New FCP Multicam Workflow
There's another thread on the problems with multicam editing, and as I've come up with a reasonable solution, and that thread has become unnavigable, I'll post the workflow here.
Basically, the issue with FCP's multiclip function is that it depends on having single clips for each camera angle, i.e. camera's can't start and stop. But unless you're shooting in a very controlled environment, that's simply not possible.
Invariably, you end up with a bunch of synched clips on a timeline, one track per camera, and any number of audio tracks recorded separately. But there's no way to edit it. The only way to hack a normal multicam timeline into FCP is to export every single track as a QT file and reimport each track, and turn them into a multiclip. That results in massive duplication of data, and is so manual as make tape-capture seem fun.
I've wrecked my brain for several days for how to use something like QuicKeys to make this tolerable, and here is what I think is the best solution.
In this workflow, you have all your sources on e.g. tracks 1-9, and you build your final edit on track 10. If you have 4 tracks source, you build your final edit on track 5.
You make a QuicKeys macro for each source track. For track 1, the macro is:
Cmd-Shift-A: Deselect everything
Q: Move to the In point (must be mapped, I use Avid mapping. Substitute your own mapping).
Cmd-Keypad-1: Enable Auto-Select on track 1
Cmd-C: Copy to clipboard
Cmd-Keypad-1: Disable Auto-Select on track 1
Option-F9: Enable Auto-Select on track 5 (*see below)
Option-F9: Disable Auto-Select on track 5
Q: Move to the In point (again, change to your key mapping for that feature).
G: Clear locators (again my Avid mapping, use your own key for that).
The reason that I'm using Option-F9 to select the top track is that the top tracking might be #5 in one timeline and #15 in another. But I don't want to have to change that in QuicKeys, so if your top track changes, you only have to change a single key in FCP's key map.
You then make the same script for each source track, only substituting Cmd-Keypad-1 for Cmd-Keypad-2 etc.
This macro is mapped to Ctrl-Keypad-1 through Ctrl-Keypad-9 in QuicKeys.
Before you start editing, you make sure that all Auto Select buttons are disabled.
Then, to view source footage, you select the source track and hit Ctrl-S, which soloes the source media. Always make sure to Ctrl-S again before you move on to another track, because the Ctrl-S feature literally changes your Track Enable/Disable function.
When you find some footage you like on Track #2, you set in and out points, and then hit Ctrl-Keypad-2. The footage between the locators is copied to the top track.
Then while viewing the top track, I want to cut into something, so I make an In point, then Ctrl-S to solo Track #4. I set an Out-point, Ctrl-S again to go out of Solo, and Ctrl-Keypad-4 to copy it up.
All the time, you simply Ctrl-S to view source footage (Ctrl-S again to get out of Solo), set In and Out points, and Ctrl-Keypad-# to copy it up.
And so on. This is actually quite fast to work with.
As a little extra note, it works well to label all the clips on track 1 simply '1' and all clips on track 2 simply '2', because then the final edit up on track 10 simply says 1-3-2-3-1-5-2.
Hope you like it!
Your workaround is interesting and I hope it helps others in similar situations but, let's be clear, and let's also understand that you do not need to take anything said here personally, it's not Apple's problem.
Multicamera editing in FCP is a bit cumbersome (and I hope they fix all the silliness of the setup and improve the interface someday) but, even with the warts, it works spectacularly well. You seem to be insisting on cramming an unusual function on top of a perfectly useful and functional feature that was designed to accomplish a specific but quite different task.
If multicamera editing is a function you require on a regular basis, you should work hard to acquire appropriately synced footage. Folks have been shooting multiple camera events for a couple of decades; the tools exist, the resources are available, the workflow is well-established, the conventional techniques are completely bombproof. Varying from those established systems is not really necessary, it's a conscious and probably misplaced decision. And your fix-it-in-post Quickeys routine, however cool it is, is still a post production file assembly and modification hack,and it carries the inescapable price tags of inefficiency, effort, expense, and brittleness.
Process design is never about discontent or gut feelings. It is always a question of metrics. Is the new version of a process easier, safer, cheaper, or more efficient than the previous version or of the Dumb Old Way of Doing Things? If the answer to some of those items can be assessed objectively as positive, then you have a process that is a candidate for adoption. However, all good new processes are subject to redesign. You must reexamine your process in a few months to see what can be improved.
I can tell you this: the procedures in place for shooting multiple cameras are in place because they have been proven to be effective, efficient, cheap, and utterly reliable.
No, you are absolutely dead wrong. It's is practically impossible to acquire footage where no single camera ever breaks, especially when you're tapeless, with various crews roaming. It's just a fantasy, and it doesn't matter how much you idealize it. You can ONLY have footage this well synced when you completely control the event, and can ask EVERY SINGLE CAMERA to stop, and ask the EVENT to stop for syncing.
It's a total fantasy.
Sorry Per but your dead wrong, we do it every week, we record a 3 hour program twice each weekend directly from our cameras into 3 editors plus 2 Xdcam decks that can record only 45 minutes at a time. When the footage from the Xdcam breaks we take the next clip make a cut in the multicam timeline and drag it over the angle or camera in the Viewer that needs to be preplaced, a small window pops up and gives us several options one of which is replace it. Simple as long as you don't have to do this 100 times.
It's still a highly controlled circumstance, you seem to be recording straight into external decks as well, and you seem to be able to stop the production for sync.
That's fine, if you completely can stop the production and there are no live moments that only happen once, unpredictably, and can't be recreated, then yes, you're totally fine, that is certainly ideal.
But take any reality-type programming, and you no longer control the event. To ask the event to stop for sync is the equivalent of asking a bank robbery to pause while you change P2 cards and resync.
Stopping the event is the same as losing the event. Period. It's very simple. It's like a photographer who takes the picture long after everyone has lost their natural smile.
Therefore, you instead just have a bunch of coverage, and you can't know when it starts or stops. Invariably, all this ends up on a timeline where each track is an angle, and everything is synced up, but there's no way to edit it.
I'm not extrapolating my reality-type programming to encompass your highly controlled programming, so please don't extrapolate in the other direction as well. I fully understand that you can use multiclip if every angle is one clip.
But it's bizarre that it's so hard to understand that reality-type programming by its very nature records an unpredictable event, and you might be recording for over 20 hours before it happens, but when it does, a camera has to be rolling, because you can't stop and ask the event to happen one more time.
Therefore, all you can do is keep as many cameras rolling as much of the time as possible and get a separate multi-channel audio-recording, and then you KNOW that you have coverage of any key event that happens.
I don't know of any camera that can record for hours on end without ever stopping -- possibly a P2 workflow with some MASSIVE real-time card-swapping regiment. Maybe some gigantic TV-station multi-channel HD recorder and wireless HD feeds from the cameras. But this is getting ridiculous.
Read the forums, you'll find an endless list of people having the same problem. I'm trying to solve it.
And by the way, I will definitely try to squeeze as much coverage as possible into the one-clip-per-angle regiment, because clearly that has obvious advantages. There *are* parts of this shooting that I *can* control, so at least I can benefit from multiclip for those circumstances.
By the way, Premiere Pro allows you to edit multiclip with a sequence as a source, but limits it to 4 angles, probably for no real reason other than picking a number. While I don't have the balls of steel to do a large project in Premiere Pro, I think that there's some realization in their choice of doing it like that.
I've read an article where someone describes at great length a technical workaround that makes it possible to have multiple clips per angle in Avid.
So it's actually only FCP where it's completely impossible without exporting and reimporting each single angle with duplication of data.
Whether Avid or Premiere Pro, their implementations have enough drawbacks that it doesn't make sense to for example switch back to Avid.
But as David points out, there are improvements more desperately needed, so this feature is unlikely to see improvement.
Did you look at Multiclip Sequences in the FCP manual , as I suggested in the Avid forum?
It seems like an attempt by Apple to address your problem (Avid has no such option)
just curious, as I have never needed to use it.... I do a great deal of multicam editing, but it is of the more conventional nature.
I think you should write a detailed feature request to both Apple and Avid about this...the advent of tapeless media is changing how things are shot...it would be cool if the tools would change too. It would not seem to be impossible to implement if they were aware a sizable group of people wanted it.
Yes, I'm writing a detailed feature request. My hopes are low, though, as they probably should be, but then at least I tried.
As someone who multi-cam edits on FCP all the time I have to disagree with you on this.
When it comes to multi-cam I am one of the lucky ones - most of the time I work with material shot in the ideal situation you described.
But often enough to cause what grey hair I still have to go white I come up against the same problem Per described.
The multi-cam feature in FCP works very well and is quite elegant in many ways -- as long as you have the right material - but an editor can't control the material he is given and while FCP's "clip" based multi-cam is ingenious it offers exactly zero improvements over a timeline based multi-cam, such as Premiere Pro has now or Discreet's now orphaned Edit* had 5 years ago. There is nothing a timeline based system can't do that a clip system can, but there are many things you can't do with a clip based multi-cam editor.
I wish apple would re-think their approach. It's not at the top of my wish list, access to multi-core rendering from within FCP has that honor, and I doubt that Apple will do anything. But to claim that FCP works "wonderfully well" in all situations and that Per is wrong in his analysis does not make any sense.
If anyone is still reading along in this thread, I'm loving this QuicKeys workflow, it actually results in nearly natural multicam editing.
Look at this picture:
The workflow is that you simply Ctrl-S to view the footage on the lower tracks (the angles), then you Set In, Set Out, and Ctrl-Keypad-1 (or the number of the track), and it's copied up on track 5.
With Show Duplicates enabled for the timeline, a by-product is that you can even see which source is currently showing.
I've worked with this for a little while and I'm picking up some speed, so I'm definitely going to do it this way. I can have any number sources, they can start stop as they please, and they can be any mix of formats. Yay!