Weird Gamma Issue

TimTyler wrote on 11/5/2007, 10:09 AM

Here's a PNG output from the Vegas 8a Preview window.

Everything was shot 30p with matching HVX200's, but the first five images were shot to tape and the bottom right image was shot to P2. The DV tapes were dumped to disk using Vegas' vidcap60.exe.

If you open the PNG in photoshop you can see that the darkest black areas in the taped images go down to the low 20's in RGB, while the P2 image goes down to 3 or 4 or 5.

Is there an explanation for this?

Comments

marts wrote on 11/5/2007, 10:55 AM
theoretically it might be related to the use of DV codec - the resulting files from tapes captured using Vegas might be recognized as a Vegas DV files and thus decoded to Studio RGB, while the P2 files will be decoded with a different codec to Computer RGB

Glenn Chan has made a nice overview on this topic http://glennchan.info/articles/vegas/v8color/v8color.htm
TimTyler wrote on 11/5/2007, 1:45 PM
I think you're right. It might just be a preview issue.
GlennChan wrote on 11/5/2007, 9:36 PM
It'll likely render like that too. You might need to add a color corrector "computer RGB to studio RGB" preset if that codec (for the P2 clips) is decoding to computer RGB.

2- In Vegas 8, the computer RGB <--> studio RGB presets don't do what they say in 1.000 compositing mode. (Anyways it's explained in my article.)
TimTyler wrote on 11/5/2007, 11:34 PM
I'd much rather have Vegas actually see and use the 0-255 color for the DV clips so I can render the real blacks and whites, not some color-corrected facimile.

Is there any way to do that?
GlennChan wrote on 11/6/2007, 12:09 AM
You could apply the "studio RGB to computer RGB" preset on all your DV clips. So you convert everything to computer RGB.

2- Converting everything to studio RGB is advantageous in some cases. e.g. if you want to pull the superwhites into legal range, you can do that with studio RGB. (Though the clips that came off the p2 cards can't do that if they decode to computer RGB right off the bat.)
TimTyler wrote on 11/6/2007, 12:47 PM
Thx, Glenn.

But since that's color correcting, is it in effect crunching the blacks instead of using the real blacks that are in fact part of the DV media?

Is it simply stretching the 16-235 luminance to fit the 1-255 Computer RGB range?

Does the true 1-15 and 235-255 luminance information exists within the DV media, and is Vegas refusing to use it since it wants DV to be Studio RGB?
GlennChan wrote on 11/6/2007, 1:28 PM
Not sure if this clarifies things but:

1- Computer RGB range is 0-255.

2- DV records to Y'CbCr color space. Many values in Y'CbCr color space cannot be represented in computer R'G'B' space... e.g. superwhites. Studio R'G'B' color space can represent more colors, including most superwhites.

3- I'm not sure what you mean by real black and real white.

4- Before outputting, you make sure that the levels get converted appropriately (if need be). So if you are working with studio RGB levels, you make sure that the black and white levels map correctly for your output format.
TimTyler wrote on 11/7/2007, 7:46 AM
> I'm not sure what you mean by real black and real white.

I mean that the cameras recorded RGB values from 0-255, and Vegas seems only able to work with values16-235 for the tapes that were imported with Vidcap.

So what happened to the video information that fit into the 0-15 and 236-255 range? Is it ignored, discarded, or squeezed into the 16-235 values?

Does the CC "Studio to RGB" preset simply stretch the 16-235 information to fit a 0-255 color space?

I'd like to work with the 255 distinct luminance values originally recorded, not 220 values that have been stretched to fill the 255 space.
GlennChan wrote on 11/7/2007, 11:00 AM
I mean that the cameras recorded RGB values from 0-255
They don't. They record Y'CbCr values. The Y' component is luma... legal range for luma is 16-235. Pretty much all cameras I've seen will record values above 235(Y')... many people call those values superwhites.

I'd like to work with the 255 distinct luminance values originally recorded
Working with studio RGB levels allows you to do that. The 16-235 Y' range gets mapped to 16-235 RGB... and the stuff from 0-15 and 236-255 Y' gets mapped to RGB similarly. So they are there.

But if you converted/mapped 16-235 Y' to 0-255 RGB right off the bat, then values outside the 16-235Y' range get clipped. Which is a problem for the P2 clips I think... since it looks like they are decoding to computer RGB.