Weird issue - separate Video machine jump-completes cues

[FL K]

Hi wonderful community (and Figure 53),

I preface this message as being a preliminary step to poll if anyone else has experienced this particular problem (as I have not), before going down the support road; and that sadly, I was not on site when it happened, so I am relaying the story with the information I got from the stage manager during the show, as it happened;

The scenario:

- A primary QLab 5 workspace on an intel iMac (exact details like software version tba when I can get into the space next) is calling

- a video QLab 5 workspace (5.6.3) on an M1 Mac Studio 24Core CPU/60Core GPU/64

G/1TB, outputting a single stage via syphon to Madmapper for output,

- using explicit network cues via a QLab 5 network patch using "start cue and load next", in a list of top-level timeline cues all numbered with the cue numbers being called from the primary workspace.

- the video content is usually a video that starts with the completed static background of the previous scene, contains the transition and builds the state for the next scene (hold last frame)

- the fading out/stopping of the previous scene is managed on the video machine, sometimes via explicit fade-outs, but mostly via "fade and stop peers over 0.5 secs", and the new video being in the same layer, but because being called later, being on top.

- This scenario has worked reliably and without any issues for all of the tech and the whole first week of performances

- yesterday, the SM came back for the second week of performances, booted up both machines and did some pre-show checks and ran cues up until scene 2, apparently all worked fine, then set to pre-show and ran the show. around scene 3 she noticed that cues did not go through their transition but when called, showed up complete (in the active cues window with the progress bar showing the cue completed.

- I advised her to fully panic the video workspace in an opportune moment and re-run the cue (or run the next) - but the problem persisted; whenever a cue was fired on the video Mac, it jumped to completion, and she would manually drag it (in the active cues window) drag it back to start quickly, so that the transition could take place visibly.

- Like this, she limped through to interval (it was decided that this was more desirable than a show stop, not here to discuss the politics of that), where she rebooted the video Mac, went through all the same steps of setting it up, tested it, and the problem was gone; it worked without issue until the end of the show, and after clearance, she cue-to-cued through the beginning of the show with both Macs with no apparent issue.

While this might be generally good news for the remaining shows (indicating a strange outlier rather than something that keeps happening frequently), the disconcerting thing is that a) the problem did not show itself in tests pre-show, b) did not have a discernible cause, and was c) nothing that I ever experienced in my long use of QLab before.

So just thought I'd ask around here, in case someone has come across this issue before but was able to detect a root cause, and might not have reported it to support - please enlighten me :). I will separately also report to support now.

Cheers,

Freddy

[Paul]

I have occasionally seen video cues not fade in when they should; this was during programming the show and I would delete the cue and remake it, as life was too short to figure this out.

I would make all the video fades explicit instead of relying on "fade and stop peers".  Sounds like it may have got confused about what "peers" were in this instance (cues in the same group?); making an explicit fade cue means the SM can see the fade working (or not) in the active cues panel. Removes a little of the uncertainty. 

You can also check the logs (on either machine) if you enable OSC logging in Workspace Status (CMD-Shift-W) and if necessary the SM can copy the logs to save them elsewhere. Although in this case I don't think it's a network issue.