Hi everyone,
I have request 90 mins documentary film finishing. ( Some touch up, Title. Grade…)
Anyone have experience long form work in flame? Any tips ?
Especially Storage management.
Hey there Peter. I don’t do those for my day job but I’ve done maybe 5 very low budget feature length docs / films over the last few years all in Flame. I do the online, grade, and minor cleanup / simple compositing. Everything worked great on my end, no big issues that I can remember so I’d say you can have some confidence that it can and has been done.
Workflow-wise, it’s probably important to not cache if you can get away with it since the volume of media is quite large. If you come back with any specific questions or hurdles that pop up along the way I’d be happy to chat more!
Great answer! Thanks Jeff
I did a full length feature in Inferno, back in 2000, conform, grade, VFX, end roller. All worked fine, but obviously, quite time consuming.
Oh inferno… miss you!
Done some back in the day, I remember underestimating the time it takes to do a simple export…
I used to do a fair amount of long form online in flame. Not full grade, that was done within the facility in Resolve. If there were VFX that needed to be done pre-grade or go to outside vendors, those plates were pulled in Resolve.
Rarely did I re-conform the whole film from individual clips in flame. Usually had grade export a single mov that was notched on the flame timeline and then if there were sequences that needed a lot of work, I’d get those shots as individual renders and only re-conform that portion. The biggest example of needing to do this was documentaries with a lot of pan and scan stills or archival footage that needed field merging or other cleanup.
Exporting the whole timeline is time consuming, faster than realtime, but I wouldn’t say double closer to 1.25x.
The other thing to watch out for is if the timeline has a burned in letterbox. If you make one change to anything on the track below, it un-renders the letterbox. Not a huge deal on a 30 second commercial but can be if you are trying to do real-time playback in a supervised session.
One trick around this I found was to notch (right+click the track head on the right, copy transitions to LB track) the letterbox itself. That way the letterbox will only need to be re-rendered on the section you are working on.
The one big hiccup to all of this is HDR projects. What we were finding very quickly with HDR, is it just made more sense for the timeline to live in Resolve. Individual shots would be pulled in resolve and might be worked on in Flame (or by a vendor in whatever software). But titles etc all made more sense living in resolve so they could be adjusted as part of the grade.
I love long form projekts in flame.
You have to have an appetite for success and catastrophe in equal measure.
Don’t be surprised when your adequate equipment is actually less than adequate.
Derive the number of frames by dividing the total running time by the frame rate and multiplying that number by the byte weight of a single frame.
Add the byte weight of your audio.
You need between 3 and four times this quantity of storage before you even get complicated.
Think about all the things that can go wrong and try to come up with solutions to those problems before they happen.
Never do someone else’s vanity projects for free.
I hate to say it - but I will. Do it in Resolve! ive done a few in Resolve and the grading is far beyond anything else. as is the pipeline with multiple camera formats! Sorry
I’ve done it in most of the apps that are around. Flame is totally capable of it. But you do want to spend enough time thinking through the workflow and learning the parts of the workflow that will keep you efficient and keep you sane. Do a test, read the manual, make a plan before you dive in. Especially know how you will handle the look vs. individual clips, and how you will handle revision and notes. That’s the hill you die on otherwise. Also what color management you’ll use. ACES 2.0 is a good choice for multi-source material, but find out what the sources are that you have all the IDTs lined up or know what to use.
You want to have a standardized image node tree. You want to make extensive use of the storyboard, groups, and TL-FX explorer. Consider stacking image nodes rather than doing everything in a single one. Get familiar with the image processing pipeline. And know your go-to Matchboxes. If you are concerned about performance, totally fine to break it up into 3-4 reels. And if you do that much grading, at least a simple control surface such as a Tangent Wave 2 is worth the tactile control it gives you.
I would not do the titles in Flame. While the new text tool may be tempting, this is the domain of Adobe InDesign and AfterEffects. Everything else is not worth it (except for a simple title card, but even then…)
Would I use Flame, Resolve, Premiere, Mistika, or Baselight to do it? Well, it depends a bit on what your clean-up needs to be. Personally, long form color, especially on a Doc Baselight wins every day. Premiere dies on secondaries. Resolve is the fastest and handles the most formats withou fuzz, but is uninspiring and the tools work like Google products. The first 80% are great, the rest is crap. Flame is the perfect place if you have a decent amount of finishing and cleanup task, with the broadest set of tools in an integrated workflow.
As opposed to scripted which have a lot of consistent and well shot camera material, docs are a trip to the flee market most days. They will test the quality of your tools and process more than other projects because of that. That’s where those remaining 20% matter. X-Grade has much better math than Resolve’s tools, and Flame’s keyers, especially the diamond keyers are not efficient nor are they best in class.
In terms of storage - long form is too big to fit on NVMe. Make sure you have a good spinning disk RAID, like RAID 5 with 8 drives, connected via TB if on Mac, eSata if on Linux. That’s plenty fast and you can get 30-40TB at reasonable cost. Make sure you have a render drive that’s separate from your footage drive - never read and write to the same spinning drive.
Do consider transcoding some weird codecs into ProRes ahead of time. It’s worth the time in prep.
Wow! all the amazing answers and opinions! Huge Thanks to all!