- Joined
- May 19, 2019
- Messages
- 194
- Reactions
- 206
Hi,
I'm about one month in to my first drone (MP) and I'm trying to learn as much as i can about post-processing of images (and eventually video) to get optimal results. After a lot of youtube advice I settled on D-Cinelike +1-1-1 for all the good reasons you hear about on youtube (and in here). I'm using GIMP to post-process my images. I've watched a lot of tutorials and understand that because of the limitations of the 8 bit sensor in the MP, D-Cinelike is not a true 'log' capability but it gets acceptable results out the camera and will improve a lot more than the other styles (except D-Log) when post processed.
So the fundamental question that remains after many youtube vids is: what are the generic corrections needed to any D-Cinelike image that should result in an image that is ready for final tweaks?
My experiments in GIMP show that in the Curves view my image usually fits within 5% - 92% of the histogram range, so by adjusting the Curve to those values I get the image to lose a lot of its flatness and the contrast pops right out. Did I get those numbers right? Do I need to also lift or drop the mid-tones by a few percent? I'm guessing yes, but its tough to figure out how much in general.
The image also needs more saturation - I find that raising that from 1.0 to somewhere in 1.14 - 1.20 range seems about right. Thoughts?
Hardest part for me is Color Balance - does D-Cinelike intrinsically alter the color balance such that there is a standard tweak needed in post? I'm guessing a small lift in red is needed, but welcome opinions.
Now, I know that everyone has a LUT to sell or give away, but I want to learn. I have tried a gazillion LUTS (in G'MC QT plugin) but only about 3 of the hundreds in there seem to do the right things to a D-Cinelike image. Advise welcome, but I'm not looking for 'use this LUT' answers unless I understand what the LUT is actually doing in terms of levels, saturation, balance, etc.
I'm about one month in to my first drone (MP) and I'm trying to learn as much as i can about post-processing of images (and eventually video) to get optimal results. After a lot of youtube advice I settled on D-Cinelike +1-1-1 for all the good reasons you hear about on youtube (and in here). I'm using GIMP to post-process my images. I've watched a lot of tutorials and understand that because of the limitations of the 8 bit sensor in the MP, D-Cinelike is not a true 'log' capability but it gets acceptable results out the camera and will improve a lot more than the other styles (except D-Log) when post processed.
So the fundamental question that remains after many youtube vids is: what are the generic corrections needed to any D-Cinelike image that should result in an image that is ready for final tweaks?
My experiments in GIMP show that in the Curves view my image usually fits within 5% - 92% of the histogram range, so by adjusting the Curve to those values I get the image to lose a lot of its flatness and the contrast pops right out. Did I get those numbers right? Do I need to also lift or drop the mid-tones by a few percent? I'm guessing yes, but its tough to figure out how much in general.
The image also needs more saturation - I find that raising that from 1.0 to somewhere in 1.14 - 1.20 range seems about right. Thoughts?
Hardest part for me is Color Balance - does D-Cinelike intrinsically alter the color balance such that there is a standard tweak needed in post? I'm guessing a small lift in red is needed, but welcome opinions.
Now, I know that everyone has a LUT to sell or give away, but I want to learn. I have tried a gazillion LUTS (in G'MC QT plugin) but only about 3 of the hundreds in there seem to do the right things to a D-Cinelike image. Advise welcome, but I'm not looking for 'use this LUT' answers unless I understand what the LUT is actually doing in terms of levels, saturation, balance, etc.