Genlock issues - support ticket unanswered for nearly 3 weeks

[Nick Burridge]

Hey guys,

Not normally one to complain, hugely satisfied with our RED stuff thus far, but I've been dealing with some genlock/dropped frame/audio issues on behalf of a client with 2x Scarlets, and aside from the initial reply to the ticket filing (filed on June 11th, answered on the 13th) there has been no reply to any of the further details I've sent through. My client filed the ticket on the 11th, RED replied on the 13th, I sent through further information on the 15th, then nothing. Not even a "we're looking into it" - sent a reminder on the 26th, and still nothing. This isn't typical of the RED support we've had in the past, so I'm hopeful we've accidentally just fallen through the cracks, but still - myself and my client are up in front of a sizable production, and when issues like this happen without resolution we all (including RED) look bad. Hoping to get this resolved soon.

For those interested, the issue number is RED #71185 and the details are quoted below:

I'm handling the technical side of things for Julian on this issue - I was able to reproduce the issue (redescribed below) using two different Scarlets (#503 and #870, build 3.0.7 - firmware was updated using the 'force_' prefix so should have done a complete reset) with two different Ambient Lockits attached.

The setup:
Scarlet-X: 4K HD, timebase 25fps, frame rate 25fps, powered using GlobalMediaPro R-model (RED firmware) bricks. Settings > Setup > GPIO/Sync is set to 'Genlock'
Ambient Lockit ACL 203: set to 1080p/25fps/25fps, supplying both genlock and timecode to the camera using a RED 3BNC-to-000 sync cable. TC, GEN, and SYNC are lit green
Wooden Camera A-box: takes a feed from the sound recordist's mixer and supplies mono/single channel guide audio to be recorded with the R3Ds

The issue:
The camera appears to drop frames on a cycle (the cycle differs between the lockits/cameras), causing a noticeable 'hitch' in the video playback. As the video drops frames, the guide audio gets further and further out of sync (the offboard recorded audio also doesn't sync after aligning by slate or by timecode).
Some examples from both Scarlets (Camera A = #503, Camera B = #870) processed in REDCINE-X to h264 Quicktimes:
http://www.halcyondigital.co.nz/halcyonftp/tbr/A012_C001_0614P0.mov panning test, demonstrates dropped frames/'hitch' in video
http://www.halcyondigital.co.nz/halcyonftp/tbr/A012_C002_0614YK.mov slate test, demonstrates guide audio drifting further and further out
http://www.halcyondigital.co.nz/halcyonftp/tbr/B012_C001_06138K.mov panning test, demonstrates dropped frames/'hitch' in video
http://www.halcyondigital.co.nz/halcyonftp/tbr/B012_C002_0613PK.mov slate test, demonstrates guide audio drifting further and further out
and the log file from Camera B (Camera A didn't generate one as I didn't power the camera down) is here.

The 'solution':
After unplugging genlock and switching Settings > Setup > GPIO/Sync to 'Off' the issue vanishes. Timecode is left plugged in. TC is lit green, GEN and SYNC are grey.

Whilst genlock is unnecessary for this setup (just timecode is needed for sync sound) and production is continuing it would be good to get to the bottom of this issue as it sounds like what is described here, here, and here.

 

[Gunleik Groven]

So... This is in no way to be considered a Red official answer, but...
We have found that Red+Genlock+not sufficient testing of which Genlock box gives what kind of signal = problems.
Thus, we have ended up not using Genlock on all the shows I have been in touch with.

That is not to say that it is an impossible combo. Just that we have seen different issues on R1 and Epic that have lead us to turn it off.

The issue you describe is close to what Phillip Bloom saw, and what I have seen on the Epic.

Jam works very well, though, so if you are not recording very long takes, you should be fine.

If you need Genlock to work... I guess you should be talking privately to Brook... Or get the response you need from the Red team about "certified" lockit box brand and configurations.

I have never managed to - or really bothered to - solve this issue, butbI have a hunch that it might be a p vs psf thingy for the Epic @ 1080, possibly related to 24 vs 25 as well...

But I am totally aware that I may be waaaaaaay off on this hunch.

Cheers and good luck and it would be good to know what config you end up with, if you solve it.

G

 

[Nick Grow]

Nick, the last communication I see regarding tech support ticket [RED #71185] was to you on June 26th requesting to send us a log file from SCARLET-X #503 for diagnosis. Were you able to send that to us?

NICK

 

[Nick Burridge]

Nick, the last communication I see regarding tech support ticket [RED #71185] was to you on June 26th requesting to send us a log file from SCARLET-X #503 for diagnosis. Were you able to send that to us?

NICK

Hi Nick, thanks for responding - weirdly I can't find any trace of that reply from RED in my inbox, spambox, or anywhere else. It appears the interwebs have eaten it.
As for the log file, as mentioned Camera A didn't generate one during testing/diagnosis, as I didn't power the camera down and forgot to opt for a manual generation of log file through the maintenance menu. I do have the log file from Camera B (#870) which also demonstrated the issue, is that not sufficient?
Regardless I'll try and get log files from the post house for the bad Camera A (#503) mags from the first days of shooting when this problem occurred (prior to testing) - they have no on-set data management which makes this a bit harder.

So... This is in no way to be considered a Red official answer, but...
We have found that Red+Genlock+not sufficient testing of which Genlock box gives what kind of signal = problems.
Thus, we have ended up not using Genlock on all the shows I have been in touch with.

Cheers Gunleik, during my poking round Reduser trying to resolve this I did see you had similar issues, we're also just running with TC Jam only and no Genlock, seems to be fine - in fact it's what I would have had them doing in the first place if I'd been involved with the testing prior to production, but only got brought in after they'd lost 2 days of shooting to this issue. Will definitely let you know if we resolve it!

 

[Nick Burridge]

I've received the log files from the post house, and have just emailed them through to support(at)red(dot)com, but copying it here as well as email communication seems to be a little unreliable.

I've sourced the log files from #503 for the first 2 days of shooting where the issue presented - I don't believe all of the clips on these mags had the issue, as production tells me they were unplugging/replugging genlock all the time as it was causing the PIX240 video assist system issues as well when genlock was connected. But the very first setup on Camera A/Roll 1 had the issue, which then reappeared intermittently depending on production's setup (not being on set it's difficult for me to know what they changed and there's no RED-savvy person there to say "yes we turned genlock off from here").

Log files zipped and linked below:
https://rcpt.yousendit.com/1588727275/010fa56a693e752d4336fb29f871e588

 

[Nick Grow]

Nick, I confirm we received the link to the log file you just sent and I forwarded you that email from June 26th for reference. We will get this sorted for you.

NICK