Nomad Help - original audio file sounds stretched or slow??

[Jason Manzano]

I also posted this in the official zaxcom forum, but thought I would double my chances for a quick answer.

I had a client call me today about a single one of my files (out of several) sounding slow. I was sure it was an import problem on their part, at worst a mirroring problem, but when I went back and listened to the original .marf file on the nomad it does sound as if it were stretched or slowed down. I have never experienced anything like this. I have had the nomad since it was first released and have been on version 7.27 for at least a few months with no problems. The file was 24bit 48k 23.98. No other files sound like this and no setting changes were made for that take. I need to know why this happened so I can prevent it and know if I can I save this file for my client. I will, of course, start by updating firmware. Any suggestions?

[Philip Perkins]

Maybe load the file into WaveAgent and see if the TC metadata is what you think it is, and if the file duration makes sense.  Do you have another take with the same talent speaking so you could compare the relative pitch of their voices and the BG?   How much slowness are we talking here?  Did your client complain because the file would not sync to picture?  If you sent a ref feed to camera you could compare the sound of that audio with your file and know for sure if something is odd with your recording.

[Jason Manzano]

Thanks Philip.  I just loaded into wave agent.  Everything looks fine.  It is the expected TC, bit rate, sample rate.  File length is more difficult because I do not remember exactly how long the interview was.  I am audibly comparing it to other files and it definitely sounds like they are talking about 50% slower along with the typical under-modulated lower pitch quality. I did send a reference to camera (and am suggesting they use for this track).  I am now asking them to upload me the reference track for comparison.

[Philip Perkins]

Wow, 50%?  That's really huge.  You may have an intermittent sample clock issue in your recorder if that's true.  I'll be interested in what you hear by way of a comparison between the ref track and your file!

[Jason Manzano]

ok. I way over-estimated that. I knocked it down to 85% and it sounded right.  Still don't have the ref track to compare.  Not sure if my math is right but that could theoretically be caused by a 55.2k sample rate?

[MartinTheMixer]

I really can't to see how this one ends. We all must not be thinking about some obvious thing.

Right? When I think about a problem, I try to figure out how one could make this happen on purpose. 

So, everyone, how could we make this event happen on purpose?  

 

[Philip Perkins]

I really want to hear about the ref track comparison before I speculate further.   This doesn't sound like an editor ingest issue, like the old FCP bogus pulldown problem re TC, since the OP is playing it is Wave Agent and hears the problem there.  It would be great to know how much longer the problem file is than the ref track, if that is the case.

[Wandering Ear]

Does the duration in wave agent correspond to the duration in the sound report?  

[Jason Manzano]

great question, thought you might have had something there, but the length in both wave agent and the sound report is the same

[Jason Manzano]

Talked to Zaxcom's software engineer.  This seems to be a unique problem.  Sounds like it booted up at a different sample rate for that one take (it was powered down before and after take).  I am going to use the reference file (when I receive it) to match length and send it back to my client.  I was recommended to put a couple seconds of tone at the head of takes for a while to use as an internal reference if it happens again.  Hopefully with my firmware update this won't happen again, if it does then it needs to be serviced.

 

Thanks all for your input!

[Philip Perkins]

....scary. 

[Wandering Ear]

That is a bit unnerving that could happen and there would be no way to know until post.  

[old school]

No ideas but it is an interesting problem. I hope it has a useful resolution.

CrewC