Adrian Sancho
Forum Replies Created
-
>>The only compressed options for 4:4:4 are HDcamSR (which is obviously restricted to tape) and CineForm.<< Are these lossless systems? Even at roughly 2:1 lossless ratios I wouldn't think it detrimental for comp work. It's only in the compositor where you'd need it (unless you want to torture yourself keying in your editor :-D), and after you're finished with the comp you can get back to a working format.
-
Why not just put your room environments on planes in a modeler? Matchmove to your live action.
-
>>I’m most concerned with getting a realistic composite between the staged actor and the real environment…<< Isn't it easier to synthesize the "real environment"? Then you can matchmove the keyed actor (which essentially becomes the "real environment") to make a perfect sync, no?
-
>>you mean keep at least one camera running….in other words stagger the reload times, right?<< Sorry for the late response, but you have it right. Stagger. If so, you'll need at least two house audio feeds running to two cameras you can stagger the audio to as well, otherwise you'll have to be a bit creative with your audio mix. Been there, done that too.:-) It's actually not hideous if you can sync and blend for a good amount of time. If you need to sync disparate audio, and you an run VST plug-ins on your audio (video) editor, check out the Phase Bug: https://www.betabugsaudio.com/plugs.php
Scroll down to the bottom and get the Phase Bug (and anything else you think you might like/need). This freebie can be a life-saver when syncing audio from multiple cameras (or other multiple audio sources).
-
Hmmm, it’s surprising that in this day and age you still can’t get 4:4:4 digital capture without an elaborate setup. I remember a few years back JVC planned to release a camera with either 1/2 inch or 2/3 inch sensors that had a port that would allow you access to full 4:4:4 data output (the camera itself recorded to a compressed format). This was to be a model in the ~$25k range. I don’t know what ever happened to that. Apparently it never materialized.
I have to say that I find the question of “What kind of imagery?” as odd. My experiences with keying have shown me that, short of a talking head, you need full color depth if you hope to get any kind of decent key from a chroma reference for most motion footage. If you work at it enough (re: roto spline masks) you can pull luma keys, but of course it’s tedious. Even with a talking head you sometimes need at least 4:2:2 depth.
At this time I don’t actually have any keying projects, I’m merely researching so that I have a better perspective on whatever resources I may have at my disposal down the road apiece when things hit the fan. It’s a bit depressing to see things are still locked in to big budgets for full depth keying capabilities.
-
Thanks for your reply. I believe the Black Magic interface is 4:2:2. That can work for some things, but certain types imagery need full color depth. Sometimes you can pull off luma mattes, but that’s tricky. Sometimes you need full-depth chroma.
Thanks again.
-
Yeah if you can’t sync the cameras keep ’em running. Even if you have to change tape it’s no big deal to sync as long as they’re running. Definitely get a house mix on at least one camera. I once had to sync stop and go footage off 3 cameras once and it took over 16 hours just to sync! Keep ’em runnin’…
-
Thanks for your reply.
Are there no simpler alternatives or alternate approaches to shooting full depth color for keying?
-
Typically speech recognition software works best when it’s “trained”. This usually involves reading predefined test that the program “knows” and uses it to decipher your speech patterns and create a profile of you.
Needless to say, this is next to impossible when doing videos like you do, unless you can somehow get the speakers to read at least a short verse the program can analyze. Even that sounds like a long shot.
I’ve never worked with Dragon, so I don’t know how “smart” it is out of the box. It’s been a while since I’ve dealt with speech recognition, and it would be very nice indeed if a program like Dragon can work right out of the box, or if Dragon has the ability to be “taught” from existing audio instead of predefined text reads. Then it would be a great boon to speech recognition indeed.
-
>>We get the best result exporting as flv @ 2mb/s (2000 kbps) 720P…<< Does that allow you a 10-minute video at 720P? What audio settings are you using? Just curious.