Has anyone got any tips they could share for working in Davinci Resolve? I am just getting into it, and it seems fairly good for stop motion. Interested to hear from anyone who has some experience with the program.
Hans, you clearly know a lot about this, and I am struggling to understand all that you have written.
I suspect that what you say is a counsel of perfection. I tried doing some shots in RAW and the files come out overwhelmingly large. So I shoot using Large Fine JPEGs, which are manageable, and I suspect many others do the same. I am not aware that the JPEGs have to be processed again, except for the final render, but am a bit uncertain about the workflow. You mention exporting as tiffs and then importing - but I was under the impression that DR does away with all that, with its attendant loss of quality. It surely does not render between Fusion and the Edit page?
Does this sound about right for the workflow?
Import JPEG sequence into DR project (do I set the project to 1920x1080 or UHD and does this make a difference when importing the JPEGs to work on? Some videos suggest it doesn't matter until the time for selecting to export.) Create a Fusion composition and put the sequence into that.
Do rig removal with masks in Fusion, this would be at full Res and not requiring smooth playback. Then select optimized media or proxy media for editing, once the sequence is put into the timeline.
Then do color work, audio etc. And finally export.
If I need to move clips around, as in to other programs, I understand ProRes422 is what is needed to maintain quality.
As this is all so complex I have tended to find a path through the maze that seems to work for me and stuck to it, although it might not be optimal. Having spent a very long time getting my head around PPro and AE, I am hoping that DR will not take quite so long, but there are a lot of things that function differently.
Any advice is extremely welcome!
Hans Jacob Wagner said:
Well that's why you use proxies, when you are satisfied with how it looks running with the proxies, then you turn on the originals in the render, and leave it to work on the output. Don't expect it to do much else than crash if you try working with highres originals in the program. Same goes with fusion. As for using .jpg as originals - it's a highly destructive compression, and should only be used as a low. It's a bit like having a betacam recording making a vhs copy to edit, and output to vhs.
The jpg algorithm is making the images increasingly worse each time it's being applied. I would rather import raw if possible or else 64bit tif. Make your mind up regarding the resolution (like 1920x1080) and scale the images accordingly. Maybe you can get away with doing the colour grading at this point as well, and export the graded images as 24 bit tif files.Then you get the most out of the colours. Then import those for editing using proxies. The early downscaling and colour grading, will be making the editing a lot easier. Just remember that if you plan to truck in or make other subtle camera moves, do so before the downscale and that goes for green screening as well, do those effects before you export to 24bit. Scene by scene. Working in 4K in raw or 64 bit calls for dedicated hardware. Jpg only stores 256 levels for each RGB channel, 64bit 1024, and raw is direct camera dump, it doesn't get better, changes from camera to camera, and should be handled with in floating point resolution in fusion or resolve. Don't use jpg in the production, save that for an internet output like Mpg-4
Simon Tytherleigh said:
One thing I have been discovering is that Davinci Resolve allows you to bring in jpeg sequences, create Optimized Media proxies, do all the work needed and then export from the original jpegs. This should mean that it is a simple matter to export at 4K or HD, and that one does not need to use a video intermediate, so there should be the absolute minimum of loss of quality. I have always been concerned that the need to create video footage just to be able to edit it introduces a loss of quality from the original.
When I generated optimized media for one clip DR showed that the original image sequence was about 500Mb but the proxy clip was about 25Mb. I tried to add some effects to the original, and it soon began to stutter badly on my 2015 iMac. But the proxy plays like a breeze.
Still learning, so if anyone knows better about this, please correct me!
This video looks like it has some useful stuff for stop motion animators, using the Paint mode and Planar Tracker in Fusion to paint something out.
Great for those bits that get slightly moved during a moving shot. I think I would just create a simple mask for a static one, though.
Sort of responding to a thred over on FB, but I wanted to put this link here so it doesn't get lost.
Here is a link to a youtube tutorial on animating masks in the Fusion page. It is particularly relevant to stop motion from about 5:00 onwards, and shows where the controls are hidden!
Hmm, thought it would show a thumbnail, but it didn't.
Thanks for posting that link Simon, his channel is a goldmine of Resolve Tutorials.
At last I have found the solution for rig removal in DR! It is obvious really, but my previous experience with AE had led me to think of Fusion as a standalone program, when it is a page integrated into Resolve. I expect rig removal can all be done inside Fusion, but I am still on lesson 101 and learning quite slowly.
You can stack clips (images and sequences) in the Edit page, with the top video layer being the one visible. Then click on the one that needs the rig removing (the top one, with the clean plate below) and go to the Fusion page. The rest is explained in this video, from about 3:50 onwards.
It all reminds me so much of learning AE, when I would suddenly grasp something after months of puzzling... Anyway, I hope it helps someone!
That was very helpful and seems relatively easy.