Q: Strategies to speed up rendering in compressor
Hi Gang,
Background:
I'm using an old 2008-era Quad Core (intel) mac pro with 5gb RAM and 2.67ghz processor. I have a lot of drive space spread amongst thee external and one internal drive, but my boot drive is getting rather full. I use FCS 3. I have OS 10.6.8.
I have a large number of projects to work on which are basically 90 minute long lectures with an inset of a screen. The video I am using generally consists of two AVC files in HD, with one of them shrunk down as an "inset" in my FCP sequence. It's a very simple sequence with very few filters applied. At most I use a few cross fades and maybe a color-3-way applied to both clips.
The problem:
Despite this rather simple setup, when I send the project to compressor, it takes a VERY long time to render.
Here's what strategies I've employed so far to help speed things up:
- Told Qmaster (in System Preferences) to use multiple instances of compressor for my mulitple core machine (I use 2 of the four I have).
- Exported the project using "make quicktime movie", (which itself takes a few hours) and then importing that into compressor- i.e. NOT using "send to compressor" from FCP.
- Turning off Frame Controls in Compressor (under "inspector")
- Rendering out to a disk that has plenty of room on it.
- Turning off as many other applications as I can
Despite this, my rendering time in Compressor is very, very slow. For a single 90 minute project, being exported out to H.264 (8mb/s roughly), the process takes about 15 hours or so, which is rather prohibitive. To go to an SD DVD format (mpeg-2 usually) takes about the same amount of time if not longer (It's hard to tell sometimes, since I'm usually not at the computer or asleep when it finally finishes.)
What, if any, additional strategies can I employ to speed things up? Would converting the files I use to a certain codec in MPEG Streamclip help? Would it reduce the total time for my beginning-end workflow? IS FCS X a better option? Would breaking up the long file into smaller "chunks" and then rendering the whole mess as a "batch" help?
Any thoughts on this would be appreciated.
Power Mac, Mac OS X (10.6), 4 core intel
Posted on Feb 26, 2013 8:02 AM
So you are telling us that you currently have NOT converted those files in your project to ProRes? You are editing H.264 mixed with HDV (HDV is fine in FCP...it has a codec for that)...mixed with AVC (unconverted) and then some as AIC (which is also fine...that's an Apple codec)??
Yeah...THAT is your problem. The workflow you are currently using is what you'd do if you were using Adobe Premiere. Edit native, then encode to your final format. FCP doesn't work that way. You MUST...must must must must MUST...convert ALL of your footage to an edit friendly codec before editing. That means convert to ProRes. All those H.264 files, all the AVC files. If the DSLR is a Canon 7D or 5D...you can use Log and Transfer if you install the EOS Log and Transfer plugin. I have a tutorial for that process here:
http://www.youtube.com/watch?v=AhDnyWFAM10
This also works with AVCHD...most tapeless formats actually.
>For instance, if I have two different resolutions, for instance, should I re-encode one of the two to match, or would that just be a waste of time?
Not a waste of time...good practice. Unless you are mixing 720p with 1080p...you dont' need to. But if you are mixing SD and HD...good to convert the SD to HD. Mixing frame rates? Yeah, you must convert that...Compressor does a good job.
>Or If I have two different codecs, should I re-encode to match?
Yes....uniformity is good. But, if you capture HDV native via firewire, using that on a ProRes sequence with other ProRes footage...fine. It'll work in real time without rendering, and you can just render when you are done editing and it will be fine. But yeah, all those tapeless formats need to be ONE format....one codec.
>which codec should I chose to make things run most smoothly? Would proRes be best?
Yes...ProRes.
Posted on Feb 26, 2013 11:54 AM