Displacement of Audio compared with waveform and video


We are facing problems with a misplacement of the actual audio being heard and the visual waveform that follows a video. This problem appears usually in long videos of 5 minutes or so, maybe it is an audio sampling rate problem or a buffer size issue.
Can you please give us some information about this?
We are using our own custom audio_manual recipe to annotate audio based on a video feed, with the waveform down below the video. So both videos and waveform are getting out of sync with the audio being heard after 3-4 minutes or more.
Thank you.

Thanks for the report! Could you share some more details on how you're loading the video data? Are you sending it with the tasks as a base64-encoded string, or are you loading it from a URL? And is the waveform out of sync, too, or just the audio playback? Also, which browser are you using?

Some background on how it works under the hood: if you're using a video, the "source of truth" is the embedded video, which is just a regular HTML5 video. The audio from that is then loaded into the audio player – essentially like this. So my current guess is similar to yours and that it could be related to the way the video is buffered and the audio widget not being able to "catch up" fast enough.

Hey Ines,
We are sending the tasks with a base64-encoded string. The waveform is out of sync with the audible audio.
I figured it might be the surfer embedded player, probably because the files are a bit large to handle maybe.
Do you have a work-around for this ?

Thank you,