Welcome to our community

Be a part of something great, join today!

Register 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 :)

RED One / Mysterium Spectral Response

Chromaticity

Chromaticity

Graeme Nattress said:
Displaying a chromaticity plot based upon the xy chromaticities of the RGB to XYZ transformation is very misleading as to a camera "gamut". That sort of plot is reasonably meaningful for an output device, but not so for a input device.

Graeme
Click to expand...

Hmmmm Graeme the CIE xyY color space plot is a common way of comparing the color gamut of devices that produce and reproduce color information such as a camera to the RGB color space for example.

I have often wondered why this is not standard information provided by all camera manufacturers.

Doesn't this plot give an idea of the reproduceable colors that your camera is capable, and isn't there value in comparing this with other devices as well as film?

Or am I confused or missing something?

Regards

Michael
 
Is this why Red One is so bad under Tungsten, because the Blue channel is so weak?


Is the massive crosstalk in the Green channel a side effect of Bayer?
 
Zaphod said:
Is this why Red One is so bad under Tungsten, because the Blue channel is so weak?
Click to expand...
No, this is why Red is so fantastic in daylight.:biggrin:

Zaphod said:
Is the massive crosstalk in the Green channel a side effect of Bayer?
Click to expand...

Massive? Whad do you mean? Where do you see that? There is no (significant) crosstalk that I can see. Look at the response of film stocks if you like to see lack of colour separation between emulsions. The colour separation of Mysterium is extrceptionally good and incomparably better than film and most other cameras.

Are you working for Sony or have you not used Red, or both?
 
Every sensor, including human eye has a "massiv" crostalk. The reason is that in real word there is no fillter which curve is bluff enough.
From the curve red supply we can see that this sensor is one of the best in the category of spectral responce. In additional the sensity of the blue chanal is fine.
every camerea must find its methode to handle the " crossover" areas. As more ideal curve you have as much easy it becomes for the regarding signalprocessing to get the real color information
BTW, you would be supprised it you watch the spectral response of the human eye in green and red receptors. When i was in highschool I never understand why humans can see a color like orange
This curve does not give you an information of the photon sensity.
So its difficult to judge how noisy the picture is.
 
I was very interested to see the spectral curves, thanks for posting this!
I did some crude spectral measurements back in January, as you see here.
http://www.bealecorner.org/red/test-080108/page2.html

Of course I didn't have access to raw sensor data, just what was in the .R3D file. What I saw then, and do not see in Red's data posted here, is a significant response in the blue channel around 650 nm. So was that just some measurement artifact of my own, or did something change in the way the sensor data is interpreted into RGB since firmware version #12 ?
 
Or maybe the top-secret, Area 51, sensor, the "Black Red"... using that new ultra-light-sensitive, black silicon stuff. haha. :)
 
Blue into purple, and green/magenta sensitivity

Blue into purple, and green/magenta sensitivity

Michael Lindsay said:
On a similar note has anyone seen skewy colours (blue ending up purple being the most obvious) using the Rec 709 matrix (no problem with Camera or Redspace setting).
Click to expand...

Yes, the more the scene's color balance shifts away from 3200K (at least upwards), the more purple the blues get. The camera's color rendering varies a lot depending on the light used. We've seen the high green/magenta sensitivity, too.

Switching the matrix to "Camera RGB" or REDSpace makes all of these issues better (even after boosting the saturation to compensate for Camera RGB's more muted rendering). I've put up some comparisons (from five different cameras under six different lighting conditions) at http://provideocoalition.com/index.php/awilt/story/green_magenta/, mostly focused on R1 imagery processed using the default Rec.709 setting.

Looking at the Rec.709 results was a bit scary, but my initial explorations of Camera RGB make me feel a lot better, especially as I have a show coming up where color accuracy is paramount (and, I hope, universal). Some secondary grading may be needed in the basic soup, but Camera RGB looks like it'll deliver more accurate and consistent results.

I have more explorations to do, of course, and I'm curious what other folks are finding out and how they're handling color space selection.
 
Just my personal opinion, but I would recommend steering clear of choosing rec709 as the colour space in either Red Alert or Redcine, and use Camera RGB instead. Using rec709 as the output LUT is obviously fine. The rec709 colour space does very strange things, particularly on the magenta/green axis, and although it may make colours look 'pleasingly vivid' under some circumstances, it will make them look very wrong under others. If you want vivid greens in your image, do it in the grade!
 
Adam Wilt said:
Yes, the more the scene's color balance shifts away from 3200K (at least upwards), the more purple the blues get.
Click to expand...

That would correlate somewhat with what is known in the color science community for sometime that the standard tables of colorimetric tables in blue region are not accurate. 3200k has little blue. Other temperatures have more blue. That is why color spaces derived from standard colorimetric tables are severely twisted in blue-purple region. A frequent problem seen is that even if one moves along the constant chroma vector blue and purple change between each other.

It is also a fact that several related important experiments by one of most famous color scientists, David McAdam (Kodak) who had amazing contributions to color science, were done by just him and perhaps another person.

Several spaces were invented to solve this (and other) issue(s) including CIE94, and DE2000 for small distance color calculation. Such notions are based upon correlations of color values with each other. However, trying to improve upon something which is twisted perhaps due to incorrect data in the first place, has resulted in horrendously complex distance formulae in color spaces -- so much so that several publications were written just to address the frequent mistakes people were making in implementing the distance forumalae.
 
You must log in or register to reply here.
Back
Top