Announcement

Collapse
No announcement yet.

high ISO editing

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • high ISO editing

    I have photo taken at 1000 ISO (for good reason) which has mottled sky (and other). I would like suggestions on how to improve photo

  • #2

    ​Keep in mind, noise reduction in post processing is a bit of an art. Because every picture and camera is different, there is no one-size-fits all answer. Below are a couple of my favorite thoughts on the topic.

    Noise seems to be so much easier to see in the sky. When the noise is bad enough, fixing noise can lead to image softness. There are many ways to fix the problem, but I will guess that you are looking on ACDsee pointers.

    ​In ACDSee
    ​My favorite trick is to do selective noise reduction.
    Sky and clouds are sometimes OK with a bit of softness. Sometimes softer skies/clouds can look better. By selectively applying higher noise reduction in the sky area you can solve noise and tolerate the softness without impacting the rest of the image.

    There are many schools of thought, so YMMV with this one... Do noise reduction before other edits (RAW or TIF). Thus, the image is a bit friendlier to the ACDSee LightEQ, ColorEQ, and Dehaze algorithms. (i.e. you have less chance of exasperating noise later in the post processing steps). Personally, I think the ACDSee algorithms could be improved the data comes near the limits of the dynamic range.. Don't get me wrong, I really like these features and are glad they are there. But until the algorithms are tweaked, I try to lessen the possible impact by avoiding contributors. In my experience, reducing both color and luminosity noise first, helps.

    ​In general...
    ​Do all editing, including noise reduction, on RAW or TIF files instead of JPG. By default, JPG strip out some of the data. The result is inability to interpolate/calculate the correct color/tone which results in noise artifacts.

    Another major contributor of noise.... looking to close. For example, if the image is only going to be seen as an 8x10, looking for problems at 0.008 x 0.010 is an exercise in self-inflicted pain.

    Comment

    Working...
    X