Smpte LTC issue using Dante DVS

[Gary Parke]

I am having a random issue using LTC with Dante DVS as my time code source. It is pretty random, and happens on a few cues. We have an LTC stripe on a Nuendo machine with is Dante patched into a Mac Mini using DVS. I am running Lockstep as well to convert from LTC to Midi for our midi devices. When the issue happens, the timecode display which qlab is showing is ahead of what Lockstep is showing even though it is the same mac mini machine with the same DVS ch selected. I have a cue list which listens to LTC via my Dante input. How is it even possible for the qlab LTC to be showing ahead? If I pause Nuendo as soon as the stripe happens, and then play, it shows correct. It is hard to reproduce but I managed to get a video of it. Link attached.

https://www.youtube.com/watch?v=xK3ijbvd0G4

Thanks!

Gary

[Anthony Narciso]

Just for the record, Dante technically doesn't support using LTC. In the Audinate training power point it specifically says it doesn't do it. Now, it does work just because of the fact that LTC is jan audio file that is being transmitted, but I image its unsupported because of latency from running over a network. 

[Rich Walsh]

I don’t think I agree with that statement as it doesn’t make sense:

1. LTC _is_ audio. So either Dante transports audio or it doesn’t. It does.
2. Latency is fixed ("deterministic") across the Dante network and defaults to 1ms, which is 1/33 of a frame at 30fps. I don’t think that should preclude the ability to sync…

If it were due to Dante then it would affect _both_ TC displays as they are both connected to Dante; so this is a red herring in this case.

I assume you’ve picked up on this from here:

What does Dante NOT do?

• Sample rate conversion
• Level control
• MIDI
• SMPTE time code
• These are handled by products
• Control and other data runs alongside Dante on network

I think this means that it does not _generate_ TC.

I haven’t explored the thought I had watching the YouTube video to check if it was just plain wrong, but I wondered if there was something about QLab not detecting the format of the timecode correctly and therefore miscalculating the time to display? I have a vague memory about timecode messages including a header to define the format, but maybe that isn’t sent every message? Is QLab always ahead, and does the amount that it is ahead increase as the TC goes up (ie: if it’s 5s out at 01:00:00:00 is it 10s out at 02:00:00:00)?

Rich

[micpool]

I agree with Rich. LTC is very robust, there is no reason it should't work with any digital means of transmission.

Lockstep and QLab have almost identical responses to incoming LTC, their threshold is at -30dB. They are more intolerant of pitch variation than hardware SMPTE synchronisers and have an edge case malfunction at 0.85 of normal playback speed.where the both their timecode readers will go haywire, but in an entirely identical fashion.

There is a difference in that, if the playback speed of the source drops to say .8,  QLab will go into freewheel and use an internal clock source (although flagging "signal received is not valid timecode") while Lockstep will stop receiving

Lockstep will freewheel under other circumstances, e.g receiving 2 timecode signals with different times and in this circumstance it's freewheel and recovery when valid TC is restored  is identical to QLab

The only other difference is QLab has to have the TC format set correctly to work, LockStep detects incoming TC format automatically.

So I am baffled by Gary's problem. I suppose a critical question to assist a diagnosis is, which program is correct?, i.e showing the same time as a reader on the Nuendo machine

Is this something that happens when the timecode has been running for 5 hours or is the high timecode value because that section of timecode has a high start value?

Mic

[Gary Parke]

what we are seeing is random spots in the show which get ahead. sometime 5 secs but usually no more than 15 or so. We use timecode values starting at the 2hr mark and then because of all the start / stops we end up around the 6 hour mark. If I were to Pause the Nuendo playback in the middle of a cue where the smpte was ahead in qlab , and then hit play again, it goes back in sync and is displaying correctly.. Lockstep is always right on, but for some reason, Qlab gets ahead - On Occasion. It is cannot always be reproduced.
I do know that it is not just a display thing, but the cues which have a timecode trigger on them, trigger to the displayed qlab timecode, which sometimes is incorrect to what the actual timecode value is.

[Gary Parke]

Mic - Lockstep is always correct. We distribute LTC to lighting and to media from the Nuendo Mac and to other various time code displays in the theatre. We also have a Horita time code reader in the audio booth for smpte confirmation and qlab is the only thing occasionally gets ahead.

[micpool]

On Saturday, March 10, 2018 at 12:08:12 PM UTC, Gary Parke wrote:

what we are seeing is random spots in the show which get ahead. sometime 5 secs but usually no more than 15 or so. We use timecode values starting at the 2hr mark and then because of all the start / stops we end up around the 6 hour mark. 

I don't understand what you  are describing here, why should the number of stops or starts affect the Timecode value. The only way you can be receiving Timecode with a value if 6 Hours + is if your stripe on the Nuendo machine is encoded with that time value isn't it. Or have I misunderstood?

Mic

 

[micpool]

If Lockstep is correct then rather than spend time diagnosing a very obscure problem, I would just switch QLab to use MTC and set its source to lockstep

Mic

[Gary Parke]

We start out at the 2 hour mark. Our show is about 2hr 15 long with about 45 or so stops in the playback, just so we can roll the nuendo music cues with the LTC stripe on word cues or whatever. We do always start the next music / smpte cue at least. 5 minutes ahead of the previous one ending in case we insert a music cue with smpte in the middle of the show. This way, we don’t have to restripe from the inserted point forward. makes it easier. That is why we go from the 2 hour mark and end up after the 6 hour mark.

[Gary Parke]

Mic - I tried switching over to Lockstep as my source with MTC, but it was not reading it. Do I need a loopback or something?

[micpool]

On Saturday, March 10, 2018 at 12:24:26 PM UTC, Gary Parke wrote:

We start out at the 2 hour mark. Our show is about 2hr 15 long with about 45 or so stops in the playback, just so we can roll the nuendo music cues with the LTC stripe on word cues or whatever. We do always start the next music / smpte cue at least. 5 minutes ahead of the previous one ending in case we insert a music cue with smpte in the middle of the show. This way, we don’t have to restripe from the inserted point forward. makes it easier. That is why we go from the 2 hour mark and end up after the 6 hour mark.

OK, that makes sense, you just have continuous timecode with your cues spaced along it.

Mic

[micpool]

oh yes, sorry I forgot you had lockstep MTC output selected to an interface for other MIDI hardware. The internal Lockstep output only works if that is the selected MTC output. You would need to loop it back

Mic

[Gary Parke]

Also as far as trying to use the Lockstep MTC on my cuelist, I have Lockstep sending MTC and other program changes out a presonus audiobox to other devices in the theatre for cues and misc things. I will mess with that some more on Monday.

[Gary Parke]

Thanks. We are replying at the same time :) I will look at it some more. would be nice though if the LTC would be solid though.

On Sat, Mar 10, 2018 at 7:43 AM Gary Parke <glp...@gmail.com> wrote:

Also as far as trying to use the Lockstep MTC on my cuelist, I have Lockstep sending MTC and other program changes out a presonus audiobox to other devices in the theatre for cues and misc things. I will mess with that some more on Monday.

--
Contact support anytime: sup...@figure53.com
Follow Figure 53 on Twitter: https://twitter.com/Figure53
User Group Code of Conduct: https://figure53.com/help/code-of-conduct/
---
You received this message because you are subscribed to a topic in the Google Groups "QLab" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/qlab/tK5bG7hDyDE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to qlab+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/qlab/fae6109e-015f-4d65-8239-e5e43dc5720c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Thanks
Gary Parke

[Gary Parke]

MTC does seem to be more solid. I did some testing this morning. The issue with LTC only occurs if you stop the incoming LTC (in our case with a smpte stripe on an audio track in Nuendo) and then go forward in time to a new audio cue and then start it up. If you locate backwards, somewhere earlier in time, then qlab always seem to be in sync. it is only when you jump forward in time, like we do in our show.

Thanks

Gary

[Chris Ashworth]

Gary,

Thanks for this report and for the YouTube video — I filed it in our issue tracker in hopes we’ll be able to investigate it at some point.  (Load is high just now finishing up 4.2, but it’s in the tracker now.)

-C

[Gary Parke]

Hi Chris

Has there been any luck with this issue? I have not tried switching over to LTC yet because I don't remember seeing anything in the change log, but I may have missed it.

Thanks
Gary

[Pierre-Luc Brunet]

Hi, I'm also very curious if there has been progress on this since I'm currently specing a system where I would like to use QLab generated LTC outputted over Dante.

Thank you.

[Nigel Hogg]

I use LTC all the time over Dante DVS into a Midas M32R then out on a bus into a Jands Vista lighting desk. I have had no problems at all. I’m running Qlab on a Mac Pro, High Sierra. The only requirement is to ensure enough level on the channel but below clipping.

[Gary Parke]

The issue I was having was how qlab receives incoming time code LTC on DVS and it misreading it occasionally when I am tracking a cue list against it to trigger cues. It would be off as high as 17 seconds or so. I have a video of it in action at the top of this thread. I had to switch to lockstep and convert the ltc to mtc, go out to a midi merger and then back in to qlab and use mtc instead.

On Tue, Nov 13, 2018 at 2:42 AM 'Nigel Hogg' via QLab <ql...@googlegroups.com> wrote:

I use LTC all the time over Dante DVS into a Midas M32R then out on a bus into a Jands Vista lighting desk. I have had no problems at all. I’m running Qlab on a Mac Pro, High Sierra. The only requirement is to ensure enough level on the channel but below clipping.

--
Contact support anytime: sup...@figure53.com
Follow Figure 53 on Twitter: https://twitter.com/Figure53
User Group Code of Conduct: https://figure53.com/help/code-of-conduct/
---
You received this message because you are subscribed to a topic in the Google Groups "QLab" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/qlab/tK5bG7hDyDE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to qlab+uns...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/qlab/22f7bd8f-8a0c-4e61-8536-f40a7ed051b6%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Thanks
Gary Parke
glp...@gmail.com

[Pierre-Luc Brunet]

Thank you Nigel and Gary for sharing your experiences.

I feel better about outputting it now. :)