Hibbert6 wrote:Doing a virtual choir video. 3 minute song. Laptop is top-of-the-line Dell XPS.
25 voice choir, in 5 parts. I do the sopranos, add the altos, and things begin slowing down. By the time I finish adding the basses my laptop has crawled to a stop, the fans are blowing like crazy. My 2nd to last rendering took 4 hours, and I had to turn it off halfway through to cool. This has happened the last 3 of these I've done, but this song had more voices and the problem got way worse. Says I'm using 95% of cpu. Any suggestions?
Intel core i7-9750H 9th gen
Ram 32 GB
64-bit
Windows 10
NVIDIA GeForce GTX 1650 4GB
1TB M.2 PCle NVMe SSD
You're trying to pull 25 streams of video simultaneously on a a laptop that probably doesn't have sufficient cooling, only has 4GB of VRAM, and isn't that performant in the grand scheme of things. So you have to do something to decrease the load.
Two things come to mind...
First, if you have 25 people onscreen mostly together, they're each going to be fairly small and you don't need to have full-sized video for them being pulled into your timeline. You could pre-render each of them down to 1/5 size or whatever size they end up onscreen. That way you're still pulling 25 video streams at once, but not 25 full-sized streams. At 1/5 size, you're actually pulling 1/25 the data in each stream. (Though it is from 25 different files, which is still a fairly high demand.) If you're cutting to larger images for solos, etc, you could still bring in those individuals at the original size. The important thing is you're not trying to pull 25 full-sized video streams for most of 3 minutes.
Second, you could go the other way around: edit the altos only and output them together as a full-sized render. Then the tenors, etc, etc. That way you only have to pull four (full-sized) video streams into your timeline at once. You'd want to make you're intermediate renders higher-quality than your final product, since you're going to two-step things. Depending on how the different sections are organized in your video, you may have to use an alpha channel or a mask or something so that all four layers show onscreen at once.
The same for audio. No need to have 25 separate vocal tracks in your timeline, since they are a choir. You could do a first pass where you sync everyone up and then render just the audio and reimport that. (Audio is a much lower burden than video, but your load here is cumulative, so simplifying each part helps.)
A beefier machine would help, but then maybe you'll want to do 50 vocals and you'd still need to do things like I've suggested.
Resolve Studio 19 latest, Fusion Studio 19 latest, MacOS Sequoia latest, MacBook Pro M4 Max
Sat Dec 19, 2020 5:23 am