|
|
All the news now and in your digital future
Re: Transcoding red files
Posted by Ethan
|
Re: Transcoding red files July 19, 2009 08:20AM |
You'll get more attention by posting a new thread than by replying to an old one. So I've split this off into a new thread for you.
When you refer to "transcoding using FCP," I'm gonna assume you're talking about using log-and-transfer to convert R3D to ProRes. There are a lot of reasons not to go this way. I haven't investigated the L&T plugin for a while, so I'm a little rusty on exactly what the liabilities were, but they were significant enough for me to say "no thank you" and not look back.
For batch-converting a whole bunch of R3D media to ProRes, you really want to use Red Rushes. Set up your look in Red Alert, then use that look when batch-transcoding with Red Rushes. Since you're not going to do a DI on this (which is a valid option when shooting Red, but one that limits your flexibility) you'll be best served applying a nice, flat look which doesn't clip anything and that preserves as much dynamic range as possible by keeping the contrast to a minimum. You want, in essence, as close to a flat curve as you can get. Think of this as your "best light" pass in telecine; think of your subsequent tape-to-tape grading as your "scene-by-scene" pass.
If you're going out to SR as video, not using the data option, then your format is going to be 1920x1080 at either 23.976 or 29.97, assuming you live in NTSC-world. Note that if your material was shot at 24, it will have to be conformed to 23.976. You can choose, in Red Rushes, whether to center-cut your 2:1 material ("fit height"
or letterbox it inside the HD frame ("fit width".
There's really nothing you can do to reduce processing time; converting R3D to Quicktime, DPX or TIFF is equivalent to sending your film to the lab, and as such you have to build time into the schedule for it. You could try doing a quarter-res debayer, but you probably wouldn't be happy with the results; in my experience, when going down from 4K 2:1 to HD, most folks debayer at half-high for the best compromise between quality and processing time.
Note that your basic workflow is not a bad one! There are tons of people out there, from what I can tell, doing just what you're talking about: shooting on Red, then getting out of the R3D format and into a video format as the first step in their workflow, before finishing traditionally as if the job had been shot on SR or whatever. There are definitely some catches ? since you're transcoding out of R3D first, it's imperative that you make sure you get it right, or you'll have to do a lot of backtracking to correct any mistakes or make any changes.
When you refer to "transcoding using FCP," I'm gonna assume you're talking about using log-and-transfer to convert R3D to ProRes. There are a lot of reasons not to go this way. I haven't investigated the L&T plugin for a while, so I'm a little rusty on exactly what the liabilities were, but they were significant enough for me to say "no thank you" and not look back.
For batch-converting a whole bunch of R3D media to ProRes, you really want to use Red Rushes. Set up your look in Red Alert, then use that look when batch-transcoding with Red Rushes. Since you're not going to do a DI on this (which is a valid option when shooting Red, but one that limits your flexibility) you'll be best served applying a nice, flat look which doesn't clip anything and that preserves as much dynamic range as possible by keeping the contrast to a minimum. You want, in essence, as close to a flat curve as you can get. Think of this as your "best light" pass in telecine; think of your subsequent tape-to-tape grading as your "scene-by-scene" pass.
If you're going out to SR as video, not using the data option, then your format is going to be 1920x1080 at either 23.976 or 29.97, assuming you live in NTSC-world. Note that if your material was shot at 24, it will have to be conformed to 23.976. You can choose, in Red Rushes, whether to center-cut your 2:1 material ("fit height"
or letterbox it inside the HD frame ("fit width".
There's really nothing you can do to reduce processing time; converting R3D to Quicktime, DPX or TIFF is equivalent to sending your film to the lab, and as such you have to build time into the schedule for it. You could try doing a quarter-res debayer, but you probably wouldn't be happy with the results; in my experience, when going down from 4K 2:1 to HD, most folks debayer at half-high for the best compromise between quality and processing time.
Note that your basic workflow is not a bad one! There are tons of people out there, from what I can tell, doing just what you're talking about: shooting on Red, then getting out of the R3D format and into a video format as the first step in their workflow, before finishing traditionally as if the job had been shot on SR or whatever. There are definitely some catches ? since you're transcoding out of R3D first, it's imperative that you make sure you get it right, or you'll have to do a lot of backtracking to correct any mistakes or make any changes.
Sorry, only registered users may post in this forum.
