2 questions about 32-bit Vegas

vitalforce wrote on 9/11/2007, 10:55 AM
For the first time I am debating whether to buy a Vegas upgrade.

1. Does 32-bit processing improve (a) color correction and (b) banding around secondary color correction selections, or am I talking apples and oranges?

2. Has Sony put out a new version of Vegas in which a number of the Video FX don't work as they used to? Curves, for instance?

Comments

GlennChan wrote on 9/11/2007, 1:18 PM
1- Yes / kind of. Usually in 8-bit you don't see banding to begin with. 32-bit mode will get rid of some types of banding.

Some/?most/all? of the filters do not dither, so you can still see banding in some cases. To see this in action, add the color gradient (black to white). Apply some aggressive color curves (or levels) filter. You'll see gaps in the histogram. 32-bit mode doesn't get rid of that. If the patches of color are big enough and the source footage has no noise in it, you can see banding artifacts.

(Hmm... it looks like the gradient generator is still 8-bit.)

2- Your second question is answered in this thread:
http://www.sonycreativesoftware.com/forums/ShowMessage.asp?MessageID=546313&Replies=22
vitalforce wrote on 9/11/2007, 5:10 PM
Thanx Glenn as always.

I knew about the other thread but was concerned that Sony appeared to have put out a product knowing that the Vegas filters 'behave differently' with 32-bit processing, yet did not develop the filters accordingly so that the filters would behave as originally designed and consistently, even when 32-bit was applied. It seems astonishing to me that a new version of Vegas would now have central features no longer functioning in the way users have learned to use them, over some 5-6 years.

I have a DV feature project that I had held back from certain festival submissions thinking I might be able to enhance its look further with new V8 features, but from what I am seeing, it seems that opening such a project in Vegas Pro 8 would result in many of my filters, especially as relates to color correction, levels, curves, luminance and secondary correction, not behaving in the same way over many hundreds of timeline events?

Well, maybe I'm overthinking this. But a good white paper would certainly seem to be necessary here.
.
GlennChan wrote on 9/11/2007, 9:45 PM
Use 32-bit / 2.222 compositing gamma. It's the old school filter behaviour. (Though certain bugs/deficiencies in the filters may be fixed in V8.)
vitalforce wrote on 9/11/2007, 9:53 PM
I think I get it now. Reread the link you referred me to and well--time to order that upgrade....