Hi,

I'm wondering why I cant seem to eliminate very dark nighttime shots from my sequence, no matter the parameters used in the Long Term Workflow tab of Luminance, Contrast and Hue? Whether its Constant or not, the image filtering only seems to reduce random images out of the folder rather than targeted specific ones...

For example, in this folder I have filtered the photos down from 6583 down to 696, and yet i've still got night time images and fogging issues etc etc. Why arent the filters providing me with similarly exposed photographs in similar environmental conditions?

I used LRT about 6months ago and dont remember this issue, but then again I might've done manual culling of images prior to importing to LRT and perhaps thats why it worked well..

Surely I am doing something wrong here?

Many thanks to anyone who can provide advice.

PS. I have watched the tutorials several times.

This post was last modified: 2020-11-12, 10:04 by Gunther.

It's hard to say without really seeing how you are working and how that sequence looks like. Of course the filters work only on the preview files and there is a certain threshold where they can or can't detect changes. The key for filtering is using an appropriate reference area in order to get a blue luminance curve which clearly shows dark ares (low) and brighter areas (higher). Now you can set the smoothed lum curve to Linear and move it up down to cut off the dark or bright parts. That's how I would approach this.

OK, thanks... I guess i'll keep trying.

Hi Gunter,

Just wondering if you could point me to where I can learn how to:

'...using an appropriate reference area in order to get a blue luminance curve which clearly shows dark ares (low) and brighter areas (higher). Now you can set the smoothed lum curve to Linear and move it up down to cut off the dark or bright parts. That's how I would approach this.'

Do you mean drawing the square on the image display window with my mouse cursor?

Do you mean that I should be selecting within that square areas of high contrast in order for LRT algorithm to understand what 'dark' and bright'?

I remember I saw you do that in the longterm tutorial, but didn't find it super clear...

thanks