yingcheng007
September 28th, 2005, 03:26 AM
Hi all.
I have a question about ColorSpace node in Shake.
When I use Shake to do color space conversion , I find a problem .
This is what I do.
1) Create a Ramp image in shake . From left to right is red (R=1,G=0,B=0) to Black (R=0,G=0,B=0) .
2) Use Pixel Analyzer to analyse the pixels from left to right . I only find the R value is changing . The G value and the B value are always 0 .
3) Use a ColorSpace node to change the RGB space to YUV color space .
4) Add a ColorSpace node again and change the YUV space to RGB space .
5) Now use Pixel Analyzer to analyse the pixels from left to right again . I find the G value changes a lot ( Max to 0.0667 , about 6.67% ) , the B value is always 0 .
I know that color space change causes error . But I do the same thing in After Effects , the error is more smaller than in Shake . In AE , the B value max to 0.0039 ( about 0.39% ) and G value is always 0 . I also do it in Nuke , the error is very little too.
All the testing above are tested in both 8bit and 16bit .
Question : Dose Shake’s mathematics calculation causes the larger error in color space change ?
Can I use any node to decrease the error ?
I have a question about ColorSpace node in Shake.
When I use Shake to do color space conversion , I find a problem .
This is what I do.
1) Create a Ramp image in shake . From left to right is red (R=1,G=0,B=0) to Black (R=0,G=0,B=0) .
2) Use Pixel Analyzer to analyse the pixels from left to right . I only find the R value is changing . The G value and the B value are always 0 .
3) Use a ColorSpace node to change the RGB space to YUV color space .
4) Add a ColorSpace node again and change the YUV space to RGB space .
5) Now use Pixel Analyzer to analyse the pixels from left to right again . I find the G value changes a lot ( Max to 0.0667 , about 6.67% ) , the B value is always 0 .
I know that color space change causes error . But I do the same thing in After Effects , the error is more smaller than in Shake . In AE , the B value max to 0.0039 ( about 0.39% ) and G value is always 0 . I also do it in Nuke , the error is very little too.
All the testing above are tested in both 8bit and 16bit .
Question : Dose Shake’s mathematics calculation causes the larger error in color space change ?
Can I use any node to decrease the error ?