to believe.

Jack

 

[Deleted member 7828]

Well, for what it's worth, I went back and took some test shots with the 5D4, deliberately underexposing and then pulling up.

I can confirm that I am able to reproduce the issue that the OP and others have brought up, but it only seems to occur in certain, extreme conditions. Strongly underexposing in low-light conditions (2 stops or more) at base/low ISO (particularly if there is a lamp in the frame or just outside) seems to reproduce the issue most strongly. Oddly enough, at higher ISOs it doesn't seem to be a problem (maybe the higher noise levels obscure it?).

I never saw it in any of my real-world shots, but in my tests I'm able to see it in varying degrees - lines of green tint (occasionally red or blue) in certain dark and/or shaded areas of the image. I guess the real question is whether this is simply a quirk of the 5D4 sensor or whether it's actually a legit issue where some cameras exhibit this "phenomenon" (Canon's word for bug) while others do not. Either way, I'm not sure it's a huge deal for me - as long as it's not showing up in my everyday, real world images, I'm OK.

If Canon later comes out and ends up issuing a service advisory for this, so be it - but as far as I can see so far, it only seems to happen in extreme circumstances and it doesn't seem to be anything for me to get really concerned about...at least for now.

 

[TimoV]

I can also confirm there is something wrong in my 5Dmk4 camera and it can see in real world photos…

I take some shot at night time in the city and outside shopping mall and later look photos from my PC, all parts of dark night sky was full of dark stripes, so near to black level there are visible horizontal stripes (if boosting shadows even little it’s coming more visible), other Canon cameras I own earlier don’t have this "phenomenon".

This stripes can be visible without boosting anything in the photo, stripes changes little photo by photo and some situation there is little color tint.

Shots was taken RAW and base ISO 100 and can see in newest Lightroom and DPP.

I return this camera already to Canon (Finland) and now waiting the results…

 

[unfocused]

... If Canon later comes out and ends up issuing a service advisory for this, so be it - but as far as I can see so far, it only seems to happen in extreme circumstances and it doesn't seem to be anything for me to get really concerned about...at least for now.

I think this very effectively sums things up. No one on this forum can know whether or not this is an actual "defect" or simply a result of pushing a product beyond the limits of what it is designed for. Many of us suspect it is the latter, but I'm willing to hold that thought until we know more from Canon. I would, however, caution that simply because Canon may choose to replace or repair a camera that displays this phenomena, that doesn't necessarily confirm that it is a "defect." Many times, manufacturers will exchange a product simply because they want to keep a customer happy and it's cheaper and easier to ship out a replacement.

 

[tron]

... If Canon later comes out and ends up issuing a service advisory for this, so be it - but as far as I can see so far, it only seems to happen in extreme circumstances and it doesn't seem to be anything for me to get really concerned about...at least for now.

I think this very effectively sums things up. No one on this forum can know whether or not this is an actual "defect" or simply a result of pushing a product beyond the limits of what it is designed for. Many of us suspect it is the latter, but I'm willing to hold that thought until we know more from Canon. I would, however, caution that simply because Canon may choose to replace or repair a camera that displays this phenomena, that doesn't necessarily confirm that it is a "defect." Many times, manufacturers will exchange a product simply because they want to keep a customer happy and it's cheaper and easier to ship out a replacement.

Two things:

1. There are no "mirracle" products of course but I assume that my 5D4 should be better than my 5D3 in similar conditions. It is not.

2. The dealer will indeed replace it but he also agreed to make some tests himself using the replacement camera. I told him of the specifics plus to make things easier I sent him a usb stick with the 5D4 test raw files, the 5D3 similar ones, Canon's latest DPP and Adobe's free DNG converter. I do not know what will happen but for now I ... have hope...

 

[Jack Douglas]

I'm thankful Canon has a good service record and of course there is hope.

I just hope my new 1DX II doesn't exhibit any issues. After I got it I did get the funny rectangle down in the right bottom corner but knew immediately to check the firmware number and then upgraded. That gave me the jitters - only worked after I found out that in-camera formatting the card still left empty folders, which I had to delete.

Good luck.

Jack

 

[tron]

I'm thankful Canon has a good service record and of course there is hope.

I just hope my new 1DX II doesn't exhibit any issues. After I got it I did get the funny rectangle down in the right bottom corner but knew immediately to check the firmware number and then upgraded. That gave me the jitters - only worked after I found out that in-camera formatting the card still left empty folders, which I had to delete.

Good luck.

Jack

Thanks! As far as 1DxII you are talking obviously of the Sandisk bug which has been handled with a firmware update. The only possible issue I have read about is auto focusing during servo AF but I guess by now - as a bird shooter - you will have verified that your camera is fine

 

[Jack Douglas]

I'm thankful Canon has a good service record and of course there is hope.

I just hope my new 1DX II doesn't exhibit any issues. After I got it I did get the funny rectangle down in the right bottom corner but knew immediately to check the firmware number and then upgraded. That gave me the jitters - only worked after I found out that in-camera formatting the card still left empty folders, which I had to delete.

Good luck.

Jack

Thanks! As far as 1DxII you are talking obviously of the Sandisk bug which has been handled with a firmware update. The only possible issue I have read about is auto focusing during servo AF but I guess by now - as a bird shooter - you will have verified that your camera is fine

Hard to verify if the camera is fine since I'm not fine.

Jack

 

[tron]

I'm thankful Canon has a good service record and of course there is hope.

I just hope my new 1DX II doesn't exhibit any issues. After I got it I did get the funny rectangle down in the right bottom corner but knew immediately to check the firmware number and then upgraded. That gave me the jitters - only worked after I found out that in-camera formatting the card still left empty folders, which I had to delete.

Good luck.

Jack

Thanks! As far as 1DxII you are talking obviously of the Sandisk bug which has been handled with a firmware update. The only possible issue I have read about is auto focusing during servo AF but I guess by now - as a bird shooter - you will have verified that your camera is fine

Hard to verify if the camera is fine since I'm not fine.

Jack

? ? ? ? ? ?

 

[Jack Douglas]

Fine as in a good photographer.

Jack

 

[tron]

Fine as in a good photographer.

Jack

You are more than a good photographer as all your posted photos verify. So sharp photos, fine camera, fine photographer...

 

[MichaelTheMaven]

Thank you Kevin for your help, I never head back from you so I went ahead with my video and just posted it on Youtube. Hopefully we can get some better numbers to understand how many cameras are affected and get a solution or firmware update or something from Canon. I don't want this to become a thing where they deny it exists, and if we make enough noise maybe they will look at it.

Thank you guys for helping with the tests!

Here is the video on youtube:

https://youtu.be/R4c78yuHgMY

 

[ilumo]

not having any issue with mine... as far as I can tell. I was reading on dpreview that it may be some sort of radio interference? has anyone tested that theory out? I took a few test shots and they look clean as a whistle even at +5 stops.

 

[ilumo]

oops. here are the other two shots. the underexposed shot and the 100% crop of the middle portion thats pushed the most.

 

[LSXPhotog]

Thank you Kevin for your help, I never head back from you so I went ahead with my video and just posted it on Youtube. Hopefully we can get some better numbers to understand how many cameras are affected and get a solution or firmware update or something from Canon. I don't want this to become a thing where they deny it exists, and if we make enough noise maybe they will look at it.

Thank you guys for helping with the tests!

Here is the video on youtube:

https://youtu.be/R4c78yuHgMY

Hey Micahel, sorry I never got those files over to you. I have been really busy and was also doing some testing on my own. I even went ahead and bought a SECOND 5D Mark IV to test against my original 5D Mark IV.

Head-to-head, the cameras both exhibit the exact same issue. The new one so a slightly lesser extent, but not anything I would say was an improvement, but it consistently has cleaner files. I have also discovered the problem is with boosting shadows and not so much the exposure. The problem persists at 100 ISO when you push 2.5-3 stops and is heavily exaggerated to the point of being comically unusable at 4 and 5 stops. ISO 50 is more resistant to the problem and only shows the streaking at around 4 stops - which is good news.

What I didn't expect to find was that the new camera produces different colors when the images are matched to the same white balance. The new 5D are slightly more orange at the same color temperature/tint. Just something I noticed.

Now that I've confirmed both cameras have the problem. I'll be sending the new one back to B&H and the old one to Canon, as they requested.

I don't think every 5D Mark IV has this problem, but I think there might be more than we thought...

- Kevin

I don't think Canon is going to do anything about this, unfortunately.

 

[LSXPhotog]

Also want to add:

1.) Dual-Pixel RAW had absolutely no impact on the file's performance in regards to the streaking.

2.) Seriously, guys, the camera is absolutely unusable when you push a file 3-5 stops. We waited 4 years for a camera with more dynamic range and then when you attempt to explore that dynamic range, you find out your camera can't handle it and others can. I'm going to keep my camera and keep using it, but this is not the news any of us want to hear/see. Most of the people I'm seeing that are making a big stink about this not being a problem don't even own the camera. LOL

 

[ilumo]

Also want to add:

1.) Dual-Pixel RAW had absolutely no impact on the file's performance in regards to the streaking.

2.) Seriously, guys, the camera is absolutely unusable when you push a file 3-5 stops. We waited 4 years for a camera with more dynamic range and then when you attempt to explore that dynamic range, you find out your camera can't handle it and others can. I'm going to keep my camera and keep using it, but this is not the news any of us want to hear/see. Most of the people I'm seeing that are making a big stink about this not being a problem don't even own the camera. LOL

you know what... I think I might have replicated that "banding" problem. hah.
Here's a black frame. 20 sec exposure at base iso. pushed 9 stops in post. 100% crop

 

[Jack Douglas]

Ha, ha, ha, you should be a comedian.

Jack