Nice results! I like how the dark shot of the fighting ring got adapted to the DVD look. The DVD having crushed shadow detail in that shot but the corrected shot has the same colors but without the crushed blacks.

You_Too said: Nice results!

Yeah, I thought so.  Given that the Criterion DVD benefited from extensive color correction by Robert Harris, I thought this might be one instance in which the script might struggle; but, skipping through the movie, I'm amazed at how good the function's adjustments are. 

You_Too said:  I like how the dark shot of the fighting ring got adapted to the DVD look. The DVD having crushed shadow detail in that shot but the corrected shot has the same colors but without the crushed blacks.

I was looking at just the color, but now that you mention it the blacks do look a bit crushed on the DVD.

thanks for sharing Chewtobacca!

-G

Thanks, but the script was more forgiving to Chewtobacca's sources than some other color mismatches I've seen. I guess this technique (matching averages) works better when the colors haven't been fucked with too much between the sources.

-G

g-force said: I'll make an updated version after Christmas that has all the options of both scripts, and maybe a YUV version as well.

Hey, g-force.  Are you still planning to work on a YUV version?  It would be really useful. 

By the way, does the option to determine the correction from two different sources and then apply it to a third get round the numerous colorspace conversions in the script?

You mentioned adding the option to balance the black and white levels to an updated version of the script.  Would it be possible to write a script that matches only black and white levels i.e. makes no color changes?  Could this be used to correct a source with a brightness-and-contrast problem by using another source or does it not work like that?

Chewtobacca said:

g-force said: I'll make an updated version after Christmas that has all the options of both scripts, and maybe a YUV version as well.

Hey, g-force.  Are you still planning to work on a YUV version?  It would be really useful. 

By the way, does the option to determine the correction from two different sources and then apply it to a third get round the numerous colorspace conversions in the script?

You mentioned adding the option to balance the black and white levels to an updated version of the script.  Would it be possible to write a script that matches only black and white levels i.e. makes no color changes?  Could this be used to correct a source with a brightness-and-contrast problem by using another source or does it not work like that?

Dude, didn't see this until AntcuFaalb pointed it out. Sorry, I've been busy, and I still owe you a response to your e-mail. All these requests are definitely possible, but I just have to find the time to get it done.

-G

AntcuFaalb said:

g-force: Based on your experience, do you suspect that a YUV version would match better than an RGB version?

 No, as a matter of fact, that's how this all started. Do a google search for YcUrVematch ;)

g-force said:  Sorry, I've been busy

No apology is necessary.  I know the feeling!

The real reason I asked about the YUV version is to circumvent the colorspace conversions.  Going to RGB and back once is fine, but the multiple conversions still worry me a bit, which is why I wondered if the existing option to apply the difference between two clips to a third would be one solution.

It's a shame Avisynth is 8-bit only.  A 32-bit workflow would help here.

By the way, I don't think I've sent you any e-mails that you haven't replied to.  :-)

oops, the e-mail request was actually from AntcuFaalb.

Could someone tell me if this the correct way to apply the difference between two clips to a third?

a=Avisource()

b=Avisource()

c=Avisource()

RGBMatch(b,a,c)

Just checking...

Chewtobacca said:

Could someone tell me if this the correct way to apply the difference between two clips to a third?

a=Avisource()

b=Avisource()

c=Avisource()

RGBMatch(b,a,c)

Just checking...

In your case (B,A,C), B would be your color reference, A would be the clip it would try to colormatch to B and it would then apply those color changes to C.