I want to Capture the audio being played by other apps using AudioPlaybackCapture API and load it into a classifier but I cant load the AudioRecord directly to TensorAudio because it throws unsupported format error. my only chance is to convert AudioRecord to short array or float array and load it into classifier.

here are some part of my AudioCaptureService.java

mediaProjectionManager = (MediaProjectionManager) getSystemService(Context.MEDIA_PROJECTION_SERVICE);
        mediaProjection = mediaProjectionManager.getMediaProjection(RESULT_CODE, MPDATA);

        AudioPlaybackCaptureConfiguration config = new AudioPlaybackCaptureConfiguration.Builder(mediaProjection)
                .addMatchingUsage(AudioAttributes.USAGE_MEDIA)
                .addMatchingUsage(AudioAttributes.USAGE_GAME)
                .build();

        AudioRecord tensorAR = classifier.createAudioRecord();
        record = new AudioRecord.Builder().setAudioFormat(tensorAR.getFormat())
                .setAudioPlaybackCaptureConfig(config)
                .build();

        record.startRecording();

tensor.load(record);
List<Classifications> output = classifier.classify(tensor);

I tried this code from chatGPT but no luck

public float[] readAudioData() {
        audioRecord.read(buffer, 0, buffer.length);
        float[] floatBuffer = new float[buffer.length];
        for (int i = 0; i < buffer.length; i++) {
            floatBuffer[i] = buffer[i] / 32768f; // Convert short to float (-1.0 to 1.0 range)
        }
        return floatBuffer;
    }

Thank you so much