[Marc Wielage]

My opinion is that laptops aren’t ideal for Resolve. I think you need a heftier CPU, more GPUs, and much more RAM. Resolve is demanding when it comes to hardware.

I do think you can use lower-res proxies, like ProRes 422 HD or something, and that can work up to a point.

[Marc Wielage]

Eric, how complex is the project? Multiple timelines, long segments?

If so, my advice would be to see if simplifying the project helps solve the session-loading problem. Maybe take a 1-hour timeline and split it up, or a 10-timeline project and split it into two 5-timeline projects.

That is weird that it will work locally but not from the server.

[Marc Wielage]

Is it just one frame that got bumped, or a bunch of frames? I wonder if just a manual PTZR reposition would solve the problem. I do this all the time on normal longform jobs, particularly when there’s a matched-frame edit and everything moves 1/4″ one way or the other.

[Marc Wielage]

I would check normal SMPTE bars and see if they align in the calibration boxes.

I also don’t recommend internal scopes for anything — you’ll get more accurate results (in general) with external scopes.

[Marc Wielage]

Yes, no issues with 4K source files (including 444 and R3D), though I’m monitoring in HD.

[Marc Wielage]

You have to pay easyDCP for a license in order to create a DCP within Resolve. And there’s a second license you have to get in order to play back the DCP file in order to check it.

[Marc Wielage]

[Paul Campbell] “Buy hey, since you brought it up, would you mind describing briefly your workflow with this project you mention? Are you transcoding the H264 stock stuff to PR so that it will play nicely in your timeline without giving your processor a migraine? How was the other footage compressed that isn’t H264? (Or maybe it was shot with a serious rig, and there was no compression to begin with??)”
I have no idea what camera was used to shoot the material. It could have been a reasonable camera, and for whatever reason the stock footage company opted to sell their material only as 8-bit H.264 2K and 4K. I’m taking it all and transcoding to the exact same resolution in ProRes 422HQ (for 8-bit) or ProRes 444 (for the 10-bit material). I’ve gotten a few oddball QuickTime Photo ad QuickTime PNG files over the years, even some QuickTime Animation (which is a gnarly format), and always converted them to 444.

Given enough horsepower and storage, I’d gladly use DPX for everything. I think ProRes 444 is a good compromise for my clients and certainly doesn’t compromise the material at all. It makes the workflow simpler and is one less thing to have to worry about. I have a whole set of instructions for the editor when it comes to turning over the material for color, but the problem nowadays is a) everybody’s in a hurry, b) a lot of people don’t read, and c) I basically get told, “well, this is the way we did it the last time” and I just deal with it as best I can.

There are some other H.264 issues, like video range vs. data range clips, and you have to stay on top of that as well.

[Marc Wielage]

[Michael McCune] “And I find myself with– yes, it’s true– two or three dozen nodes for a single shot. Say, three people, each with perhaps three masked areas for face, eyes and clothes. Plus some background objects to be lightened or dimmed. Using almost everything: layer, parallel and key mixers, pre-clip vignettes, hundreds of keyframes. “
I find that can work in very complex situations, provided you have lots of time and budget to cover it. If you’re trying to get a 12-shot :30 second commercial done in 3-4 hours, it’d be tough. Impossible with a feature film that has to be done in 2 weeks (which is my usual gig), or a network TV series that has 20 hours for 45 minutes of material. I can and have done 20+ nodes for very unusual situations, but I think sometimes it’s good to take a step back and question whether there might be a simpler way of achieving the same results.

On the other hand, films like The Revenant were done with upwards of a dozen tracking windows relighting every single shot with thousands of keyframes, many of which were manually roto’d by hand, using a half-dozen people, each working 10-12 hours a day, 6-7 days a week, for months on end. So as long as the time and budget is there, go for it.

[Marc Wielage]

[Paul Campbell] “I’ve used this analogy before, but transcoding H.264 to PR would be the video equivalent of converting an mp3 back to wav, yes? You can’t get the original uncompressed information back, all you’re doing is creating a larger file size, but it’s a necessary step if you want to color grade since you’ll at least be grading in 10-bit space, yes?”
No. It’s like pouring 8 gallons of water in a 10-gallon bucket: there’s no extra water there. But it is less stress on the computer and it will play more easily, since the CPU doesn’t have to work as hard in decoding the H.264 compression. You don’t get any more bits that way, even if you pad it out to 10-bit. It’ll still “step” in gradiated-edge or clipped situations.

[Marc Wielage]

[Michael Sasser] “I’ve created text in resolve, then made it a compound clip. Now it is just completely black except for the text. It blocks the clips below it. “
Did you adjust the compositing controls so that the black drops out and the background image pops through?