Imported audio slowly drifts. Why?

I recently filmed high school students doing their practical music exam on a cheap Nikon compact camera. The microphone on the camera was horrible on its own, so I used a good Zoom recorder to capture WAV at the same time.

But after importing the Zoom audio, if I sync it to the nearest 0.01 of a second at the beginning of the video by the end of an eight minute piece it has drifted by 0.02 to 0.04 seconds and no longer syncs. It causes an echo against the original audio, and makes their finger movements on the instrument seem ‘spongy’.

Is this something to do with the way Shortcut imports the audio?
Or is it something to be expected when recording and importing separate audio?

NB: I got around the problem by cutting only the Zoom audio at an opportune place in the middle of the pieces, and shifting the audio to cause a transition that, in the right place, can’t be heard. I’d still like to know why the audio drifts though.

This happens because the clocks in the two devices (Nikon camera and Zoom recorder) are not synchronized. That means that one will be capturing the audio slightly faster than the other device. This is normal operation. Some professional devices have a means to synchronize the clocks to an external clock. But consumer devices would not have such a feature.

You may get by with this for short clips where the clocks do not drift by very much. But you shouldn’t expect it to work for longer clips. Your work-around of clipping and shifting one source is probably a good approach. You could also try changing the speed of one clip to try to match the other clip. You might be able to get them to stay synchronized longer by speeding up the slower clip.

As @brian wrote, professional equipment has gen lock and/or LTC inputs that keeps all the equipment synchronised.
There are a few things that you can do to minimise drift with a domestic setup such as keeping recordings short and making sure that both your camera and audio recorder have the same sample rate.

Thanks for the clarifications.

The audio sample rate of the Nikon was 24000, and the Zoom was 44100.
They also used different codecs (AAC and PCM respectively).

I just noticed the ‘sync slider’ in the properties box.I was just zooming into the timeline and sliding the clip left or right. The slider might easier and more accurate; I’ll try that next time.

Thanks.

Hello:
You can use an external audio editor such as Audacity, to try to correct the drift of the sound.
This free software has explanatory documentation about the different effects and analysis tools that it provides.
You can drag the video directly into the software and you get the audio track.
Then you can add the zoom microphone track and adjust it to match both.
To import the audio included in the video, you must install the FFmpeg library.
It is also useful to install the LAME MP3 encoding library if you want to export your projects in mp3 format.
Other available plugins allow more useful functions.
https://www.audacityteam.org/download/plug-ins/

To understand how Audacity works, you can read the online manual.

https://manual.audacityteam.org/

This can also be useful :

Change Speed
Change tempo

The resulting audio project in Audacity is exported in WAV format and added as an audio track in Shotcut.
In Audacity you can create a mix of both sources (audio in the video and audio of the zoom microphone). Additionally you can also normalize, amplify, etc.
Ideally, however, make sure that the sampling rate, frequency and codec are equal at the time of recording (whenever possible)

If I were in your situation I would simply get a better mic for the camera and not bother with trying to sync a separate sound track. Hopefully the camera has an input for an external mic, and that input has a proper mic preamplifier.

I have a Sony camcorder and a Canon camcorder which both have those things. I can connect an outboard mic and get wonderful audio.

Here is a mic suggestion provided you do not have a long cable run.

Interfacing a mic to a consumer camcorder can be tricky, especially when stereo is involved.

This topic was automatically closed after 90 days. New replies are no longer allowed.