using SC 26.2.26 on Win10
I have a 1,5 hr footage (15GB) taken in 4k, 30fps with an iPhone 15. According to Mediainfo it was encoded with HEVC (*.mov) in VBR and has an HDR Profile of “Dolby Vision, Version 1.0, Profile 8.4, dvhe.08.07, BL+RPU, no metadata compression, HLG compatible“
I recorded in 4k because I plan to zoom in and pan on the remote stage. For code size reasons I want to export in 720p, mp4-Format.
Is there anything I can/should (not) do with this “HDR” to best preserve quality of the original, also on non-Apple devices?
Since I don’t live in Apple-land I am not experienced with this kind of format.
Thanks for any suggestions.
@Samse26 I know so far that HDR footage shall be converted to SDR to avoid pale colors. To avoid conversion i shoot only in SDR. In my thread to this specific topic @brian said „unless you have specific workflows that need it (HDR)“
So, just out of curiosity: Is there anything „useful“ i can do in HDR?
@shotcut Thanks for the reference, I should have spotted it myself.
The bad thing is that I recorded the video with someone elses IPhone. I made sure to record in 4k because I knew I would zoom and pan, but since I don’t know IPhones I was not aware of the HDR option. Furthermore the owner of the phone is just a click-and-leave-me-alone person, he didn’t even know how to transfer the file out of Apple-land. So no chance of asking him to export the record in in SDR (as brian suggests in the HDR article). So I am stuck with this HDR thing now.
convert to edit friendly with convert colorspace on. Even in the lowest quality mode I know this would take ages for my 15GB file, so I tried converting just a short sample using the “use sub-clip” option. In souce window I trimmed the file down to 1 min and started conversion. After >20min at 90% CPU load but without any progress in the job window I aborted the conversion. What am I doing wrong here?
using the LUT suggestion: Downloaded 3-NBCU_HLG2SDR_DL_Adobe_v1.2.cube from the referenced location and applied a LUT(3D) filter to a small clip on the timeline. But I saw no difference in colors at all with or without LUT enabled. If it matters, I guess I should, no?
Bottom line: Conversion to edit-friendly with or without color conversion is impractical due to file size (and CPU power ) if I want to get the job finished in this life…
Since applying a LUT appears to make no difference, I assume that HDR/color is not an issue in my case. It’s a concert recorded in a church last week, with fixed camera position. Only difference in lighting is that in the beginning (7pm DST last week) there was a little more daylight than in the end (8:30 pm) so there is a change in color hue.
I will make more tests exporting parts from beginning and end to see how the colors behave. I may have to reduce my ambitions for making nice end credits with transitions, if that should mess up colors.
key takeaway: whenever I am again to record with an IPhone I will switch HDR off !!
so it looks as if nothing was done, out-file size was indeed 0. However I has put start and end markers in the source window about 1 min apart, so expected it would convert this chunk.
When I started conversion without “using sub-clip” it looks as it really started doing something: after I aborted the job, a ~800MB file had been created.
Is the use sub-clip option broken (using v26.01.30)?
Maybe. I’m assuming here that your final use is an SDR video (because if you’re making HDR videos it’s obvious why you’d shoot in HDR to begin with).
Shooting HDR gives you more room to adjust brightness/contrast. Sure, you can adjust SDR footage, but the original footage lost all details in areas darker or lighter than SDR can represent, so even after adjustment there’s no detail there. HDR footage retains detail better in the darker and lighter areas, so when you bring those areas up/down, there’s actually some detail there.
Also, in some cameras (e.g. my EOS R6 Mark II), SDR is always 8-bit and HDR is always 10-bit, so even if you don’t benefit from HDR, you might benefit from the greater bit depth. An obvious use case is if you’re producing 10-bit video. A potentially less obvious use is also for tonal adjustments (brightness, contrast, saturation, etc.). If you’ve ever seen photos or footage where (say) the sky has noticeable bands in it, that’s often the result of tonal adjustments at too low a bit depth. Those extra two bits make the problem much less likely.
) if I want to get the job finished in this life…