32bit float vs 8 bit

[Joe Carney]

Still no word of 10bit support. sigh.
Everything else sounds great.

[Dan Crandall]

10-bit? ...... how about 32-bit?

-Dan

[Jon McGuffin]

Yeah, something isn't right with this "32-bit video signal path" they are talking about. This is not the same thing as true 10-bit video, I'm fairly certain of that.

Jon

[Mark Holmes]

Originally Posted by Jon McGuffin

Yeah, something isn't right with this "32-bit video signal path" they are talking about. This is not the same thing as true 10-bit video, I'm fairly certain of that.

Jon

What? What's not right about it? It's the same color space Adobe After Effects works in. It's not like it's impossible.

[Steven Thomas]

Jon,
I think it's the real thing.
http://www.videoguys.com/vegaspro8.html
and to think all I was asking for was 10bit.

"32-bit Floating Point Video Processing
Surpass traditional 10-bit standards with 32-bit floating point video processing. Take advantage of greater color range for more vivid colors, reduced gradient banding and posterization for smoother color transitions, linear light capability for optically correct compositing, and many other precision enhancements."

[Jon McGuffin]

32-bit floating point video processing is the not the same as a 10-bit colorspace. It's like comparing a 64-bit CPU to 24-bit color display on your monitor (which is 8-bit RGB, and 32-bit color is 12-bit RGB).

The words in their language is "floating point video processing". This would indicate more of a color processing performance advantage and their wording following this supports that.

The fact they are comparing their 32-bit to 10-bit "standards" has me thrown a little bit and maybe questioning my skeptisicm. Either they are pulling a Creative Labs when creative tried to sell us that their X-Fi Crystalizer was so powerfull it would increase the resolution of the original music (which is bogus) or there is something they are doing that I don't quite understand.

Jon

[Glenn Chan]

32-bit color is 12-bit RGB

Huh??? 32-bit color usually means 4 channels (RGB + alpha) with 8bits/channel.

[Jon McGuffin]

Originally Posted by Glenn Chan

Huh??? 32-bit color usually means 4 channels (RGB + alpha) with 8bits/channel.

So is Vegas 8 a '32-bit' capable application or a '12-bit' application?

Jon

[Jon McGuffin]

I'm telling you, something feels very fish about the claims Vegas is making..

To suggest that Vegas was ONLY an 8-bit color application is crazy. 8-bit color is only capable of 256 colors. Remember the old VGA standard?
See here:
http://www.cambridgeincolour.com/tut.../bit-depth.htm

To now go from 8-bit all the way to 32-bit when SGI and the like all work in a 10-bit or 12-bit environment is fishy.

I'm pretty sure what is going on here is we're talking about bit depth per color channel which means vegas is/was an application working in 24-bits 16,700,000 colorspace (8-bits X RGB or 8-bits X YUV).

Now maybe it's capable of processing colors in 32 bits which is a "fancy" way of saying it processes in the 12-bit realm.

But still... why aren't they advertising this as a "12-bit color with 32-bit processing"

The floating point remark refers to mathmatical calculations not color density...

Jon

[Glenn Chan]

Jon.... 3 x 12 = 36. Not 32.

To suggest that Vegas was ONLY an 8-bit color application is crazy. 8-bit color is only capable of 256 colors. Remember the old VGA standard?
See here:

Vegas 7 and previous versions were always capable of 8 bit / CHANNEL, not 8 bits / pixel.
8 bits / channel
is
24 bits / pixel
assuming three color channels (red, green, and blue). 8 bits/channel X 3 channels = 24 bits.

[Jon McGuffin]

Glenn,

Is that not what I said my my very next line? That we were talking bit level per channel. So what exactly does vegas mean by saying its 32-bits?

[David Newman]

When discussing computer graphics, bits-per-pixel was used as there is only one color system 4:4:4:(4) RGB. In video systems, channel depth was used as bit-per-pixel is more confusing. RGB 4:4:4 is 24-bit in computer graphics terms and 8-bit in video terms, then consider 4:2:2 which is still 8-bit in video terms, but an averages 16-bits per pixel. 4:2:0 is 12-bits-per-pixel, yet still an 8-bit video system. You can start to see why bits-per-pixel is not used in video systems, as it requires a deeper understanding of the color sub-sampling. 10-bit 4:2:2 YUV is for pro broadcast applications, with an average of 20-bits per pixel, it looks significant better and more flexible than a computer graphic 24-bit image, as it has 4 times the number of luminance levels.

[Douglas Spotted Eagle]

You'll first want to understand the difference between processing bits and color bits.

[Glenn Chan]

10-bit 4:2:2 YUV is for pro broadcast applications, with an average of 20-bits per pixel, it looks significant better and more flexible than a computer graphic 24-bit image, as it has 4 times the number of luminance levels.

Technically it doesn't. ;) Only the values 64 to 940 have legal luma levels. The values below 64 should not be reproducible (because it calls for negative light).

They also aren't luminance levels per se, since video uses luma. See
http://poynton.com/papers/YUV_and_lu...e_harmful.html
or
http://glennchan.info/articles/techn...ma/chroma1.htm
(my article has pictures and shows flaws with 4:2:2 chroma subsampling)

2- In my opinion, 8-bit R'G'B' (no subsampling) is superior to 10-bit 4:2:2 Y'CbCr as chroma subsampling can produce visible artifacts (see the pictures in my article). Here the distinction between luma and luminance is actually important. I prefer using the correct terms as the difference can be somewhat significant (assuming you are bothered by chroma subsampling artifacts).

But anyways, this is perhaps getting overly overly complicated. :D

[Mark A. Foley]

Glen,
Whats your take on the new color processing or are you under a NDA?