[Matthew Griffin]

Yeah, I’m trying to avoid nested sequences like the plague — everytime I try to use them in a fancy way they become the root of headaches.

Some folks setup a shortcut key or script to force copy-paste instead of nested sequences as subclips — I can’t remember how to do that.

But Andreas’s method shouldn’t require them as I’ll be sending back and forth “selects” sequences that include that while offline for the annotator can be directly reconnected by me on the receiving end. Provided the code remains solid.

Here’s what I mean by selects reels: the footage that seems the most useful, along with a few alternative shots — as little as I can stand for rough translation, and as much as I need to make the edit decisions.

And I think I’ll be very old-school and divide my editorial sequences into under 20min reels. The sound and music folks will thank me — they can get to work sooner with some locked reels!

And will go ahead and let the director + producers stamp markers on the V2 lowres clip as an additional layer of metadata that I can investigate without messing up my translation details.

I talked to four editors who did the traditional V3 text generator subtitles, who constantly have to turn that track off when the number of realtime effects hits the limit that causes FCP to force all of them to render. Switching from dropshadow to outline seems to be a good workaround to reduce the processing overhead, but still this sounds like a disappointing way to do things now that we have the powerful tool of XML to track and re-route metadata of all sorts.

[Matthew Griffin]

Andy,
Well, the reason I’m experimenting with the XML routes are the possibilities of using a tool other than FCP in the mix for the director to translate. And he is running 6.0.1 or 6.0.2 rather than 6.0.3 like I am.

But mainly the issue is that there is a LOT of footage and I’ll only have selects ready a few at a time. So the 1 XML to 1 sequence seems like a canny workflow. It just seemed clearer.

Well….

Actually, that’s still not a reason not to send a series of FCP project files, particularly if I prep them like Andreas suggests. I’ll admit that I’m very curious about the XML elements and want to learn more about it — a lean XML file archived with a lowres video clip seems somehow more “elegant”

So you know what? I’ll do both: xml export-then-import and new FCP seq per selects sequence — Andreas’ double-tracks as way to work.

Have you tried something that works better?

I’m appreciative that you have kicked in an option that cuts out all extra software before building the final titles — like a paired-coding, someone reminding me that there is a clearer, tighter way to do things.

Matt

[Matthew Griffin]

I really like this last idea. I noticed somewhere else someone used a trick to copy all of the markers to the various angles of a multitrack clip by marking one, duplicating it the number of times for the other angles, and then Replacing the duplicates with the other angles.

This has the advantage of ensuring that there are no TC change issues that send the markers whizzing around within the clips.

Then, with the reconnected clips back in my sequence, ready to be copy/pasted from the selects sequences into the editorial sequences, I will have the translations I need as clip markers and won’t have to render all the time.

So all that remains is choosing the best of the solutions for going from Markers to a subtitle generation software — I am testing SubBits and would love to look at your XML routes, AnnotationEdit, BelleNuit, and a few more options to pick what works best.

Has anyone had experience going from Clip Markers to subs? Obviously, Andreas has the XML chops to get what he needs from metadata, but I’m not sure markers is the usual route.

I tried to give the demo of Traffic a go but … haven’t had time to figure it out. Pretty cool, though.

And the additional great thing about using your suggestion, Andreas, is that the director can also mark other notes on the lowres clip making this for a great method to exchange both annotations and translations without creating extra work for anyone.

I am going to test and report back here.

[Matthew Griffin]

Essentially, FCP believed the lognote that the audio was 30fps and performed a pulldown to play each cluster of samples more quickly so that the TC Rate would match 25fps. Unfortunately, this was an unnecessary compensation so the 30fps stuff (for long clips) finished slightly early on the longer takes (barely noticeable otherwise).

At least I think so — I’ve just fixed this problem with the help of this thread and can’t check again to make sure.

Anyhow, this action wasn’t native-to-seq-settings playback so there were green bars on the audio track suggesting to render them for best performance.

I didn’t actually solve why this problem appeared, but the solution to import the audio and prep it in a vanilla PAL project worked perfectly!

Matt

[Matthew Griffin]

I’ll be darned. going to a basic PAL Easy Setup and project worked! all files came in as 25fps with no problems, and I could make them dual mono and drag them into my main project without a problem.

So thanks so much! I somehow got caught up trying to fix this darned document instead of dropping it and doing my work in a project file with no history — good idea going forward!

Matt

[Matthew Griffin]

Thanks for the recommendation. Am taking a look at your solution this morning!
Matt

[Matthew Griffin]

This does sounds like a good option — especially to get me through the editorial process before fussing with subtitle appearance issues.

Thanks!
Matt

[Matthew Griffin]

ah. I’m afraid I don’t know what it was — I think the recording device is the Tascam compact flash recorder. In any case, it recorded .wav not broadcast wave files on a compact flash card, which the recordist delivered to me with as a folder each for the shoot days. There are xml files for markers.xml and settings.xml but in this specific case these things don’t matter because there was no smartslate or tc generated by camera.

So these are just wave files — 48k 16bit — lacking any timebase. For some reason FCP has decided to give them a timebase instead of following the protocol of the project’s 25fps timebase. 90% of the files come through as TC Rate 25fps just fine. Those that are marked as 30fps are then out of sync with the audio tracks with picture because they are being falsely pulled down for 25fps.

I’m basically hassling with this now so my sound team doesn’t have to hassle with it when I deliver these frustrating out of sync dialogue tracks cut into seq. Trying to get the world truly back to 25fps timebase family.

Thanks for your curiosity! Does this help suggest a solution?

[Matthew Griffin]

Definitely 48k. Not sure what that would effect Timebase, but I am very happy to check anything you folks suggest!

Some of the post-sync tracks are mono, some stereo — I cannot determine any logic by which it chooses which files should stick stubbornly to TC Rate of 30!

I am tempted….with much back up….to change the TC Rate element by hand in the FCP Project files opening in BBEdit or similar. That is likely to be a bad idea, but if I can’t think of anything else, I’d rather do an annoying process 45 times by hand and know the issue is solved rather than keeping these 30fps Timebase audio files in Merges where they sound like they are drifting out of sync.

Thanks for suggesting I check the sample rate. Other suggestions?

[Matthew Griffin]

I didn’t know that I could do “clip modify” on a number of tracks at once — thanks for the help!
Unfortunately, some of these tracks are coming in as 30fps instead of 25fps — these are ones that were in the project when I re-opened it in 6.0.3. I need to resync these anyway, so I’m going to just blow out these files from the project and see if that helps.

But if there were to be a secret setting for correcting the 30fps back to 25 fps I wouldn’t have greenbars on those audio clips in the first place!

matt