Too many Dead Pixels in new 5D3 (Bridge vs CameraRAW)

[adrjork]

Hi everyone,
I've just bought a brand new 5D3, and the first thing I did was a sensor check (to verify dead pixels). I used Film Nurture method.

In my 4-shots test (HERE) I can see 20 or 30 dead pixels yet at ISO 100, and many more at ISO 800. Anyway, I can see all these dead pixels ONLY in Bridge, while they disappear in ACR. I tried to explain this difference remembering that ACR should have a sort of dead pixels fixer (even if I didn't suppose it was automatic).
So probably Bridge shows the "truth".

At this point I can choose: "mapping" the sensor's dead pixels with the in-camera manual sensor cleaning method (so keeping the camera with its patchy sensor), or returning the camera to Amazon asking for a replacement.

Before taking a decision, I'd like to know your opinions. Is it normal that a brand new camera has that amount of dead pixels? (Moreover, ACR is not in my workflow, so I can't eventually take advantage of its "corrections".)

Thanks in advance for your help.

[Ray-uk]

You are worrying too much, the test in the video is flawed, if you carried out the test as shown in the video then you should have got black frames and not the multi-coloured frames that you have shown, even then it is a very poor test and will not show a dead pixel which usually appears as a black spot.

The remapping procedure that they describe is very much open to question, even Canon will tell you that it doesn't work. What happens is that the pixels that seem to be hot/stuck/dead can change position on each shot or even disappear at times and people think they have fixed it when they have done nothing.

Just use your new camera and stop worrying about things.

[adrjork]

Thank you Ray-uk. I'll follow your advice.
Anyway - just to know - if the test in the video is not good, so which is a good test to verify dead pixels by myself?
Thanks again.

[Ray-uk · Accepted solution]

There is a free program called "dead pixel test" which you can download from photo freeware net which apparently gives you a count of problem pixels but I have never tried it so can't say if it is any good.

Even so as I said before you will find that problem pixels can show up in different places on different exposures so it is only a guide if the same pixels show as faulty every time.

[adrjork]

Thank you so much!

[TCampbell]

I’m just using my iPad so I’m unable to view the contents of .zip files... but generally speaking there are loads of reasons for “noise” and “dead” or “stuck” pixels just one thing.

If you’re getting a different amount at ISO 800 then you get at ISO 100 then you don’t have dead pixels... you just have noise (and maybe some pattern noise).

There’s a completely different way to measure “noise” based on determining the mean & std deviation of all the pixels in a “dark frame” (a frame shot with the body cap or lens cap on the camera so that nothing in the image is real light (aka “signal”)).

True stuck pixels would occur at any ISO and it would always be the same few pixels.  If this isn’t what you are seeing then youy’re probably seeing some kind of noise.

[adrjork]

Thanks TCampbell for your reply.

Actually ALL the visible colored dots at ISO 100 are still visible at higher ISO. Simply, at higher ISO I can see more dots (or probably I count more dots because they are more evident).

Based on your description, I'd say that it's definitively a dead pixel issue.

[Peter]

Checked the raw files. Just a few hot and dead pixels. Use your "manual sensor cleaning method" or remove in post.

[adrjork]

Thanks Peter for your check.