|
|
All the news now and in your digital future
Dupe Frame Discovery
Posted by FilmBase
|
Re: Dupe Frame Discovery August 24, 2008 07:10AM |
i'd love to take you up on that, one day!
i like Belgian beers, preferably those brewed by Trapist monks.
suits our profession, don't you think?
this process of converting the HDV to some other codec on ingest means you cant Batch Capture.
i have to say that here in the early days of the 21st Century that seems a little bit primitive, but there it is.
because the file you created in this process cant be batch captured from the original tape,
FCP is "doing you a favour" by not giving it a reel#.
the bigger truth behind this is that the file doesn't have a proper timecode track... the absence of a reel# is merely a symptom.
giving the file a reel# in FCP kind of "kick-starts" a TC track within the file, starting at zero.
without a TC track, which (to be pedantic), manifests itself as lack of a reel#, FCP treats the file as if it DIDN'T come from tape,
as if it is the only instance of that media in the universe, which in a sense it is.
so if you ever came to MEDIA MANGE that file, FCP would NOT trim it of any excess media, as it believes it should not destroy something so precious.
FWIW, by giving the media a reel# you are giving FCP the go-ahead to trim media in any MM process.
i take it youlve been doing this for a while.
captureng your HDV as AIC or ProRes,
cutting away,
and then exporting, delivering, and possibly blowing off the media and starting all over again.
if that's the case, and if you ever needed to re-visit one of your projects,
you'd sort of need to jump thru a hoop or two to re-build it.
but we'll cross that bridge when we get to it.
and then you can owe me another beer!
cheers,
nick
i like Belgian beers, preferably those brewed by Trapist monks.
suits our profession, don't you think?
this process of converting the HDV to some other codec on ingest means you cant Batch Capture.
i have to say that here in the early days of the 21st Century that seems a little bit primitive, but there it is.
because the file you created in this process cant be batch captured from the original tape,
FCP is "doing you a favour" by not giving it a reel#.
the bigger truth behind this is that the file doesn't have a proper timecode track... the absence of a reel# is merely a symptom.
giving the file a reel# in FCP kind of "kick-starts" a TC track within the file, starting at zero.
without a TC track, which (to be pedantic), manifests itself as lack of a reel#, FCP treats the file as if it DIDN'T come from tape,
as if it is the only instance of that media in the universe, which in a sense it is.
so if you ever came to MEDIA MANGE that file, FCP would NOT trim it of any excess media, as it believes it should not destroy something so precious.
FWIW, by giving the media a reel# you are giving FCP the go-ahead to trim media in any MM process.
i take it youlve been doing this for a while.
captureng your HDV as AIC or ProRes,
cutting away,
and then exporting, delivering, and possibly blowing off the media and starting all over again.
if that's the case, and if you ever needed to re-visit one of your projects,
you'd sort of need to jump thru a hoop or two to re-build it.
but we'll cross that bridge when we get to it.
and then you can owe me another beer!
cheers,
nick
|
Re: Dupe Frame Discovery August 24, 2008 09:07AM |
Everything Nick says is right on, if a little bit vivid. (Crossing bridges and jumping through hoops; that stuff is way too energetic for a Sunday morning.)
We were just talking about this over in another thread: Final Cut, to its credit, lets you do a lot of things that you might need to do sometimes, but that you really shouldn't do unless you have to. "Capture now" from a controllable device is one of them.
If you wire your camera's SDI output straight into your Kona board, then "Capture now" is a dream. You just mash a button, and the live feed from your camera is going straight to your framestore in whatever format you chose, in real time. It's fantastic. But if your material is on tape, then it's kind of silly to use "Capture now," since doing so ? as Nick says ? gives Final Cut no information at all about where that footage came from, or how to get it back. If you "Log and capture," you tell Final Cut that this individual frame of footage came from tape so-n-so at timecode point such-n-such. With just those two pieces of information, Final Cut can, one minute or five years from now, batch re-capture that footage in any format just by telling you which tape it needs.
In fact, you technically don't even need to capture your footage at all! There's nothing stopping you from using "Log and capture" to log your footage without capturing it, telling Final Cut which bits of which tapes you want to use and having Final Cut treat those bits as offline media. I can't imagine why you'd ever want to do this ? it'd be like cutting a motion-picture negative in the dark, or editing your whole show using only Microsoft Excel ? but the point is that you can. The actual pictures on your framestore are irrelevant as long as Final Cut has the tape-and-timecode information.
This, of course, is how you can do an offline in Final Cut. You can capture the frames off your tape to some absurdly low-resolution format just so you can see what you're cutting. Because Final Cut knows exactly where those frames came from ? which tapes, and from between which timecode in- and out-points ? you can go through all the motions of editing with low-res material, then take your media offline and batch-recapture just the bits you used at full resolution for finishing.
That workflow doesn't mean nearly as much as it used to in these days of 20- or 25-Mbps HDV and multi-terabyte portable framestores, but it's still the model around which Final Cut is built. And it's still useful if, for example, you shoot an entire feature on HDCAM SR with a 20:1 shooting ratio ? that'd be around fifteen terabytes of footage if you captured it all uncompressed. In that case, you'd log-and-capture everything at low resolution, cut it into your timeline, then do a select-all-make-offline and batch-recapture just the parts you need, with handles. That's how features have been edited for years, and they're still cut this way today. (Except, of course, when you're cutting a film without doing a DI, your finished product is a low-resolution offline cut with an edit list for your neg cutter, and there's no online conform. Which sort of serves to underscore just why this workflow exists in the first place. The reason non-linear editing was first used for creative editor was to produce not a video of any kind, but rather an edit decision list for a neg cutter or an online linear videotape editor to execute on the actual footage.)
Technology has improved to the point where few of us will ever need to do a real offline edit and conform, because we're shooting compressed material and finishing right in Final Cut. But it still helps to understand why Final Cut works the way it does. If nothing else, it just makes a whole lot more sense if you know how it was originally meant to be used. Then you can make educated choices about whether to stick to a very traditional post workflow or to do things differently.
We were just talking about this over in another thread: Final Cut, to its credit, lets you do a lot of things that you might need to do sometimes, but that you really shouldn't do unless you have to. "Capture now" from a controllable device is one of them.
If you wire your camera's SDI output straight into your Kona board, then "Capture now" is a dream. You just mash a button, and the live feed from your camera is going straight to your framestore in whatever format you chose, in real time. It's fantastic. But if your material is on tape, then it's kind of silly to use "Capture now," since doing so ? as Nick says ? gives Final Cut no information at all about where that footage came from, or how to get it back. If you "Log and capture," you tell Final Cut that this individual frame of footage came from tape so-n-so at timecode point such-n-such. With just those two pieces of information, Final Cut can, one minute or five years from now, batch re-capture that footage in any format just by telling you which tape it needs.
In fact, you technically don't even need to capture your footage at all! There's nothing stopping you from using "Log and capture" to log your footage without capturing it, telling Final Cut which bits of which tapes you want to use and having Final Cut treat those bits as offline media. I can't imagine why you'd ever want to do this ? it'd be like cutting a motion-picture negative in the dark, or editing your whole show using only Microsoft Excel ? but the point is that you can. The actual pictures on your framestore are irrelevant as long as Final Cut has the tape-and-timecode information.
This, of course, is how you can do an offline in Final Cut. You can capture the frames off your tape to some absurdly low-resolution format just so you can see what you're cutting. Because Final Cut knows exactly where those frames came from ? which tapes, and from between which timecode in- and out-points ? you can go through all the motions of editing with low-res material, then take your media offline and batch-recapture just the bits you used at full resolution for finishing.
That workflow doesn't mean nearly as much as it used to in these days of 20- or 25-Mbps HDV and multi-terabyte portable framestores, but it's still the model around which Final Cut is built. And it's still useful if, for example, you shoot an entire feature on HDCAM SR with a 20:1 shooting ratio ? that'd be around fifteen terabytes of footage if you captured it all uncompressed. In that case, you'd log-and-capture everything at low resolution, cut it into your timeline, then do a select-all-make-offline and batch-recapture just the parts you need, with handles. That's how features have been edited for years, and they're still cut this way today. (Except, of course, when you're cutting a film without doing a DI, your finished product is a low-resolution offline cut with an edit list for your neg cutter, and there's no online conform. Which sort of serves to underscore just why this workflow exists in the first place. The reason non-linear editing was first used for creative editor was to produce not a video of any kind, but rather an edit decision list for a neg cutter or an online linear videotape editor to execute on the actual footage.)
Technology has improved to the point where few of us will ever need to do a real offline edit and conform, because we're shooting compressed material and finishing right in Final Cut. But it still helps to understand why Final Cut works the way it does. If nothing else, it just makes a whole lot more sense if you know how it was originally meant to be used. Then you can make educated choices about whether to stick to a very traditional post workflow or to do things differently.
|
Re: Dupe Frame Discovery August 24, 2008 11:45AM |
>the bigger truth behind this is that the file doesn't have a proper timecode track... the >absence of a reel# is merely a symptom.
>giving the file a reel# in FCP kind of "kick-starts" a TC track within the file, starting at zero.
Great point. A while ago there was a mention of subclips going out of sync or something and the fix was to insert a reel number. This may be the reason for that- the absence of a TC track.
>That workflow doesn't mean nearly as much as it used to in these days of 20- or 25-Mbps
>HDV and multi-terabyte portable framestores, but it's still the model around which Final
>Cut is built.
That is also the way many NLEs are supposed to be built, especially in the earlier days of AVID, where capturing at online resolutions was quite a dream for many and hard disks were small. On the other hand, many editors have been trained to observe TC integrity, mainly because of the offline/online workflow. It's not a bad practice at all, as maintaining accurate timecode allows you a back up route to recapture from tape should anything go wrong with the footage (file corruption, drive failure, etc).
To add on: Running Capture Now maintains the timecode on tape. Running Capture now (as a non controllable device) ignores the timecode.
www.strypesinpost.com
>giving the file a reel# in FCP kind of "kick-starts" a TC track within the file, starting at zero.
Great point. A while ago there was a mention of subclips going out of sync or something and the fix was to insert a reel number. This may be the reason for that- the absence of a TC track.
>That workflow doesn't mean nearly as much as it used to in these days of 20- or 25-Mbps
>HDV and multi-terabyte portable framestores, but it's still the model around which Final
>Cut is built.
That is also the way many NLEs are supposed to be built, especially in the earlier days of AVID, where capturing at online resolutions was quite a dream for many and hard disks were small. On the other hand, many editors have been trained to observe TC integrity, mainly because of the offline/online workflow. It's not a bad practice at all, as maintaining accurate timecode allows you a back up route to recapture from tape should anything go wrong with the footage (file corruption, drive failure, etc).
To add on: Running Capture Now maintains the timecode on tape. Running Capture now (as a non controllable device) ignores the timecode.
www.strypesinpost.com
|
Re: Dupe Frame Discovery August 24, 2008 11:55AM |
> Running Capture Now maintains the timecode on tape. Running Capture now (as a non
> controllable device) ignores the timecode.
Yes. Although I did have a strange thing this past week when an assistant used Easy Setup (Kona Uncompressed 10-bit SD), then changed Device Control to FireWire NTSC, played the DigiBeta tape and clicked Capture Now. The video captured at 10-bit, but the timecode was missing -- it started at 00:00:00:00, which clued me in to the problem. But what surprised me was that the capture happened at all -- normally if you don't have a deck attached by FireWire, I was under the impression that it should have refused to capture unless Device Control was set to Non-Controllable.
www.derekmok.com
> controllable device) ignores the timecode.
Yes. Although I did have a strange thing this past week when an assistant used Easy Setup (Kona Uncompressed 10-bit SD), then changed Device Control to FireWire NTSC, played the DigiBeta tape and clicked Capture Now. The video captured at 10-bit, but the timecode was missing -- it started at 00:00:00:00, which clued me in to the problem. But what surprised me was that the capture happened at all -- normally if you don't have a deck attached by FireWire, I was under the impression that it should have refused to capture unless Device Control was set to Non-Controllable.
www.derekmok.com
|
Re: Dupe Frame Discovery August 24, 2008 01:44PM |
> I shoot HDV to MiniDV
Just call it HDV. MiniDV is a tape that can store both HDV and DV footage.
Nick's point is interesting, as I was under the impression that FCP automatically adds a timecode track in the qt container.. So it doesn't without a reel name? Or are we talking about something else?
>What I discovered is the long-ass-pain-in-the-butt way that works for me.
The capture lags a little as the processor is doing the on-the-fly transcoding from HDV to ProRes. Ceteris Paribus, captures made under the same conditions and settings will produce the same result. Though I won't be sure if it is a reliable fall back in the long run, since a guy mentioned that the timecode seems different as compared to a capture with a controllable device (the firewire prores capture seems slower). Would macs with different processors/hard disk/interfaces result in non frame accurate recaptures, especially since timecode data isn't used to maintain frame accuracy?
www.strypesinpost.com
Just call it HDV.
MiniDV is a tape that can store both HDV and DV footage.
Nick's point is interesting, as I was under the impression that FCP automatically adds a timecode track in the qt container.. So it doesn't without a reel name? Or are we talking about something else?
>What I discovered is the long-ass-pain-in-the-butt way that works for me.
The capture lags a little as the processor is doing the on-the-fly transcoding from HDV to ProRes. Ceteris Paribus, captures made under the same conditions and settings will produce the same result. Though I won't be sure if it is a reliable fall back in the long run, since a guy mentioned that the timecode seems different as compared to a capture with a controllable device (the firewire prores capture seems slower). Would macs with different processors/hard disk/interfaces result in non frame accurate recaptures, especially since timecode data isn't used to maintain frame accuracy?
www.strypesinpost.com
|
Re: Dupe Frame Discovery August 25, 2008 03:29AM |
>I DO get timecode when I capture in ProRes and it does match what is on the tape
Hmm... I got to run tests on this when i get the chance to work with HDV again. Then again, frame accuracy could be likely, as the clip is broken up with scene detection. Very smart move by Apple.
www.strypesinpost.com
Hmm... I got to run tests on this when i get the chance to work with HDV again. Then again, frame accuracy could be likely, as the clip is broken up with scene detection. Very smart move by Apple.
www.strypesinpost.com
Sorry, only registered users may post in this forum.
