Multiple LTC streams || Wordclock for a/v sync

[Juan Aboites]

Hi List,

Firstly thanks to Chris and to all developers and users for making such great great and reliable software - been using Qlab for years with great success.

I'm looking for approaches for an installation which may have multiple playback timelines. A separate vendor is handling video, so I'd like to see if a single workspace could accept two or more separate streams of LTC (one per cue list/input channel). I tried this out by sending TC from Qlab out physical outputs and then back in, but only one cue list seems to accept TC at a time (the other stays waiting). Since we're not talking about transport following, this should presumably work?

Also, has anyone had any experience maintaining a/v sync over a longer program with two machines that share only word-clock? So if video is being run on a watchout system and audio in Qlab, would a common clock be sufficient to keep the two transports together?

Apologies if any of this is unclear. Many Thanks!

Juan

[Pierre-Luc Brunet]

Hi Juan,

I was just thinking about that myself as I had a similar need for a show recently. With the case where you have Watchout running video and Qlab running audio (using the same word clock) it's hard to say if it would stay in sync but I'm not convinced it would. We would have to know how Watchout handles its clock timing and know if it handles it the same way QLab does. There is no actual system there in place to make sure things will play at the same time.

Is your show(s) gonna be the exact same every time or will there be live elements and different timings from time to time?
If it's going to be an exact replica every time, you might want to run your audio on software like Ableton Live, Pro Tools, Logic or Digital Performer and run timecode to them to make sure audio and video will stay in sync over long periods of time. If you need to throw in unpredictable live elements while having a base that is the same, then a combination of QLab, Lockstep and Live could work.

Live can chase MTC and also receive midi triggers to launch other unpredicted audio or midi tracks. From Watchout you can output LTC though you will have to convert it to MTC for Live to take. For the triggered events, you could send a UDP signal from Watchout to QLab which would then convert it to Midi to go into Live.

Watchout can also play multichannel wav files so that might be another option (which could be simpler.)

But now the first thing to know, is to know more about the installation.
Is your show(s) gonna be the exact same every time or will there be live elements and different timings from time to time?
Are there multiple rooms to deal with? Is this why your bringing multiple timelines?

Let us know.

[Juan Aboites]

Hi Pierre,

Thanks for your notes. My responses are inline:

But now the first thing to know, is to know more about the installation.
Is your show(s) gonna be the exact same every time or will there be live elements and different timings from time to time?
Are there multiple rooms to deal with? Is this why your bringing multiple timelines?

 

Yes, it's an installation with several rooms. Only one room plays a straightforward a/v program which does not change, but it does require sync to be within a 1/4 frame. Ultimately playing directly from Watchout might be the best solution, but I'm just thinking of ways of streamlining the workflow in the install/mixing stage. For that reason I was wondering if advocating for a common wordclock between both systems would achieve this -- Very well might try Live as you suggest, Pierre (PT is out of budget I think).

 

We would have to know how Watchout handles its clock timing and know if it handles it the same way QLab does. There is no actual system there in place to make sure things will play at the same time.

 

How does Qlab handle clock timing? In other words, what's required of any software to duplicate how Qlab handles scheduling in this sense? I'm assuming that If I were to play separate video and audio files together within Qlab, both would resolve to the same clock (the audio interface one?)...

 

Are there multiple rooms to deal with? Is this why your bringing multiple timelines? 

The other rooms have programs that don't change, but when they start does (there are motion/proximity sensors). The programs are short and don't involve video (mostly lights), so sync isn't really an issue in the same way, they just need to start together. No issues there, as LTC triggers would work fine for this, but there will likely be several rooms and so instances where different ranges of timecode would run simultaneously (room 1 starts at 1:00:00:00, room 2 starts at 2:00:00:00 etc). In my limited experimentation, Qlab would only listen to one LTC stream at a time (I tried sending LTC to different Cue Lists on different inputs). Do I need to use different trigger types for each room?

I hope this is clarifying!

Many Thanks!

Juan

[Andy Lang]

On Thu, Oct 2, 2014 at 11:47 AM, Juan Aboites <jab...@gmail.com> wrote:
> How does Qlab handle clock timing? In other words, what's required of any
> software to duplicate how Qlab handles scheduling in this sense? I'm
> assuming that If I were to play separate video and audio files together
> within Qlab, both would resolve to the same clock (the audio interface
> one?)...

This is correct, with the caveat that the video file must have an
audio track in it (it can just be digital silence), and it's audio
must be patched to the same audio device as the audio cue. Then, as
long as the cues are part of the same sequence (i.e., part of a "start
all" group or linked with auto-continues), they'll be locked together
to the audio device's clock. If the video does not have an audio
track, it clocks to the computer's system clock, so there's no
guarantee of sync.

QLab is not capable of slaving to external timecode, for a number of
reasons, so it can only use incoming LTC as a trigger to start a cue.
Once that happens, the cue is freewheeling with regards to the
incoming LTC. For this reason, we generally recommend that QLab be the
LTC master, since (if it's patched through the same audio device) the
LTC and audio will have sample accurate sync.

I don't have the experience with Watchout to know if it's capable of
varispeeding to chase incoming timecode or not, so I can't answer if
it will work going in that direction.

I'm comparing notes with my teammates here on the multiple LTC
question, so stay tuned for more on that…

Thanks!

-Andy

[Pierre-Luc Brunet]

 For this reason, we generally recommend that QLab be the 

LTC master, since (if it's patched through the same audio device) the
LTC and audio will have sample accurate sync.

That's a good idea. If you're able to set things up this way and obviously test them before hand, this might be a good solution for you. 

This way you can also send multiple timecode streams to different outputs simultaneously. You could also burn your LTC timecode into fixed audio track and play them back as regular audio files in QLab. And to trigger the lighting consoles, you could also use good old Midi Show Control being sent from QLab.

For your Watchout/Audio sync room, by burning LTC to fixed audio tracks you could also put them into Live or another DAW, using your required audio sequence session but adding that one LTC track that is being sent to Watchout so it can follow it when you play audio. Basically have your editing audio software with the audio for this video and add the LTC track to it. Then your DAW become the Timecode master for creation (but it could also be for the whole installation) or you could later on export a multi-channel audio file of the video's audio + the LTC track and play it in QLab on cue.

Many possibilities. Are those good solutions for you? This sounds like fun!

[Juan Aboites]

Many possibilities. Are those good solutions for you? This sounds like fun!

Should be fun indeed! Some combination of these will get us what we need ... Thanks all!