Image exposure/brightness adjustment

Answered

Comments

6 comments

  • Avatar
    Wishgranter
    Hi Jon

    RAW adjustments...

    its best if use some good RAW workflow "develop" sw like DXO Optics where you do all the adjustments to the image before processing in RC. Its because at least for now use use standard Windows RAW settings loader that are not the best when we talk about quality.

    So you prepare perfect images corrections in DXO Optics and just then use the data in RC.
    Its depend on projects but mostly properly developed RAWs to TIF-JPGs with 8bits are OK

    The 16-32 bit image workflow is mostly used in VFX ( games ) workflows.. where it can be really used in production. If need orthos, renders and etc then the 8bit are in most cases OK.
    0
    Comment actions Permalink
  • Avatar
    Götz Echtenacher
    Does reducing 16-24 to 8 bit improve performance at all?
    Hope that isn't too much of a noob question... :D
    0
    Comment actions Permalink
  • Avatar
    Jon
    Thanks for the tips.

    I currently use Adobe lightroom to do the corrections - it has been working for now, I've not had a chance to test noise reduction and contrast adjustment to see how it impacts alignment.
    This might be a strange question, but if I have an 8bit JPG that looks dark, are there less feature points added?

    I'll attach example which shows an exposure from the camera that tries to make sure highlights do not blow-out.
    Ideally a camera would shoot HDR but this would require a tripod.
    I have processed this image and reduced contrast / changed the curves to keep detail in highlights, but bring up the shadows.

    Ultimately does this affect the RC algorithms? this is certainly the case in photoscan.
    0
    Comment actions Permalink
  • Avatar
    Jon
    Götz Echtenacher wrote:
    Does reducing 16-24 to 8 bit improve performance at all?
    Hope that isn't too much of a noob question... :D


    I believe it does not affect performance too much because an 8 bit JPG still needs to be uncompressed internally in most situations.
    Maybe memory consumption might drop, but I'm curious about this too.
    0
    Comment actions Permalink
  • Avatar
    Wishgranter
    its a 8 or 16 (32)bit per channnel

    for the speed its almost 3 times more data so even recon time get higher processing 16 bit imgs. From our internal testing the 8bit vs 16bit are in most cases with just small improvements. its highly advisable to make more images if possible and some of them edited for better color rendering = play with the contrast and brightness a bit.
    0
    Comment actions Permalink
  • Avatar
    Neil Nafus
    @Jon I tested Lightroom, Photoshop and DXO. Of those, DXO seems to deliver, hands down, the best results with the least amount of work. Both Lightroom and Photoshop perform some sharpening that seems to introduce a decent amount of noise in the results. I spent a lot of time testing various settings and the first results from DXO were better than anything from either of those.

    I also found that with my camera array, in cases where I had shutter shadow, Photoshop and Lightroom produced different results for those images from the rest in the set. The results from DX0 were very consistent. FWIW
    0
    Comment actions Permalink

Please sign in to leave a comment.