Need to optimize Alexa-FCP workflow
I'm currently on a shoot with 2 Alexa cameras. One camera has the codex OB recorder for big totals and vfx shots, but 90% is ProRes4444 LogC. However, I feel that my workflow is too slow for my ProRe files.
MacPro 2x3.2GHz Quad core
eSata interface to 2x RAID5 enclosures.
Sonnet QIO to read the SxS
NVIDIA GeForce GTX285
Pomfort Silverstack SET 1.3.1
I also have FCPX and compressor 4 installed, but don't use it (yet).
So, my workflow is offloading PR4444 in Silverstack, FirstLight grade (LUT) in Silverstack, eksport FCP xml for each reel, import xml in FCP, send to compressor, render out ProRes422 (proxy) for edit.
Up to this point everything sounds pretty good.
My problem is render time combined with Compressor failing to render.
I need to render during night, and every morning I have messages in compressor like:
•Failed: QuickTime Error: -50
•Failed: Final Cut Pro generated an error and unexpectedly quit. (FCP is actually not quitting)
• In general I also get one two clips each batch thats black.
Sometimes it's all good, but some mornings I have to start render 3-4 reels (≈30 clips) all over again.
This is pain when I don't got the time to ensure that the second batch will render properly.
So my question is if someone have a workaround that will work with the FCP xmls from Silverstack and combine FCP7 (or X if possible) with Compressor 4 that I hear have faster render and uses more of my cores? I also have the graphics card that does not do anything right now. Of course I need my original filenames and proper timecode.
I will for sure invest in a Colorfront or similar system in the future, but right now I need it to work with what I have.
Will really appreciate any suggestions or solutions for my worries!
Thanks for reading!
This may not directly answer your question, but here is a workflow on a current Alexa job for a series of big commercials. 2 cameras, double-system audio, ProRes4444 and some 5D in the mix.
Editing in FCP 7 on a Mac Pro 12-core/16GB RAM/ATI5870 connected via FC to a SAN.
I've imported all the Alexa files as native ProRes4444. Applied the Nick Shaw LUT filters to correct to 709 for editing/viewing. Tweaked matrix settings (in the filter) by eye to match the cameras and adjust for green removal.
I've synced/merged the clips with the audio, then created synced 2-video-track sequences to see the two camera angles side-by-side. So I'm starting out the edit with a set of "selects" sequences showing my coverage on all spots. At this point everything plays in real time. No renders, no transcodes, etc. (Note - the 5D footage is transcoded to ProRes using MPEG Streamclip.)
The one caveat is that this project has been very crash-prone and subject to out-of-memory errors. I suspect it's because I'm pushing the system pretty hard. I had done an earlier flight of spots for the same client, which went a bit smoother. No FCP issues, but I only had 1 camera, onboard audio and was using the FCP color corrector instead of the LUT filter.
In the end, the project will be graded in Baselight, so I'll strip out all filters and send an EDL plus the QT camera files. When I get the graded clips back, I'll relink and add graphics and mixed audio back for output.
Hope that helps.
Oliver Peters Post Production Services, LLC
I've used the Shaw LUT before, but the DP wants more extreme corrections for the edit than just a logc to 709 correction. You say "tweaked the matrix" what exactly do you mean? Can I adjust colors in the Nick Shaw plugin? (I don't see it).
My job is only to backup to RAID and LTO, then do a firstlight and transcode the editing files. Then the logger does the sync job after I'm done (on a different system).
My main concern is how unstable compressor behaves with the renders - its pain when the render is ≈5:1/4:1...
[Andreas Herzog Grimso] "but the DP wants more extreme corrections for the edit than just a logc to 709 correction"
Too bad. My own opinion is that this tends to be pointless, because those corrections get redone anyway at the grading stage. So it's a lot of extra effort that never really gets used. No disrespect intended towards the DoP, since he should get the look he's going after. I'm just not sure the editing stage is where that ought to be.
I prefer it when the DoP determines a look as a reference for color-correction, but leaves the dailies relatively neutral so as not to encumber the post workflow. Especially since you mentioned that the Arriraw files were going to be used for the VFX shots.
[Andreas Herzog Grimso] "You say "tweaked the matrix" what exactly do you mean? Can I adjust colors in the Nick Shaw plugin?"
Not really. There is a film matrix on the camera that is intended when film-out is the target. However, sometimes this is engaged when it shouldn't be. In the Nick Shaw LUT (newest version) there's a selection to apply the matrix setting by "add", "remove" or "as is". You can also toggle his additional matrix on or off.
[Andreas Herzog Grimso] "My main concern is how unstable compressor behaves with the renders"
Good luck. I've given up on Compressor for batch processing.
Oliver Peters Post Production Services, LLC
I see your point regarding "grade" at this early stage...
As long as I need to transcode I must go via compressor. All 24fps images becomes 23.98 when I transcode out of FCP. I can do a conversion in cinematools, but I don't feel confident about that solution - and it will take even more time..
I've donated $30 to mr. Nick for the updated plugin..