FORUMS: list search recent posts

What is the difference between 32-bit (video levels) and 32-bit (full range)?

COW Forums : VEGAS Pro

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Levan Katsadze
What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 24, 2014 at 10:31:49 pm

Hello, in Sony Vegas Pro 13 Project Properties window, there are three options about pixel format.


1) 8-bit
2) 32-bit floating point (video levels)
3) 32-bit floating point (full range)



Well, as I know, 8-bit and 32-bit are basically number of colors, in this way:

8-bit color = 256 colors (2^8)
32-bit color = 4,294,967,296 colors (2^32)

But the Pixel Format is more complicated thing then just number of colors. Well, more bit produces more color precision, but I have a question, what is the difference between these two options:

32-bit floating point (video levels)
32-bit floating point (full range)

What do "video levels" and "full range" mean? Also I know that some effects/transitions/media generators don't support 32-bit format. So, I just can't use them when I am working in 32-bit format, right?

I know that using just 8-bit is very good and quite enough for general video production, and I know that 32-bit format is a headache for non-professional user, But I have a question, in what situations 32-bit format is better than 8-bit? For example, I know that 32-bit supports Alpha (transparent) channel rendering, and 8-bit doesn't, is that right?

Thank you.......


Return to posts index

Norman Black
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 24, 2014 at 10:46:10 pm
Last Edited By Norman Black on Sep 24, 2014 at 10:50:27 pm

8-bit color is 16 million colors. It is 8-bits per channel, RGB.

Video levels versus full range...Due to the history of analog video signals digital maps something compatible.

So in 8-bit, Studio RGB (aka video levels) is 16-235. Full range is 0-255.

32-bit floating point maps similarly. Video levels being a slightly smaller range than full.

Unless your camera captures 10-bit or more color, then the 32-bit levels are of no use. Nearly every camera out there is capturing 8-bit color. It is really mostly "pro" cameras that capture higher color bit depth.

8-bit does support Alpha channel rendering. Only a very few codecs support rendering the Alpha channel.

Usually the best bet for project settings is to have Vegas match your media settings. It defaults to asking this.


Return to posts index

John Rofrano
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 12:00:47 am

[Levan Katsadze] "For example, I know that 32-bit supports Alpha (transparent) channel rendering, and 8-bit doesn't, is that right?"
You are really confusing two different things. 32-bit pixel format has nothing to do with 32-bit color. Is has to do with the internal math used to calculate pixel values. In 8-bit mode Vegas uses 8-bit floating point precision in pixel calculations, in 32-bit mode Vegas uses 32-bit floating point precision to calculate pixel values. This has nothing to do with how many bits are needed to represent alpha channels.

As Norman said, if you are using footage from a camera that shoots 10-bit color you can benefit from using 32-bit floating point math when calculating pixel values. Otherwise it's a waste and can actually cause problems.

~jr

http://www.johnrofrano.com
http://www.vasst.com



Return to posts index


Norman Black
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 2:01:39 am

[John Rofrano] "As Norman said, if you are using footage from a camera that shoots 10-bit color you can benefit from using 32-bit floating point math when calculating pixel values. Otherwise it's a waste and can actually cause problems."

Even with a high bit camera source, one would likely still use 8-bit mode to edit/cut the video and only use 32-bit mode to color grade and final render.

Lets not talk about the $$$ for a video display setup that will actually display high bit color.


Return to posts index

Levan Katsadze
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 12:13:52 pm

Thank you for the reply, well now, as I understand:
.
.
.
.
.
8-bit pixel format uses 32-bit color depth (24bit RGB + 8bit Alpha = 32bit color depth)
...................... so, it is 16,777,216 colors + Alpha
...................... and the range is 16-235


32-bit (video levels) pixel format uses more than 32bit color depth (and please tell my how many?)
..................... and the range is also 16-235


32-bit (full range) pixel format uses the same color depth as 32-bit (video levels) pixel format.
..................... and the range is 0-255
.
.
.
.
.
Please correct me if I am wrong in anything. Sorry, maybe I am a total ignorant about understanding these things.


Return to posts index

John Rofrano
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 1:04:20 pm

[Levan Katsadze] "Please correct me if I am wrong in anything. Sorry, maybe I am a total ignorant about understanding these things."
I believe you are still a bit confused (no pun intended). The color depth is controlled by the source video. 8-bit video will only ever have 8-bit color depth regardless if you are using an 8-bit pixel format or a 32-bit pixel format in Vegas Pro. 10-bit video will only ever have 10-bit color depth regardless if you are using an 8-bit pixel format or a 32-bit pixel format in Vegas Pro. The pixel format has nothing to do with the color depth of the video.

The pixel format has to do with the precision of the math used to calculate color values. If you are using 10-bit source, and you are calculating with 8-bit precession, you may loose information due to rounding errors. By using 32-bit precision, you will not have any rounding errors when working with 10-bit color.

Said another way, 32-bit pixel format gives you greater precision to represent the color in your video when using formats that have a color depth that is greater than 8-bit (i.e., 10-bit, 12-bit, etc.). It does not use more color depth than the original video already has.

Video levels vs full range has to do with whether the internal 32-bit math calculations will contain the values produced by legal video levels or allow the full luminance range to be used. In other words it just constrains the min/max values used in calculations.

The bottom line is if you are working with 8-bit video there is no need for you to understand any of this. Just set your pixel format to 8-bit which is the default (for good reason).

~jr

http://www.johnrofrano.com
http://www.vasst.com



Return to posts index


Levan Katsadze
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 1:56:21 pm
Last Edited By Levan Katsadze on Sep 25, 2014 at 2:07:17 pm

[John Rofrano] "The color depth is controlled by the source video"

[John Rofrano] "The pixel format has nothing to do with the color depth of the video"

[John Rofrano] "The pixel format has to do with the precision of the math used to calculate color values."

[John Rofrano] "32-bit pixel format gives you greater precision to represent the color in your video"

[John Rofrano] "It does not use more color depth than the original video already has."

Thank you so much, now I really understand what does 32-bit pixel format mean. And also, recently I saw these links and they also helped me so much:

Understanding 32-bit format
http://blogs.adobe.com/VideoRoad/2010/06/understanding_color_processing.htm...

from Lynda.com
http://www.lynda.com/home/Player.aspx?lpk4=30903
.
.
.
These articles also helped me so much:
.
http://www.glennchan.info/articles/vegas/colorspaces/colorspaces.html
.
http://www.glennchan.info/articles/vegas/v8color/vegas-9-levels.htm
.
.
.
And, well, about "video levels vs full range", well as I understand, "full range" has range advantage (0-255 vs 16-235) than "video levels", so full range is better, right? because, as I know, "video levels" is old standard (Studio RGB) for TVs, and "full range" is new standard for computers (Computer RGB).
.
.
.


Return to posts index

Norman Black
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 3:03:58 pm

[Levan Katsadze] "8-bit pixel format uses 32-bit color depth (24bit RGB + 8bit Alpha = 32bit color depth)
...................... so, it is 16,777,216 colors + Alpha
...................... and the range is 16-235"


The range of 8-bit is always full range. It is up to you to maintain studio range, 16-235, if that is your target. This is one of the quirks of Vegas.

After a lot of learning pains, I work/edit in full range and apply a levels effects, Computer to Studio when I render.

[Levan Katsadze] "32-bit (video levels) pixel format uses more than 32bit color depth (and please tell my how many?)
..................... and the range is also 16-235


32-bit (full range) pixel format uses the same color depth as 32-bit (video levels) pixel format.
..................... and the range is 0-255"


The color depth of 32-bit floating is not precisely defined. Its precision is way more than enough to handle any bit depth of camera input anyone could ever dream up.

The "range" of 32-bit in Vegas is either video or full range.

[Levan Katsadze] "And, well, about "video levels vs full range", well as I understand, "full range" has range advantage (0-255 vs 16-235) than "video levels", so full range is better, right? because, as I know, "video levels" is old standard (Studio RGB) for TVs, and "full range" is new standard for computers (Computer RGB)."

I guess one can say full range is "better" (your words). The problem is that most video playback, including on computers, expects video levels. Therefore if you give it full range your video will not look correct on playback. Your darks and lights will be clipped and the video will have more contrast.

Cameras in the video world might capture beyond video levels and it is up to the editor to bring these within the legal range. Typically using the video scopes.


Return to posts index


John Rofrano
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 25, 2014 at 8:01:49 pm

[Levan Katsadze] "And, well, about "video levels vs full range", well as I understand, "full range" has range advantage (0-255 vs 16-235) than "video levels", so full range is better, right? because, as I know, "video levels" is old standard (Studio RGB) for TVs, and "full range" is new standard for computers (Computer RGB)."
Yea, as Norman pointed out, it depends on how you plan to deliver your video. If you are planning to deliver this video to YouTube, YouTube (and most other video sharing sites) expect to see Studio RGB. So while full range may be ideal if you only plan to watch it on your personal computer or you are preparing if for digital signage, it may not be idea to deliver to video hosting web sites.

BTW, 0-255 assumes 8-bit video. If you use 10-bit video the RGB values are 0-1023. ;-)

~jr

http://www.johnrofrano.com
http://www.vasst.com



Return to posts index


Levan Katsadze
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 26, 2014 at 4:53:05 pm
Last Edited By Levan Katsadze on Sep 26, 2014 at 5:12:07 pm

Thanks, well, if I want to use 32-bit pixel format and also want to watch the final rendered video on my computer, not on my TV, then what options should be best for this situation?

Are these the best options?

Pixel format: 32-bit floating point (full range)
Compositing gamma: 1,000 (Linear)
View transform: Off

Well, but as I see, if the "View transform" option is set to "Off", then compositing gamma 1,000 and 2,2222 just looks the same. Actually, in video preview window, the change is not visible for my eyes, but very small change is visible in RGB Parade. So, I am confused, which option (1,000 or 2,222) should be better for my situation?


And also, please tell me, why "View transform" options are needed? I saw this help page, but I still don't understand, why I should not set it to "Off"?






And, also, if the video is darken and clipped by these options, then it's not a problem, right? I can use Video levels FX or brightness or any other FX and correct the playbck.


Return to posts index

Norman Black
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 26, 2014 at 6:43:54 pm

[Levan Katsadze] "Thanks, well, if I want to use 32-bit pixel format and also want to watch the final rendered video on my computer, not on my TV, then what options should be best for this situation?"

If you watch your final rendered video on TV or computer or basically anywhere, it will be in an 8-bit video format. Your PC video card and monitor are probably incapable of 10+ bit display. Your camera(s) is probably 8-bit.

So what is the fascination with putting 32-bit floating mode in between your camera source and final output?


Return to posts index

Levan Katsadze
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 26, 2014 at 7:47:44 pm
Last Edited By Levan Katsadze on Sep 26, 2014 at 7:50:44 pm

[Norman Black] "If you watch your final rendered video on TV or computer or basically anywhere, it will be in an 8-bit video format. Your PC video card and monitor are probably incapable of 10+ bit display."

Yes, maybe the final render will be in 8-bit format, but this video shows that even if the final is in 8-bit codec, the 32-bit editing gives us much more precise and much better results.

http://www.lynda.com/home/Player.aspx?lpk4=30903

.
.
.


Return to posts index


Norman Black
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 26, 2014 at 7:59:21 pm
Last Edited By Norman Black on Sep 26, 2014 at 8:07:47 pm

Yes, I watched that video. Have fun with your 32-bit.

It has its applications, but they are very few without high bit or HDR input. Unless you do something exactly like the second example, what are you getting out of 32-bit. YOu actually can get some of that in Vegas without 32-bit. In 8-bit Vegas allows video values outside the "legal" video levels range, aka full range, and you can bring them back into that video levels range.

Sophisticated compositing, of certain things, is an application for 32-bit. Is that what you are going to do in Vegas?


Return to posts index

Levan Katsadze
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Sep 26, 2014 at 9:12:55 pm

[Norman Black] "Sophisticated compositing, of certain things, is an application for 32-bit. Is that what you are going to do in Vegas?"

Well, why not? maybe not now, but actually I want to learn things deeply, because I love video editing and generally I love digital industry. Well, maybe it's better to go to special school of video production, but in my country, there is no such thing. My country, Georgia, actually has just started making steps into digital industry. So, internet is my best friend when I sit on the computer.


Return to posts index

Levan Katsadze
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Nov 20, 2014 at 10:52:15 am

And, I have just one more question. It is actually technical, and maybe not important, but please tell me, why most of the TVs and many display monitors doesn't support 0-15 and 235-255 video playback? Why they support only 16-235 ??? Why? what is the point of it?
.
What is the point of any technical limitation?

I understand, that if technology is not advanced enough to make something, then it is not made, but I know that technology is already advanced and manufacturers can make displays, that support full video levels, so I don't understand, why they don't make it? Why they limit it? What's the point if it?

.
.
.


Return to posts index



John Rofrano
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Nov 20, 2014 at 11:55:07 am

[Levan Katsadze] "I understand, that if technology is not advanced enough to make something, then it is not made, but I know that technology is already advanced and manufacturers can make displays, that support full video levels, so I don't understand, why they don't make it? Why they limit it? What's the point if it?"
The point is backward compatibility. The TV specification hasn't changed much since the 1920's when it was invented. Color TV was contrained by what Black & White TV's could produce so that people with B&W TV's could watch color broadcasts too. That's the same reason all broadcast is still interlaced even though most TV's don't require it. Backward compatibility has held the entire broadcast industry back for almost 100 years now.

~jr

http://www.johnrofrano.com
http://www.vasst.com



Return to posts index

Alex Ka
Re: What is the difference between 32-bit (video levels) and 32-bit (full range)?
on Apr 4, 2017 at 7:53:05 pm

Hey Levan,

I felt more confusion reading the ensuing discussion vs. your straightforward question.

The short answer, good for 95% of cases, is: 32-bit floating point (video levels).

Why not 8 bit:
  • Quality wise, much less risk of banding as a result of color manipulation.
  • Speed penalty wise, it's 2017 (but has been true for years): rendering is done via GPU (and even if CPU, those are fast).

Why not 32 bit full range:
Lots of legacy reasons. Many DSLRs and recorders will even have the vids tagged full range in the container, but the data is video levels. How can you check: in Vegas top menu, pick View -> Window -> Video Scopes. Select Histogram. It'll chart the output brightness levels in the current frame from darkest (left) to brightest (right). For a normal contrast scene (not dominated by highlights or shadows), switch between video levels and full range and observe the histogram. If it's completely pegged at both ends, you've got the wrong 32 bit mode.

What does the 95% cover:
DSLR or digital camcorder originated video, edited and played on a computer or mobile device or a flat panel tv or a digital projector (directly or via youtube/vimeo

What the remaining 5% might cover:

HDR video (likely full range, with ACES RRT sRGB, but again, check the histogram); raw video (Red, CineAlta and the like); analog tape, CRTs and film.

Al
P.S. If you hear you should not go somewhere, or concern yourself with something, that's your hint exactly where to try to go and what to want to investigate further. :-)


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]