用于录制音频的 Android AudioRecord 和 MediaRecorder

If you want to do your analysis while recording is still in progress, you need to use AudioRecord, as MediaRecorder automatically records into a file. AudioRecord has the disadvantage, that after calling startRecording() you need to poll the data yourself from the AudioRecord instance. Also, you must read and process the data fast enough such that the internal buffer is not overrun (look in the logcat output, AudioRecord will tell you when that happens).

AudioRecorderer first saves data in minBuffer then it is copied from there to the temporary buffer, in MediaRecorder it is copied to files. In AudioRecorder we need the api setRecordPosition() to copy the saved data at required position, whereas in MediaRecorder the file pointer is does this job to set the position of the marker. AudioRecorder can be used to those apps which run on an emulator this can be done by providing low sample rate such as 8000, while using MediaRecorder the audio cannot be recorded using emulator. In AudioRecord the screen sleep after sometime, while in MediaRecorder the screen does not sleep.

As I understand MediaRecorder is a black box which gives compressed audio file on the output and AudioRecorder gives you just raw sound stream and you have to compress it by yourself.

MediaRecorder gives you the max amptitude from last call of getMaxAmplitude() method so you can implement a sound visualizer for example.

So in most cases MediaRecorder is the best choice except those in which you should make some complicated sound processing and you need access to the raw audio stream.