FORUMS: list search recent posts

mindlessly obsessed with 16-bit

COW Forums : Autodesk Maya

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
SSGGMM
mindlessly obsessed with 16-bit
on Oct 16, 2006 at 11:27:25 pm

I gotta quick question I may already know the answer to. Is there any possibility, ANY possibility at all, that mental ray or renderman (for maya) will yield a different result for an 8-bit TIFF colour map than it will for a 16-bit map? I'm not talking about render time, I'm talking about... wait for it... resolution.

I've read tons of articles stating the difference as being ZERO. As in, there's no difference whatsoever. 16-bit is only meant for large prints, or something. But isn't rendering kinda like printing? I can't seem to wrap my little brain around this. I got this nagging voice in my head: "Think of the size of the screen at the local film fest. Don't risk it. Go with 16-bit." So I did. And I STILL do.

I really need someone to set me straight here. I can hear my RAM weeping at night from the abuse.

Many thanks,
Serge


Return to posts index

Jason Brummett
Re: mindlessly obsessed with 16-bit
on Oct 17, 2006 at 2:11:25 am

16 bit has nothing to do with print size or as your insinuating ~ resolution.

8 bit means that the file has an R G B color spectrum with 255 different colors to represent.

16 bit means that there are (Don't know for sure); a larger color space to work with.

This really only aparent for compositing tasks where a larger color space e.g. 16 bit vs. 8 bit (Color resolution) will allow a compositor to manipulate that image more than an 8 bit. For 99% of compositing tasks though 8 bit is fine.

Resolution is how many pixels horizontally and vertically are in an image, eg. 320x240, 640x480, 1920x1080 etc.

Only 1 or 2 rendering passes would ever need to be 16 bit and that is for specific tasks.

Jason Brummett
Studio Z Inc.


Return to posts index

SSGGMM
Re: mindlessly obsessed with 16-bit
on Oct 17, 2006 at 2:57:34 am

Yes, of course. I'm in the bad habit of using the word "resolution" to mean "quality". Even that's too vague. I'll get my lingo down eventually. Newbie.

"Only 1 or 2 rendering passes would ever need to be 16 bit and that is for specific tasks."

Can you think of an example of such a task?

S.


Return to posts index


Jason Brummett
Re: mindlessly obsessed with 16-bit
on Oct 17, 2006 at 4:31:59 am

Depth of Field pass.

Jason Brummett
Studio Z Inc.


Return to posts index

LestDoggy
Re: mindlessly obsessed with 16-bit
on Oct 18, 2006 at 8:03:30 pm

check this video out... this guy is explaining 32 bit hdr color for after effects 7 but he expains
8 and 16... it kinda helped understand this stuff too...

http://www.adobe.com/products/aftereffects/newfeatures.html






Return to posts index

Lars Bunch
Re: mindlessly obsessed with 16-bit
on Oct 22, 2006 at 4:26:58 pm

Hi,

I tend to do most of my rendering in 16bit unless there's a significant reason to do it in 8 bit. My main reason for this is that subtle transitions of color will tend to band in 8bit.

I suppose this is less of an issue in rendering a complex image of a 3D scene, but if you have a smooth sky that is a medium blue at the bottom of the frame and a lighter blue at the top, you might see bands forming between one value of blue and the next.

Anyway, this is an issue that crops up more in compositing programs such as Shake and AE.

Also if you have to do color correction down the line or you want greater flexibility in adjusting the final "exposure" of the shot, then 16 bit or 32 bit will yeild a much nicer result.

A practical demonstration can be seen in Photoshop. If you take a 8 bit image and do some really extreme levels adjustments, often the histogram will end up choppy, with blank spaces within the curve. This means there are NO pixels within the blank range of values in that image. If you do the same thing on a 16 bit image and convert it to 8 bit, the histogram will usually have a smooth curve of values from 0 to 255.

Anyway, if you are going to do a lot of post processing and compositing to the animation or if you are seeing banding in the image, definitly go with 16 bit color. If not, 8 bit is fine.

Hope this helps,

Lars


Return to posts index


Sean Fennell
Re: mindlessly obsessed with 16-bit
on Oct 22, 2006 at 5:43:35 pm

Don't mind if I chime in here about a couple of details regarding 8-bit and 16-bit color depth formats. There is some practical history behind this. 8-bit yeilds 256 colors per color channel (R,G & B) starting with black being a color, and for anyone who cares that goes right down to how bit processing is scaled up in terms of hardware (2 ^ 8 = 256). 16-bit yields 65536 colors per color channel (2 ^ 16 = 65536). Obviously this yields far more color representation than 8 bits. This allows 16-bit to offer smoother transition in colors than 8-bit does. It also allows more flexibility in compositing operations.

Now that the math is out of the way, how about some history. The film industry is where digital imaging dawned. Photoshop was actually a creation of Visual Effects Supervisor John Knoll of ILM on his work for The Young indiana Jones. Anyway, when the digital revolution for film was developing the standard color depth was 10-bit which was log-timed to get more range in the darks, since the contrast in the darks of film were more important than the brights. So most film suite digital editing software used imagery in the 10-bit format. 8 & 16-bit sort of came around as an effect of Video and TV on the digital revolution. 16-bit has more color fidelity, but surprisingly, most film log algorithms are still geared toward 10-bit color representation. So 16 and 32 bit have become a staple of pure digital technology, such as 3-d visual effects and video games. They're full potential will not be realized until we have HDR displays on everyones desk and HDR tv's in everyones homes.

Just some fun facts. But since most software doesn't really output 10-bit anymore...if you're on your way to film with your digital content, its best to stick to 16-bit rendering, since film can handle the color range and video cannot. Video is stuck with 8-bit because of how the data is processed to get to screen. Film is emulsion and not a fixed resolution or color pallate.

Thought that would be fun information for everyone. And I haven't posted in months. heh


Return to posts index

Jason Brummett
Re: mindlessly obsessed with 16-bit
on Oct 22, 2006 at 6:47:27 pm

Thanks Sean,

I ment to say originally 256 (Always say 255)thanks for correcting me! Also as I re-read my post I should be more careful in what I'm saying!

Color spectrums above 8 bit are used a lot more than 1% of the time. I'll try to be more careful :)

Jason Brummett
Studio Z Inc.


Return to posts index

SSGGMM
Re: mindlessly obsessed with 16-bit
on Oct 25, 2006 at 8:07:14 pm

Very helpful. All of this. My original colour maps were photographed in 35mm and scanned at 16-bit. While I don't anticipate transfering to film (and are thus condemned to the video projector's range, if I understood corectly) I think I'll keep the 16-bit anyway. My distributors haven't mentioned anything, but it would be good if they had the option of a print. Am I making sense?

Anyway, I guess my original question was this: Can mental ray tell the difference? Can AE tell the difference? I haven't explored all the settings yet.

Serge


Return to posts index


Sean Fennell
Re: mindlessly obsessed with 16-bit
on Oct 26, 2006 at 4:19:42 pm

Yes they can tell the difference. You're eye may not be able to tell the difference depending on the number of details and wide ranging of colors, but AE and Mental Ray know the difference. There is so much more color information packed into the image that at its core it no longer looks like an 8-bit image anymore. Be careful though, if you output from AE you still have the option to output to an 8-bit format, so be sure you pick your output format to match your input if you want to keep the 16-bit color.


Return to posts index

SSGGMM
Re: mindlessly obsessed with 16-bit
on Oct 26, 2006 at 8:58:34 pm

Many thanks, Sean.



Return to posts index

Luke Goddard
Re: mindlessly obsessed with 16-bit
on Dec 24, 2006 at 10:25:36 am

Also remember that the monitor you are colour correcting these images on is 8-bit, so you won't be able to see all that detail straight away, only when it gets printed back to film. However, to see the difference between 16-bit and 8-bit, open up a 16-bit file in photoshop, open the histogram window and play with Levels. Now save out an 8-bit copy of the original file and try it again. See that comb effect in the histogram window? That's banding.

Hope this helps,
- Luke


Return to posts index

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