Guaranteed Sync

[Jason Crystal]

Hello QLabbers,

It's been awhile since I've done a show where I need a group of cues to be guaranteed to sync with very high precision. Back in "the day", this involved some pre-load cues, some checking of "guarantee sync" checkboxes, and various other hand-waving.

What is the current (QLab v4) status of multiple tracks playing in exact sync? Specifically:

  • In a Fire All Simultaneously group, can I assume all audio tracks start precisely at the same time?

  • In a Fire All Simultaneously group, can I assume internal pre-waits will behave identically from playback instance to instance? In other words "pre-wait 1.0 seconds" will always cause the pre-waited cue to enter at the same moment relative to cues that began at the top of the group?

  • Is there any latency introduced by nesting additional Fire All Simultaneously groups within Fire All Simultaneously groups?

  • Is there any need to pre-load a large group of Audio cues that need to play synchronously?

This assumes all files on the same drive a the same sample rate, etc.

Thanks! Just updating my personal firmware.

-Jason

[Chris Ashworth]

Hi Jason,

Nice to hear from you!

Answers inline below:

On October 31, 2018 at 2:18:19 PM, Jason Crystal (jcry...@gmail.com) wrote:

Hello QLabbers,

It's been awhile since I've done a show where I need a group of cues to be guaranteed to sync with very high precision. Back in "the day", this involved some pre-load cues, some checking of "guarantee sync" checkboxes, and various other hand-waving.

What is the current (QLab v4) status of multiple tracks playing in exact sync? Specifically:

  • In a Fire All Simultaneously group, can I assume all audio tracks start precisely at the same time?

Yes - with the important note that this applies only to audio tracks that start at time zero within the group. (Any audio in cues with a pre-wait will not be subject to the same sync guarantee.)

You may want to enable the auto-load option for the cues to help ensure it happens as quickly as possible, but even if they are not loaded they will (eventually) start at the same time and remain in sync.

  • In a Fire All Simultaneously group, can I assume internal pre-waits will behave identically from playback instance to instance? In other words "pre-wait 1.0 seconds" will always cause the pre-waited cue to enter at the same moment relative to cues that began at the top of the group?

It will generally be within a few milliseconds, but it will not be guaranteed to be the exact same sample each time.

  • Is there any latency introduced by nesting additional Fire All Simultaneously groups within Fire All Simultaneously groups?

No latency is introduced by sub-grouping. (To guarantee sync within the sub-groups, no pre-wait is also required for synced cues in the sub-groups.)

  • Is there any need to pre-load a large group of Audio cues that need to play synchronously?

Possibly, but not for sync reasons, just for speed-to-start reasons.

This assumes all files on the same drive a the same sample rate, etc.

Thanks! Just updating my personal firmware.

-Jason

Cheers,

C

[Jason Crystal]

This is all brilliant. Thank you!

-Jason

[micpool]

Here's an interesting experiment. I've only done it on one machine so your mileage may vary. Also results may be affected by order of workspaces etc. so these are by no means definitive results for all systems. 

I think most of the results are in line with Chris's description of what is meant to happen. QLab2 may be maintaining Sync between all cue types at the expense of starting speed, dropped frames at top of cues, etc.

I created a workspace in QLab 2 with some audio files (a stereo wav, an mp3, and a 24 ch wav on an external firewire drive) and some video files (a Prores file with stereo soundtrack on an external firewire drive and an h264 without soundtrack)

I then made QLab 3 and 4 versions of these workspaces and set the MIDI controls to control them all simultaneously.

QLab 2 audio files have guaranteed sync checked.

Experiment 1; No Preloading

QLab 2 has taken an additional 1.5s to start but has all the files bang in sync.

QLab 3 +4 Started quicker but have a 3s anomaly between files with a soundtrack and those that don't (including h264 without soundtrack)

Experiment 2 No Preload /10 Pause Resume cycles

QLab2 a second behind but all files in absolute sync

QLab 3+4 anomaly between tracks with soundtrack and  those without.

Experiment 3 All preloaded

Practically identical start times and sync reasonably tight in all versions.

QLab 2  marginally fastest to start and all tracks in exact  Sync

QLab 3 all tracks in exact sync

QLab4 Tracks with soundtrack in Sync but 60ms behind tracks with no soundtrack

(No audio device overloads in logs)

and finally here are 2 QLab 4.4.3 workspaces 

The workspace on the left has all files on a single SSD, the workspace on the right is mixed drives

The all SSD workspace now has a small  discrepancy between Wavs and mp3s which remains even if they are preloaded. 

Mic