Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Hot pixels

Tresch said:
I find it concerning that the sensor could have dead pixels problems and therefore I kindly ask you to make some more tests (black shading...) and give your results on this thread. :-)
Click to expand...

Well, it's not something that you see in footage with *normal* frame rates or ASA settings so i wouldn't categorize it as a problem yet. However i have done some long exposures were i was able to see thousands of hot pixels in the frame. It was scary.

sutherlandprojects said:
I think I also heard once that the black shading might do some dead pixel compensation as well, not sure if this is true or imagined though.
Click to expand...

Jannard said:
If it were a dead pixel issue, it would show up under all circumstances. Do the recommended Black-cal and you should be fine. As mentioned, remember to do it again when you come out of that mode.
Click to expand...

Yes commander. I will test again some long exposures with black shading to see if the problem is gone.

Thanks for the help.
Miltos
 
!@#$... I've shot a lot of stuff with the RED, and I made the (false) assumption that a black shading calibration is good for all shooting modes. I just finished shooting two weeks in Israel with plenty of long exposure varispeed and timelapse stuff.

I was able to lower the effects of the hot pixels by turning off OLPF compensation in RedAlert before exporting (actually there's a bug in RedAlert with long-exposure stuff that turns any noise into a pure white pixel).

Jim, Stuart or others - could I put in a request to improve the calibration routine so that the camera can be depended on for all shooting modes once it's been calibrated? e.g. have the calibration routine also run through some frames at long exposure times to find the deep noise that's not obvious in normal shooting? This could also potentially help with high-ISO shooting scenarios.

This is one of those frustrating things of the camera still really being in Beta...
 
Tervor beat me to it, but I'd like to request the same. Would it be possible to do more extensive black shading to encompass a wider range of exposure settings? I'm thinking a "layered" calibration file or something more like a 3D LUT in nature. Of course, I'm just guessing here as I don't know the inner workings of the RED One... Don't know if it's even possible or practical.
 
Tresch said:
I don't know the implications of dead pixels on a CMO as there will be a debayer applied on the dead pixel and the sum of 1 pixel value is created by surronding pixels.

I find it concerning that the sensor could have dead pixels problems and therefore I kindly ask you to make some more tests (black shading...) and give your results on this thread. :-)
Click to expand...

All sensors have such pixels. You should not be concerned.

If troublemaker pixels are adjacent then that is problematic in many situations, not just deBayering. Also please not that there are other type of troublemaker pixels than just dead. Dead pixels are actually easier. The harder ones are those that I might say "dark", for the lack of a better term, as they have some response, not just the right response. It appears the pixels in the problem mentioned are "dark" as they are appearing with a higher exposure time (low frame rate) or gain, and not surfacing under normal operation.
 
As most already say. There is no sensor without hot/ death pixel.
Hotpixel compensation is one of the difficult processing in the primar sensor data processing.
Every company has its secret way how to handle it.
At the other side there are differnet ways to handle hotspot in long exposition like in a DSLR or in a timelapsmode of an cam. and in the short expositure area less than 100ms for a frame.
I think REDS blacklide shading is optimzed for film funktion.
Maybe RED in a future build can supply two storage two calibrations one for film and one for timelaps.
BTW Hotspixel is ame like noise. Its increase with the sensor tmep.
 
Joofa said:
All sensors have such pixels. You should not be concerned.
Click to expand...

???

Sorry I don't understand you!

Joofa, I also quoted "all sensor have dead pixels issues" but I don't understand you why you are saying "you should not be concerned"...

Thanks PeterM for your comprehensive post.

I wonder what exactly the black shading is doing in the RED One camera? So we (users) could know when it makes more sense applying a black shading. On some camera the dead pixels are automatically counted and killed during the "booting" process of the camera.


Patrick

PS: I had 2 camera bodies (Canon XLH-1) I had to replace because of that issue. It was no more possible to kill those dead pixels through camera software.
 
Hot pixels are part of life with digital sensors. Software & bayer math takes them out of the equation.

What we need is for RED to improve their calibration so we don't have to think about "will there be hot pixels or not in my current shooting mode?"
 
Tresch said:
???

Sorry I don't understand you!

Joofa, I also quoted "all sensor have dead pixels issues" but I don't understand you why you are saying "you should not be concerned"...
Click to expand...

Sorry, Tresch, I think I misinterpreted you. I thought you were expressing concern in the following line:

Tresch said:
I find it concerning that the sensor could have dead pixels problems and ...

Patrick
Click to expand...

Anyway, sorry for being rude.
 
Tresch said:
I find it concerning that the sensor could have dead pixels problems and therefore I kindly ask you to make some more tests (black shading...) and give your results on this thread. :-)
Click to expand...

Today i had some quiet time and it was a good opportunity to try out Build 17 and do some more testing on the hot pixel problem.

As Jim suggested so reassuringly to us:

Jannard said:
Do the recommended Black-cal and you should be fine. As mentioned, remember to do it again when you come out of that mode.Jim
Click to expand...

So, i set Varispeed to 1fps and the milky way appeared on the LCD as expected. If i hadn't shot some great footage yesterday i could have had a heart attack looking at all those hot pixels and thinking my RED is busted.

Anyway, i started the BlackShading procedure and after a couple of minutes i saw the exact same hot pixels decorating my RED LCD. I repeated the procedure again and got the same result.

After trying the same with another RED (earlier serial than my #1473) and getting similar results i can safely and reassuringly say that Black Shading is fixing something but that ain't the Hot Pixels.

Hot Pixels are still here and demand a solution.
 
Miltos Pilalitos said:
Today i had some quiet time and it was a good opportunity to try out Build 17 and do some more testing on the hot pixel problem.

As Jim suggested so reassuringly to us:



So, i set Varispeed to 1fps and the milky way appeared on the LCD as expected. If i hadn't shot some great footage yesterday i could have had a heart attack looking at all those hot pixels and thinking my RED is busted.

Anyway, i started the BlackShading procedure and after a couple of minutes i saw the exact same hot pixels decorating my RED LCD. I repeated the procedure again and got the same result.

After trying the same with another RED (earlier serial than my #1473) and getting similar results i can safely and reassuringly say that Black Shading is fixing something but that ain't the Hot Pixels.

Hot Pixels are still here and demand a solution.
Click to expand...

Same results here on several cameras. I've done the timelapse for just only a couple of productions. Thousands of hot pixels at 1/2 second exposure.

Black shading definitely doesn't solve the issue.

Finishing in SD for broadcast meant the issue wasn't a problem.
But anyone doing 1 or 2 fps on the Red should be aware of this issue, especially if you want to crop the image, or finish for HD or film.
Red should be pushed to have a fix in the next camera models.
 
When to fix/patch pixels...

When to fix/patch pixels...

Mark Pugh said:
Does anyone know of any software that can detect and correct the bad pixels in post? Can a de-speck filter go a way towards fixing it?
Click to expand...

Hot pixels or bad pixels etc, should be fixed before the de-Bayer (de-mosaic) since anything that is not good sensor data of the lenses image will alter the interpolated data affecting many pixels around the one that is off value.

After you save a TIF to cook with external filters the data has been interpolated already.

It would really be best to fix pixels in the camera since the camera uses compression that depends on the relative values of adjacent pixels of the same color field.
 
When you do a black calibration for your Timelapse, make sure your shutter is set as far away from 1/48th as possible.. to match how you are driving those pixels.
 
Do a full debayer. It reduces hot pixels considerably.
 
This has turned into a more significant issue with the new COLOR workflow. The RED tab in Color gives you no control over OLPF settings, which results in nasty artifacting on any timelapse footage. That's a PITA, because those files now have to be treated manually via REDAlert to get rid of the noise.
 
You must log in or register to reply here.
Back
Top