c++ – Writing audio data to an existing audio file using the FFmpeg API


My main task is to record the received audio data into an existing file at the request of the client.

As an example, let's say I have already created an audio file with an audio recording length of 1 hour. I want to open it, record an additional 1 hour of audio data, and close it. As a result, we should get an audio file containing a 2-hour audio track. I understand correctly that in this case, if possible, I will have to delete the previously recorded trailer (it was recorded thanks to the av_write_trailer function), write the received data to the end of the file, and when I close the audio file, record the current trailer?

This is the code snippet where I record the trailer and close the media file:

// Writing the stream trailer to an output
// media file and free the file private data.

It's important to note that the implementation of writing audio data to a new audio file works fine. The problem is exactly how to write the audio data to the existing audio file.


In its general form, this problem has no solution. The ability to append data to an existing file is essentially impossible for most audio formats because it involves a combination of many factors: the file structure should assume the possibility of appending with or without the need to match the data formats of the existing and new parts; the implementation of the corresponding container must be explicitly adapted to the gluing; in most cases, the implementation will assume reading the entire available fragment before the first byte can be appended, often incl. using a temporary intermediate file / stream.

Of the traditional data formats, there are only a few examples, such as MPEG TS, where appending is in principle provided. As a standard, gluing occurs from ready-made pieces and by a full-scale re-creation of the file with data copying.

Scroll to Top