8bit vs 32bit Settings

[Nicholas de Kock]

I've been editing in Vegas 11 in 32-bit floating point (video levels) space under project settings for some time now thinking it's a better workflow however the timeline has been slow & I've struggled with color shifting issues on 720P footage. Today out of frustration I went into all my settings in an effort to figure out why I'm struggling so much and went to 8bit (at this point I forgot I was in 32bit) - all my problems gone and smooth playback at Best (Full) with effects applied - I'm such a numbnut. I've been struggling for months, even render times are blistering fast now.

[Jeff Harper]

Nick, there are reasons people use 32 bit, I don't remember what they are, but I tried it a few years ago, and remember the images were much different, I believe they may have been much more contrasty, among other things, I didn't care for it. From what I remember it is better technically, but you have to learn how to adjust your CC and stuff and it's a whole other kind of deal.

I might need to re-visit it, but I'm happy using 8 bit. I'm a rather lazy editor, I suppose, and am content with 8 bit.

[Adam Stanislav]

Originally Posted by Jeff Harper

I don't remember what they are

One reason to use 32 bits is to avoid clipping. If one effect extends the value outside the 0-1 range, another effect may bring it back into the range. So, if one effect doubles the values of the pixels, all 8-bit values greater than 127 will result in a value greater than 255 but will be clipped to 255. And if the next effect in the chain divides everything by 2, you will only get results between 0 and 127, while with the 32-bit math you would get the original values. Now, naturally, you would not normally multiply everything by 2 just to divide it by 2 later on, but that is the type of error eight bits will cause.

The other reason is to avoid round-off errors. In eight bits each effect has to round its result to an integer. The next effect’s starting value is that integer and it can produce a different (and less desirable) result than it would without the rounding off.

For example, if an effect just multiplies everything by 1.1, then darker values will not see the effect at all (e.g., 1*1.1 = 1, 3*1.1 = 3), so only brighter values are affected. And the error gets compounded if you have a chain of effects. On the other hand, in 32 bits no such error propagates through the chain, and only the final result of the entire chain is rounded off (up or down respectively).

[Ian Stark]

Excellent explanation Adam. Thanks.

[Peter Riding]

I've also recently gone to 8 bit and seen a tremendous imporovement in both playback and rendering times on my already highly spec'd main desktop.

As I understand it you can switch between 8 bit and 32 bit at will without adverse effects. So for example if you have some tricky colour grading to do you might switch the project properties from 8 bit to 32 bit then back again to 8 bit to continue getting fast playback at Best settings. And switch your 8 bit project to 32 bit for final rendering if the editing has been particularly heavy duty.

Coming from a pro-photography background I had always taken it for granted that 32 bit was the place to be. But there again I know of very few photographers who would work on their stills in 16 bit as a default, they'd reserve that for problem images.

Pete

[Ray Turcotte]

A while back I did an video edit where I mixed in some still pictures from my DSLR into the timeline. That project was set to 8bit, but the transitions and CC on the photos were horrible. I switched to 32bit project settings and the problem was fixed. That was with VPRO 11 v411, don't know if it was a bug or not in Vegas

[Peter Riding]

Ray, I use still images from Canon 5DII's. in virtually every project and I'm not seeing that. But I do prepare them in Photoshop first so Vegas doesn't have to do anything other than handle the transitions.

Pete