Audio drift part 2

[Peter C.]

I filmed a performance on Saturday. I used a Zoom H5 to record the applause and space, while recording a board feed to a Zoom F3. The strange part the H5 was in sync for the first half of the show but in the second half it's badly out of sync. I thought I could change the speed to fix it but it looks like the speed wasn't consistent so instead of chopping it up and trying to sync it that way I'll just use one of the on camera mics. Really strange.

 

[Paul F]

Zoom doesn't specify a clock stability and it is a budget recorder, so the clock they put in it, well, who knows how stable it is. Also note that you really don't know what is drifting. You have multiple devices recorder and as I understand it, none of them are synced to one another. So the drift is relative. The camera may drift one way while the Zoom drifts the other way.

 

[Peter C.]

I know it's the H5 because the cameras sync to the Zoom F3

 

[NorBro]

If this was me, I think I would use 5-6 different 32-bit float recorders all over the place and be done with it.

Clipping doesn't exist so you don't have to worry about that. And then you'd sync only those in post - which for any kind of 1-2+ hour stage performance that's one long shot instead of dozens of different clips this is probably the most easiest thing ever for their use case.

 

[Peter C.]

I've thought about getting another F3, has 32bit but more importantly never had a clock issue. I could post the audio files but all you'd see is what I'm seeing. I've heard people complain about DJI mic 2 clock speed being off and yet I had a colleague use it recently for over an hour without issue. Seems to me a real hit or miss. But one thing for sure if the clock speed is changing its a nightmare to fix.

 

[cyvideo]

I've thought about getting another F3, has 32bit but more importantly never had a clock issue. I could post the audio files but all you'd see is what I'm seeing. I've heard people complain about DJI mic 2 clock speed being off and yet I had a colleague use it recently for over an hour without issue. Seems to me a real hit or miss. But one thing for sure if the clock speed is changing its a nightmare to fix.

By all means, post a camera audio file along with a zoom file and let's have a look see if there is an understanding to be gaines of what's going on... and maybe a solution ?

Chris Young

 

[Peter C.]

I uploaded mp3 of them to Dropbox, maybe it's something I'm doing that I'm not seeing.

Zoom H5 recorder that's out of sync

Dropbox

www.dropbox.com

Camera 1

Dropbox

www.dropbox.com

Board feed recorded into Zoom F3

Dropbox

www.dropbox.com

 

[Bassman2003]

If this was me, I think I would use 5-6 different 32-bit float recorders all over the place and be done with it.

Clipping doesn't exist so you don't have to worry about that. And then you'd sync only those in post - which for any kind of 1-2+ hour stage performance that's one long shot instead of dozens of different clips this is probably the most easiest thing ever for their use case.

I am not sure I agree with this. Apart from the financial side, unless these recorders are synced together, you might get 5-6 different opinions. Isn't it easiest to just put a mic on a camera that is in the room if it is just for applause and ambiance? That way the footage will be in sync and more than fine for applause. (which is what Peter did).

As a side note, I filmed a play last week where I had two mics feeding into my Zoom F6 in an orchestra pit right by the stage. No timecode. The play was 2 - 45 minute acts and the drift was pretty much nil. Even with the intermission being recorded so about 2+ hours of continuous record time. So the F6 keeps great sync. Hopefully your F3 works at a similar level.

 

[Peter C.]

I am not sure I agree with this. Apart from the financial side, unless these recorders are synced together, you might get 5-6 different opinions. Isn't it easiest to just put a mic on a camera that is in the room if it is just for applause and ambiance? That way the footage will be in sync and more than fine for applause. (which is what Peter did).

As a side note, I filmed a play last week where I had two mics feeding into my Zoom F6 in an orchestra pit right by the stage. No timecode. The play was 2 - 45 minute acts and the drift was pretty much nil. Even with the intermission being recorded so about 2+ hours of continuous record time. So the F6 keeps great sync. Hopefully your F3 works at a similar level.

Every situation is different. In this one I'm filming on the first level and the second level over hangs it so the acoustics are bad even for applause, get a weird reverb. The second level (balcony) has no place to set up a camera or recorder. Only place is a ledge where they have the projector I put the zoom H5 on a table top tripod. Could it be the heat or electromagnetic fields from the projector or lights caused and issue? Or the batteries when they get low effect the clock speed or maybe the unit itself is defective. Strange thing is it wasn't that long the first act was an hour and the second act was 30 minutes. The first act was in sync. Strange I can't figure it out.

The hardest part about mic placement is that in a theater if you want over all music and applause you don't want it near audience member if they talk or cough they'll get picked up more. The other problem it solves if you put a mic on the camera touching it or changing setting will cause noise in the mic.

 

[NorBro]

I am not sure I agree with this. Apart from the financial side, unless these recorders are synced together, you might get 5-6 different opinions. Isn't it easiest to just put a mic on a camera that is in the room if it is just for applause and ambiance? That way the footage will be in sync and more than fine for applause. (which is what Peter did).

As a side note, I filmed a play last week where I had two mics feeding into my Zoom F6 in an orchestra pit right by the stage. No timecode. The play was 2 - 45 minute acts and the drift was pretty much nil. Even with the intermission being recorded so about 2+ hours of continuous record time. So the F6 keeps great sync. Hopefully your F3 works at a similar level.

5-6 different opinions about what?

A mic on the camera would definitely be enough for that for many cases but I guess there is some sort of out-of-sync areas somewhere in the final product.

Like, if you hear clapping 5 seconds before or after you should, it's obviously a problem from a professional standpoint even if some people wouldn't care (and if the rest of the performance was fine).

I don't know what exactly is going on but the recorders would work great if the audio from all of them was identical in post and since they are not trying to match with anything else, otherwise it would be back to square one.

 

[Bassman2003]

I do not see the need to for 5-6 recorders just to have applause. The different opinions are the 5-6 recorders maybe running at different speeds. (drifting). Not huge but why so many recorders?

Peter, maybe I am not understanding the situation. If you have a direct feed from the board, then any ambient you record is going to be a smaller percentage of the final mix right? Then you have some wiggle room. Do you have a wireless mic? Maybe you could place that somewhere. I can see how a back of the house setup would not be ideal for audio if that is where your cameras are. So you put the F3 somewhere in a better location and slide the audio clips on the timeline in post to fix drift. Or can you run a long XLR for better mic placement?

If you were recording the performance with only mics, I can see more of an issue. Ambient pickup for only support has some options. Or am I off base?

 

[NorBro]

Oh, the recorders would be to independently record the performance and not use anyone else's audio or any other type of microphone.

5-6 was just a random estimation for different angles/parts of the performance, maybe could be done with 2-3, IDK.

 

[Peter C.]

I do not see the need to for 5-6 recorders just to have applause. The different opinions are the 5-6 recorders maybe running at different speeds. (drifting). Not huge but why so many recorders?

Peter, maybe I am not understanding the situation. If you have a direct feed from the board, then any ambient you record is going to be a smaller percentage of the final mix right? Then you have some wiggle room. Do you have a wireless mic? Maybe you could place that somewhere. I can see how a back of the house setup would not be ideal for audio if that is where your cameras are. So you put the F3 somewhere in a better location and slide the audio clips on the timeline in post to fix drift. Or can you run a long XLR for better mic placement?

If you were recording the performance with only mics, I can see more of an issue. Ambient pickup for only support has some options. Or am I off base?

The main issue is the drift was so bad couldn't use it or correct it since it wasn't consistent speed. I posted link of you can see for yourself. The theater is multi level too big to run cables. It's ok I just want to get to the bottom of the problem.

I also agree multiple recorder isn't what you want just need one that is properly located.

 

[Bassman2003]

I think the problem for a lot of these types of jobs is low budget blocking the use of the best/proper tools. It is frustrating.

 

[cyvideo]

I uploaded mp3 of them to Dropbox, maybe it's something I'm doing that I'm not seeing.

Zoom H5 recorder that's out of sync

Dropbox

www.dropbox.com

Camera 1

Dropbox

www.dropbox.com

Board feed recorded into Zoom F3

Dropbox

www.dropbox.com

Yes, Peter. I see your problem.

The three files synced and drift corrected in about 20 seconds in Pluraleyes. But that revealed a problem. The H5 has stopped recording twice. Once for 10:20 seconds and the second time for 13:20 seconds. The gaps were easily repaired by stealing from the camera file and adjusting the repair inserts. Final sync checks and tweaks were done in Resolve using Elastic .WAV. Check the attached before and after JPGs to see what I mean. I will PM you with a download link for the repaired files. They will be in MP3s in a ZIP file. Good luck!

Chris Young

 

[Peter C.]

Wow Chris your amazing. I would never have believe that it stopped recording. Must be a defective recorder since I pressed record and let it go for the entire event. Very strange indeed. Thanks so much you're the best.

 

[cyvideo]

Wow Chris your amazing. I would never have believe that it stopped recording. Must be a defective recorder since I pressed record and let it go for the entire event. Very strange indeed. Thanks so much you're the best.

Yes, when I saw the files out of Pluraleyes, I could see there were synced file areas and big areas with no sync. Which generally means there is no other audio to sync to. I've never seen drift issues in the ranges of 20 + seconds. I think about the worst I've seen is up to about 2 seconds in an hour. So knew there must be an issue if the software can't find a matching waveform in any of the other tracks. It was easy to pick the big gaps.

Chris Young

 

[Peter C.]

Just had time to run a test. I recorded with H5 and F3 for 2. 5 hrs no drift. My guess is the proximity to the projector and lights in the theater may have emitted electromagnetic field caused an issue. What else could it be?

 

[cyvideo]

Just had time to run a test. I recorded with H5 and F3 for 2. 5 hrs no drift. My guess is the proximity to the projector and lights in the theater may have emitted electromagnetic field caused an issue. What else could it be?

In cases like this, I say "Blame Bill Gates". Sometimes there just seems to be no definitive reason as to why these things happen. Your conclusions are just as valid as anyone else's until a specific cause, if ever, is established.

Chris Young

 

[cyvideo]

Just had time to run a test. I recorded with H5 and F3 for 2. 5 hrs no drift. My guess is the proximity to the projector and lights in the theater may have emitted electromagnetic field caused an issue. What else could it be?

Peter. Since you informed me, that it was a 29.97 camera time base, I ran the sync again but this time with a video track with your camera audio track. Having the video track made a massive difference. Now the recorder tracks had something to reference to on a 29.97 vision basis. This time, more errors, 5 all told, were picked in the H5. But everything synced up nicely. Patched the 5 breaks and all looks and sounds pretty decent I think. I'd be happy to use the v2 files in a job without too much concern. I'll PM you a link again.

Chris Young