Loading...
> >

Color depth question

Oso3DOso3D Posts: 15,085
edited April 2016 in The Commons
I've been rendering at twice intended size, taking the result and converting 8 bit to 16 bit, then halving size. It seems to me I should end up with a nice high quality image using that color depth. Am I right?
Post edited by Oso3D on

Comments

  • FishtalesFishtales Posts: 6,212

    That wouldn't work as you are trying to add colour depth that isn't there in the 8 bit image. All you are doing is spreading the colour thinner over the 16 bits.

  • Oso3DOso3D Posts: 15,085
    But wouldn't the algorithm that compresses four pixels into one have more color 'room' to choose from?
  • mjc1016mjc1016 Posts: 15,001

    No...it's going to be limited to the information contained in the 8bit image.  

  • Oso3DOso3D Posts: 15,085
    I realize. But if I have a 2000x2000 image at 8 bit, it seems like more of it will be preserved going to 1000x1000 at 16 bit than rendering 1000x1000 at 8 bit.
  • Richard HaseltineRichard Haseltine Posts: 108,063

    I agree with Timmins - in combining the four pixels the extra colour depth will be used somewhat, though that's useful only if further editing is planned (or if tonal adjustments are done before resampling but after depth conversion, since it will ensure that colour operations do not start to posterise the image). Whether the gain is discernible is another matter.

  • Oso3DOso3D Posts: 15,085
    To some extent I've been using 2x render to help make a crisper image without using a lot of gaussian pixel blurring. Although if I really want to do it right I should just use canvass and go 32 > 16.
Sign In or Register to comment.