Camera sensors don't capture colors. Color is how humans perceive light. Color is psychology. Cameras measure light in their own way, different from how human eyes do. A camera's measurements get transformed into colors so they can be displayed. The transformation is arbitrary, and need not be limited to a certain color gamut.
C200: Raw workflow comment and a question about Canon Raw Development
- Thread starter fhammond
- Start date
Camera sensors don't capture colors. Color is how humans perceive light. Color is psychology. Cameras measure light in their own way, different from how human eyes do. A camera's measurements get transformed into colors so they can be displayed. The transformation is arbitrary, and need not be limited to a certain color gamut.
morgan_moore
Major Contributor
This argument is kind of reasonable if you are a philosopher who like to question if a chair is there when you are not in the room- but of course cameras capture colours (or more importantly don't)
If you have two objects one super bright pink and the other ultra bright pink and when the camera films them it records the same digital number for each then it cannot film one of those colours - and that colour is therefore out of gamut range of the camera.
A raw colour space should be digitally huge (aces) to fully digitse what the camera can resolve.
mikemilton
Member
How about this: "Camera sensors constrain the possible range of subsequent gamuts"
The bayer matrix has 4 primary colour filters (r, g, b, g) which are fixed. The underlying pixels likely have a spectral response curve, and the luminosity bit depth constrains the number of permutations of combinations of the above. So one could enumerate and plot those permutations into the full range of possible colours. I'd be happy to call this the sensor gamut - as that seems an accurate description.
The only requirement is that all the standardized gamuts of interest fall within this range.
The bayer matrix has 4 primary colour filters (r, g, b, g) which are fixed. The underlying pixels likely have a spectral response curve, and the luminosity bit depth constrains the number of permutations of combinations of the above. So one could enumerate and plot those permutations into the full range of possible colours. I'd be happy to call this the sensor gamut - as that seems an accurate description.
The only requirement is that all the standardized gamuts of interest fall within this range.
It's science and engineering, not philosophy. A camera sensor doesn't detect light the same way a human eye does, so it can't see colors in a way that corresponds to how humans do. If you build a camera, you will necessarily spend some time deciding how to transform sensor values to color values. No transformation you come up with will be accurate for all colors. You'll need to trade off, optimizing for one particular scenario, e.g. accurate skin tones, at the expense of everything else, e.g. accuracy of highly saturated colors. Only after the transformation from sensor values to color values do you have colors, in a color space, with a gamut. That gamut depends on how you decide to do the transformation. The sensor itself does not have any gamut. There is no such thing as a raw gamut.
You can read Capture Color Analysis Gamuts, or search for scene analysis transforms, color analysis transforms, color analysis error minimization, or color analysis matrices. It will be helpful to know how the CIE 1931 color matching experiments were performed, what the color matching functions are, the CIE RGB color space, the CIE XYZ color space, the xyY color space, metamerism, and different types of metameric failure.
kohl-sudduth
Well-known member
Well said. And thank you for the paper. Have a bunch of questions, but I won't clutter up the thread with them and it'll be good to figure out the answers on my own. This paper proves a useful tool in my halting quest to being able to eventually genuinely author images instead of simply consume those pre-fabricated by a camera manufacturer. May never happen but it's worth trying.
Definitely look forward to whatever discoveries you make about how Canon is creating its 'Raw Light' Data. Regards.
AlkaliDigital
Active member
Default Today 08:27 AM Add to this user's reputation Report Post
I just received my C200 and I shot tome test RAW footage. I was able to open the .CRM files fine in Cinema RAW Development 2.0 on my laptop but my desktop doesn't "see" the .CRM files. The CRM files are seen on my computer as "execute" only files??? However, Davinci Resolve 14 sees and works with the files fine! Go figure. (except export resolution issues because I believe the free version is limited to UHD and doesn't allow the export of the full res 4096x2160)...or am I wrong?
Has anyone else had this issue with Canon's software? so strange it works perfectly in Resolve but not in Canon's Cinema Raw Development software. It is so weird.....I have uninstalled and re-installed the software but nothing seems to be working. Mac Pro Tower 5,1 Current OS
I just received my C200 and I shot tome test RAW footage. I was able to open the .CRM files fine in Cinema RAW Development 2.0 on my laptop but my desktop doesn't "see" the .CRM files. The CRM files are seen on my computer as "execute" only files??? However, Davinci Resolve 14 sees and works with the files fine! Go figure. (except export resolution issues because I believe the free version is limited to UHD and doesn't allow the export of the full res 4096x2160)...or am I wrong?
Has anyone else had this issue with Canon's software? so strange it works perfectly in Resolve but not in Canon's Cinema Raw Development software. It is so weird.....I have uninstalled and re-installed the software but nothing seems to be working. Mac Pro Tower 5,1 Current OS
Last edited:
If you work at Avid MC then you can link a raw light with AMA plugin from Canon, like video (MC 8.5), and transcode this linked file to DNxHR (UHD/4K). I have MC 8.4, so I can't test this workflow.
Alvise Tedesco
Well-known member
I can't seem to set individual white balance values. whatever white I apply goes to all clips.
What am I doing wrong?
What am I doing wrong?
Danilo Del Tufo
Well-known member
Hello guys, a silly question : can the C200 record in Full HD into SD cards, just as a menu option, because the manual says examples and info about it just with SDI signal or proxies. I'm interested to use the camera in Full HD just like the C100. Thanks!
puredrifting
Major Contributor
One of the first C200 shoots I did was with my C100 as b camera, both shooting at 1080. Works fine and in WDR, they were both relatively close as far as color correction.
puredrifting
Major Contributor
I didn't eyeball it carefully as I didn't edit it but in the final cut, I didn't notice any big difference between the two.
Danilo Del Tufo
Well-known member
When do you shot in RAW Light with C200, the LCD screen and EVF shows "Log" colors? Or they're covered by REC709 like LUT? How it works?
Both are options. You can choose to view the log directly or you can apply a REC.709 LUT to it. The waveform monitor always shows the log regardless of LUT setting.
