PDA

View Full Version : RGB -16-bit or 8-bit for rendering?


BossaNova
June 25th, 2009, 01:02 PM
If my computer is strong enough to handle 16-bit RGB should I use that for rendering or does it even really matter? Is 8-bit better not really sure what differences this makes kinda need a basic crash coarse on what this does for your artwork.. ? all I know is in 16-bit my PSD file is Huge and in 8-bit it's not so big.. Some help on this would be appreciated..

Metsys
June 25th, 2009, 05:01 PM
16-bit RGB is a high dynamic range image. With 8-bit you can only have the brightness of each channel go from 0-255, but 16-bit it goes from 0-65k. With 32-bit floating point HDR it gives you an incredible range of values.

Some of the more expensive cameras save HDR images that are great for photo editing because you don't have to worry so much about blowing out your hightlights when you take the photo. Even though the highlights are blown out to white on the display of your camera, and when you preview them in the computer (which would otherwise destroy visual information when you want to adjust the brightness), with HDR images those highlights are saved. And the same goes with the darks, so you can lighten shadows and not have strange artifacts because a good range of color information is still saved there.

Video games that use HDR will use 16-bit (half) HDR textures so that it can calculate the HDR effects in the game. Also HDR images are useful when doing displacement maps for 3D models where you need smooth transitions.

For concept art or illustrations I don't see any reason to save as 16-bit HDR images, unless there's some reason why you need to (I've never encountered one). That range is really only necessary for photography so you can edit the exposure of the photos after you have taken them without artifacts, for rendering 3D scenes and compositing for the same reasons, or for textures and rendering certain effects in your 3D software or video game.

jfrancis
June 28th, 2009, 09:36 PM
For photography I always use 16 bits. Many will argue that 16 is overkill even for photography, but I use it. For most graphics, I'd think '8 is enough'

:D

If you actually see a difference...

If you don't mind the extra storage space needed...

If PS gives you all the filters and blend modes and whatever that you need to get your work done in 16...

...then use 16. Otherwise 8 should be fine.

Stevie Doodle
July 20th, 2009, 04:31 PM
One difference you can notice between 16 bit and 8 bit is in a gradient. An 8 bit gradient is a bit choppy, while 16 bit one is more smooth.

If you're exporting in a lossy format for web it's not gonna matter, but in a lossless format for print or somethin' it will make a difference.

bizarre
July 21st, 2009, 09:26 PM
yeah, you'll have banding issues in gradiants and smooth edges in the lower bit-depth images.

banding is when instead of a smooth gradiant, you'll get these 'bands' (as the name implies) of color. much like my signature, heh.

PieterV
July 23rd, 2009, 01:04 PM
Yeah banding can be a bit of an issue sometimes when you have smooth gradients of color that are very close to eachothers (in terms of color).

However, there's a difference between how an image looks in Photoshop and how it looks once exported. Sometimes you'll have banding in Photoshop that doesn't show up at all in the exported image, it has something to do with the way PS handles the colors. So when in doubt, just export an image, it's usually just a Photoshop fluke.

bizarre
July 23rd, 2009, 03:06 PM
off-the wall q: does Photoshop use hardware acceleration? filtering???

curious...

blankslatejoe
July 31st, 2009, 11:55 PM
cs4 uses graphics card acceleration and filters the images using bicubic filtering by default, though you can edit that in the prefs.

Also, I've heard 16bit is useful in rendering/3d where the extra data can involve lighting information (like the aforementioned HDRI) that's carried through to the scene.

BossaNova
August 13th, 2009, 04:03 PM
Thanks everyone for your great info on this question.. for this last contest I used 16-bit for all my image work then compressed down to 8-bit when it was all done I'm not really sure if this made much difference in the end run of things but it sure felt like it did.. but still I don't think the longer save times are really worth it at all.. for now on I think I'm just going to stick to 8-bit.

Thanks everyone