Bit-rate questions

U

USLatin

Panamaniac
I am confused and would appreciate a little light being shed one my limited understanding of NLE's color bit-rates.

First of all... up to Vegas 7 Sony was in the 10-bit world right? I forget... for a second I was thinking it was 8-bit.

Secondly, does the HVX record in 8-bit?

If so... 8-bit is 256 possible values for each of R G B right? Is that only the Chroma value and then there is an added Luma value per pixel in the footage? Wouldn't 0 0 0 be black and 256 256 256 be white? And wouldn't grays be also possible with equal values of RGB such as 100 100 100 and such...?

I understand that the added bit-rate during CC can help if you are applying multiple color corrections so that there is less degradation from the "imperfections" of of these processes. So is it that during the rendering process the higher bit-rate is used, however when rendering for say DVDs the file being rendered ends up again being three 256 values? Right?

What's the bit-rate for DVDs? Is it the same for Blue Ray? What about the HVX?



I am sorry about my lack of knowledge if there is some horrible aberrations above this disclaimer guys :D I am still trying to pick up these things that most of you have known and worked with for a long time.
 
Hi,

Color depth and bit depth aren't necessarily the same thing, although they have a lot to do with each other. You can't have one without the other, as it were.

Eight-bit color has 2^8*2^8*2^8 colors (16,7million more or less) just like jpeg images. They share the same color depth. In NTSC land, we're ham strung with less color in the IRE range in the upper and lower scale, too. Our blacks aren't as black and our whites aren't as white, a safe limit could be expressed in 15-15-15 to 235-235-235, meaning that we waste a lot of color on both ends of the color range.

Pity ain't it?

Possibly. Until the last 10 or so years with the stuff the American consumer has been getting from the shelf, it likely hasn't mattered much. TV sets have improved greatly in the last few years.

Internal processing - bit depth that you are really referring to in both camera operations and in software processing is where great things happen.

The HVX200 is 14-Processing and even greater internal processing - 18 bit? 14-bit is crazy. That's like 4 quadrillion spaces for processing - at 18 bits I don't even know the name of the number, but it does more than provide operations for color adjustments.

Speaking of, there are two things that make color adjustments cool and smooth, one is the color space that you are using HVX200's 4:2:2 is why I'm here. DV's 4:1:1 is poo, and two the amount of data that makes up the color is how much you can 'round off' when you have the data to do it. IOW, the more bit depth you have, the more you can push it around. The mere fact that there is so much data there are so few rounding errors on the colors that you can have great color correction to your images.

DVDs are in 4:2:0 color space. Remember. Once the color is processed from a space it is then placed into another place. Give it your best colors in your best place and it should be fine in its new place.

My best.
 
nice post and I thank you for it!

ok... so from what I was able to gather from your post, our camera is processing at 14 or 18 bit but what's the bit rate of the DVCPRO HD footage it outputs?

I am aware of colorspace and understand how the averaging out of pixels works but there is still a specific bit-rate obviously for the values of each of these pixels to be expressed right? Whether a group of pixels share their R G or B value for compression the pixels still have an assigned value... am I on page or did I mess something up on this paragraph?

Also, you said: "the more color depth you have the more you can push it around", does that mean that the more colors you have the more you'd he able to bring about more details in the shadows and such since two pixels that may have both been 1 1 and 1 in 8-bit, could now different and be any given value from 0 0 0 to 16777216 16777216 16777216 in 32-bit, so when you mess with the gamma and accentuate the spread between these?

But that would only be so if your footage is recorded at a higher bit-rate...

So for those not able or planning on a camera upgrade yet, the main benefit is the ability to color correct a particular clip time after time with several layers of effects without the dangers of rounding off degradation, right? I am not sure I understood what you meant by the sentence starting with "Internal processing". How much does this provide for a skilled user? And how is the "benefit" achieved?
 
ok... :tongue:

I couldn't remember what my third question was gonna be for that last post.... and that was it... what exactly IS bit/color depth?

is it achieving a nice gamut of many different reds making up a red pillow as opposed to just a few shades of it? but does it deal with post only or also lighting?
 
Hi,

The rate and depth are vastly different and I didn't address it at all - so sorry!

DV's rate and HDV's rate are about the same - 25mbs. DVCPRo HD is 100mbs. But I really think you really know that.

That rate matters, but the rate, while it matters, has to do with things other than NLE processing. More is mostly better, most of the time, until it has to be delivered. Then it has to be thinned out to the customer's needs. If the customer can only handle a DVD, then that what he has to have - no point in giving him a thicker slab of data he can't use.

Think of internal processing as intelligent workings. All of those processes to include digital imaging and reporting of the imaging from CCDs to the recorders have to go through a lot of math and a lot of rounding. Less rounding errors, more pleasing results. Higher math functions, more horse power, just better stuff.

You also get extended dynamic range with the additional bits, too. More shadow detail. I don't know when or really 'if', but I suspect that some day we'll sort of shoot video like 'RAW' to some degree. Basically process the video in post. Cool idea, anyway.

How much is really practical is tough to call. The human eye can distinguish about 3 billion hues if I recall it right. And most broadcast sets won't get that close to that. And frankly, I really doubt that any customer will notice.
Drogar-Happy%28DBG%29.gif
 
The bit depth is how many increments per "bit" you have.

Imagine a pixel is a "bit," and you had only two choices -- 0 or 1. If the value for a pixel were 0, then it would be off. If it were 1, then it would be on. That's the basics of the most rudimentary graphics, like Pong. No color, no brightness, etc..

But then, imagine for each pixel, various levels of brightness. Every time you add a "bit," you double the number of increments you can have, so you don't just have on or off, you can have off, dark, not so dark, medium, sort of bright, getting bright, and brightest.

And so on, adding bits as you go.

That's bit depth. A very basic explanation, but the general idea.
 
ok so from your explanation David I guess color/bit-depth is the term we use to speak of the ability display a small or large range of color and brightness, while bit-rate is used to refer to the space it will require in the recording or while processing... did I get it?
 
question... will a project CC'ed in V7 need adjustments when dropped into V8? will it automatically get a bit better from the CC rounding imperfections dissapearing?

Or will the added processing bitrate allow for the curves to be pushed a tiny bit further and I'll be able to get better results but after re-doing it all?

Or will both of those things happen?



oh and BTW thanks again David and Mike!
 
Hi,

The additional color correction will only help if the CODEC used can take advantage of the bit depth.

So the answer is maybe.

The real deal is that it has the juice to do it. If the CODEC will allow it and Neo from Cineform or Raylight or something better comes along in 14-bit to make all the 4:2:2 imaging of HVX's image shine, fantastic.
 
USLatinContent said:
ok so from your explanation David I guess color/bit-depth is the term we use to speak of the ability display a small or large range of color and brightness, while bit-rate is used to refer to the space it will require in the recording or while processing... did I get it?
Click to expand...

The bit depth is the number of steps of variation you can have from max to min. (8-bit color = 256 steps for each primary color -- 0 to 255 for R, G, B.)

The bit rate is the number of bits of information per second necessary to . . . do whatever it is the file is supposed to do. So yes, the more information stored for each second, the bigger the file size.
 
USLatinContent said:
question... will a project CC'ed in V7 need adjustments when dropped into V8? will it automatically get a bit better from the CC rounding imperfections dissapearing?
Click to expand...

There will be no automatic "boost" of any kind. Only the level of control over the color information which already exists -- but no new information will be added. DV will still be 4:1:1.
 
David Jimerson said:
There will be no automatic "boost" of any kind. Only the level of control over the color information which already exists -- but no new information will be added. DV will still be 4:1:1.
Click to expand...

Hi,

DV is DV and 4:1:1 is what it is...

But... that internal processing is internal processing in software just like what is happening in cameras. You can't get better than white or black, but you can get a better quality of mid ranges in your color correction or better yet in your stair stepping in color keying.

Grass Valley, nee Canopus uses a 4:2:2 color space for DV instead of 4:1:1 for DV internally in its software and hardware, Matrox uses 4:4:4 its hardware for DV. Their original source content doesn't get better than the pixels that are brought to the table - that can't happen - but some of the correction, the keying, might be better after the fact using those tools and others similar to those tools.
 
MikeGunter said:
Hi,

DV is DV and 4:1:1 is what it is...

But... that internal processing is internal processing in software just like what is happening in cameras. You can't get better than white or black, but you can get a better quality of mid ranges in your color correction or better yet in your stair stepping in color keying.

Grass Valley, nee Canopus uses a 4:2:2 color space for DV instead of 4:1:1 for DV internally in its software and hardware, Matrox uses 4:4:4 its hardware for DV. Their original source content doesn't get better than the pixels that are brought to the table - that can't happen - but some of the correction, the keying, might be better after the fact using those tools and others similar to those tools.
Click to expand...

I know -- I said that. You can have better control over the color which exists, but you can't have more color INFORMATION than that which was already there. So, simply bringing it into the timeline without doing anything else doesn't get you anything extra, but when you manipulate it, you have finer control.
 
yea I thought so but I had to ask just to be sure I understood. Now even if you are going to output a 4:1:1 and it won't matter all that much that you have 4:2:2 and working on in in 32-bit, I still thought there may be there would be a little bump in the outcome when you are applying several effects... since you get rid of the rounding errors. Right... a small one I am sure, but you would not have to worry about doing too much to it any more... sorta like when I know that my ADR is 24/96 so I don't worry about messing with is several times with volume pitch reverb and so on to make it match... but do you guys this would be at all noticeable in a 4:1:1 output?

On a side note... I am CC'ing my project still and I am using Magic Bullet and Vegas 7... for render speed I started doing my gamma curves with the Vegas "Color Corrector" and then also used the offset... with that combo I found I could bring out more detail and still keep my (for lack of the proper word) luminosity... hehe... you know keep it as bright as it was... hehe... then after this thread I thought I should try to do the same but with Magic Bullet... even if it makes my render longer... but I can't find a way to offset the image... would I have to use a shade of neutral gray? that should do the same right? right now I am using the curve in MB and the offset from Vegas... I am also applying a bit of Vegas' unsharpening filter to sharpen the image as it needs it.

Should I forget about MB for now and just stick to Vegas?

I have After Effects here waiting to be installed too... should I try to get it going? I have so mail it in by not this but next weekend...

I might just get rid of AE and go for V8 since titles should also be better in V8 than in V7... so CC and titles being good in Vegas is there any reason to keep AE?
 
Well, you don't get rid of the rounding errors. That's impossible with digital. You just make them much smaller.
 
yea, of course... but what do you say David?

should I try to pick up AE quick or just let this one go with a MB-V7 combo for the CC?

and it MB-V7 is the right way to go with this deadline of about 1 weeks then should I try to do it mostly in MB? or don't even bother with that if submitting an SD DVD?


is Blue Ray 4:2:2? I guess it has to be, well no but it kinda does.. is it?
 
I don't think AE adds any value for color correction.

The color space of Blu-Ray depends on how you code. You have a lot of options, unlike standard DVD.
 
You must log in or register to reply here.
Back
Top