SFO Commons - Wirecast or BlackMagic producing garbled audio

RESOLVED FIXED

Status

Air Mozilla
Streaming
RESOLVED FIXED
5 years ago
3 years ago

People

(Reporter: jbraddock, Assigned: Jared Obstfeld)

Tracking

(Blocks: 1 bug)

Details

(Reporter)

Description

5 years ago
This past Thursday I was setting up an Air Mozilla event from the SFO office and when the stream went live I noticed that the audio being pushed out and recorded on encoder1.commons.voip.sfo1.mozilla.com through Wirecast was garbled (robotic) and did not sound correct.

This issue was not apparent when on a Vidyo only call - so it is something to do with either the encoder, Wirecast, the BlackMagic box or something in-between. I have reset all of the equipment in the rack for SFO Commons w/ no luck (broken audio still apparent). Please help me figure out what the issue is.
(Reporter)

Updated

5 years ago
Assignee: server-ops → nobody
Group: infra
Severity: minor → normal
Component: Server Operations → Streaming
Product: mozilla.org → Air Mozilla
QA Contact: shyam
Version: other → unspecified
(Reporter)

Updated

5 years ago
Blocks: 724024
Interesting problem.   Did you power-cycle the Clear-One boxes?
(Reporter)

Comment 2

5 years ago
Yes - I power cycled everything in the rack for SFO Commons, including the Clear-Ones. The interesting thing is that the audio is just fine when on a normal Vidyo call, but if the Encoder input is selected the sound gets garbled (robotic sounding). I can show you in person if you would like.
(Reporter)

Comment 3

5 years ago
+ Tim (AVDG)

Tim - not sure if you are able to do this through a bug, but I am going to try anyways. Would it be possible to get someone from your team out to check this issue out that is detailed above? If you need me to send an email, please let me know. Thanks!

Comment 4

5 years ago
I forwarded this to Service to dispatch the call.  I will get my service techs signed up for bugzilla as well.
(Reporter)

Comment 5

5 years ago
Thanks Tim - just let me know the date and time they are coming out (when you get that info) so I can make sure we are prepared.
(Reporter)

Comment 6

5 years ago
Update: Jared determined that the issue is being caused by an Crestron input card, which he is RMAing. Jared will be coming out on Monday, Dec 3rd to temp use an unused input card from SFO Commons as a temp fix until the replacement card come in the mail.
(Reporter)

Comment 7

5 years ago
Reference ServiceNow Request: INC0010677
(Reporter)

Comment 8

5 years ago
Hey Tim and Jared - We have replaced a Crestron sound card, updated all firmware/hardware, tried a new Black Magic box all with no luck - the audio issue is still apparent in SFO Commons when using the SFO Encoder.

Can we get someone out to SF to check out the issue? I know Jared is going to be here tomorrow (Wednesday, Feb 6th) would that be a good time? Thanks!
Assignee: nobody → jaredo
(Assignee)

Comment 9

5 years ago
Hey Joel,

Any way we can test with a different Mac Mini? It's the only thing we haven't switched out. Just a thought. If you want to schedule me there earlier tomorrow, I can do that.

Thanks!
Jared
(Reporter)

Comment 10

5 years ago
@Richard or Zandr - Is it OK if I take one of the Encoder mac minis from Bay Bridge or Golden Gate Bridge for the test Jared detailed in comment 9? Figured this would be easiest as all software is already installed.
:jbraddock- Go for it. I'd be amazed if it had anything to do with the problem, but we might as well rule it out.

When you say you replaced a "Crestron Sound Card", I'm not sure what that means. Which component was replaced?
(Assignee)

Comment 12

5 years ago
I'll answer this one. We replaced the DVI (with audio) input card in the Crestron Digital Media chassis that was passing the audio for the Black Magic. It alleviated the problem for a short time put the problem came back. I will look at the new card but it is unlikely that the new card is the issue.
(Reporter)

Comment 13

5 years ago
Jared - I have yet to test out swapping the Mac Mini, but I know that you came into check out a couple items in SFO Commons. Did you find anything that may be related to the audio issues?
(Assignee)

Comment 14

5 years ago
Hey Joel,

This issue is mysteriously fixed again, which is why we need to continue to keep an eye on it. I may need you to check it each morning for a while just to monitor the stability. I changed input cards again, which did not help. I rebooted the Crestron switcher, which didn't help. I routed a different source to that output and then switched back and that seemed to fix it. I was able to hear the noise trough the headphones on the live feed so I did not have to make a recording first. Once I fixed the issue live, I made a recording and I left it save on the desktop. If the problem resurfaces we should then switch out the Mac Mini but until then let's do routine check ups.

Thanks!
Jared
(Reporter)

Comment 15

5 years ago
Jared - Thank you for the update here. Emajee and I just tested the audio and it is working just fine as of today at about 11:40AM. We have an Air Mozilla event tomorrow, which will used the SFO Encoder and I will test again before the event. Will update when I know more.
(Reporter)

Comment 16

5 years ago
FYI: We tested the audio with AirMozilla after the SFO Commons rework was completed and the audio issue is  no longer apparent. There is an AirMozilla event on the beginning of March and if all is good after that point, I will close out this bug. Thanks for all of the help.
(Assignee)

Comment 17

5 years ago
No problem.  We did some re-routing with the upgrade and John found something buried in the Clearone that may have been causing an issue but we're not positive that was the case.  Either way, we noticed it was nice and clear so definitely keep us posted!  Good news!

Comment 18

5 years ago
Hi Mozilla Folk!

I'm run ~25 UX Research labs at Google and I've experience the same problem with Wirecast using many different hardware configurations (DeckLink Quad cards, line-in audio, audio-to-usb) (but all using one revision or the other of MacPro desktops).

I've failed at multiple attempts to reproduce this issue constantly and find it typically occurs when I least want it to (when a researcher is doing a study).

Everything I've seen and read so far seems to point me to a bug within OS X. Here's a few threads I've dug up with some Google-fu while researching a solution:

Examples:
http://forum.teamspeak.com/showthread.php/60215-Mac-OS-X-Robotic-voice
http://forum.ventrilo.com/showthread.php?t=47741
http://forum.ventrilo.com/showthread.php?t=45715
https://discussions.apple.com/thread/2166616?start=75&tstart=0
https://discussions.apple.com/thread/2166616?start=105&tstart=0 (same thread as above, different starting post)
https://code.google.com/p/chromium/issues/detail?id=158974
https://jira.skype.com/browse/SCM-747
http://community.skype.com/t5/Mac/Robotic-voices-when-using-Lion-Plantronics-headset-and-hosting/td-p/68702
http://www.macosxaudio.com/forums/viewtopic.php?f=49&t=50788
http://forums.counterpath.com/viewtopic.php?f=1&t=708&start=15
http://duc.avid.com/showthread.php?t=228224
http://forum.audacityteam.org/viewtopic.php?f=47&t=68847
http://h30434.www3.hp.com/t5/webOS-Hardware-and-Software/Phone-Skype-Quality-is-really-bad-Sounds-like-a-robot-on-the/td-p/2111484
https://code.google.com/p/chromium/issues/detail?id=157613
http://forums.logitech.com/t5/Video-Chat-Applications-Non/Robotic-audio-for-recipient-MacOS/td-p/844812
https://code.google.com/p/bigbluebutton/issues/detail?id=1357
http://forums.macrumors.com/showthread.php?t=213745


Example Audio (sounds a little bit better than the robot voice we have experienced):
https://www.youtube.com/watch?v=3ugdk89ojxU
(Reporter)

Comment 19

5 years ago
This issue has actually been resolved with the rework of the ClearOne processor. AVDG noted that there were some invalid audio routes (or something similar). Closing out bug.
Status: NEW → RESOLVED
Last Resolved: 5 years ago
Resolution: --- → FIXED
(Reporter)

Updated

4 years ago
Blocks: 893963
You need to log in before you can comment on or make changes to this bug.