Does RootEncoder support background music (MP3) in live stream?

[noho501]

Hi,

I would like to ask if RootEncoder supports adding background music (for example, from an MP3 file) to the live stream.

The idea is that when the mic is muted, viewers can still hear the background music in the stream.

Is this feature supported, or can it be done by using a custom audio source?

Thank you.

[pedroSG94]

Hello,

Do you want mix 2 audios or change the audio source to a file?
You can do both cases.

In the case of change to an audio file source you can use the current library code with AudioFileSource.
If you want mix 2 audios you will have to create your own custom audio source. You already have a method to mix 2 pcm audios here:
https://github.com/pedroSG94/RootEncoder/blob/master/encoder/src/main/java/com/pedro/encoder/input/audio/AudioUtils.kt#L7

This method mix 2 PCM buffers with this requirements:

• Same number of channels
• Same sample rate
• Same buffer size

[noho501]

I want to mix 2 audios. Thank you so much, I will try it

[noho501]

Hi @pedroSG94 ,

Thanks for the explanation. I’m still a bit confused about how to implement this in practice.

My use case:

• Mix MP3 audio + microphone and send the mixed audio to the livestream.
• Viewers should hear the music even when the mic or device speaker is muted.

Could you please share a small minimal example (or pseudo-code) showing how to mix MP3 + mic audio?

Here is my code:

    val fragment = MicAndAudioSettingFragment(isVertical)
    fragment.apply {
        onDeviceSelect = { device ->
            switchMicrophone(device)
        }

        currentAudioDeviceProvider = {
            getCurrentAudioDevice()
        }

        onMicVolumeChange = { percent ->
            val micSource = genericStream.audioSource as? MicrophoneSource
            micSource?.microphoneVolume = percent.toFloat() / 100
        }
    }

Thanks a lot for your help.

[pedroSG94]

Hello,

The way to mix the audio is similar to this:
https://github.com/pedroSG94/RootEncoder/blob/master/encoder/src/main/java/com/pedro/encoder/input/audio/MicrophoneManager.java#L278

The difference is that you have to replace audioRecordDevice to a method that return the audio from the MP3 on PCM format with the same buffer size and same sample rate than the microphone PCM buffer.

To obtain PCM buffer from a MP3 file you can try with the audio decoder class:
https://github.com/pedroSG94/RootEncoder/blob/master/encoder/src/main/java/com/pedro/encoder/input/sources/audio/AudioFileSource.kt

Maybe you need modify it to get the buffer with the same size and you also need be sure that the file sample rate is the correct.

If you still don't know how to implement it let me know it.

[noho501]

Hi,

Thanks for your explanation.
I’ve actually read through all the code you mentioned above, but to be honest it’s quite complex for me to implement correctly.

I was wondering if the library could support something simpler, for example:

val micSource = genericStream.audioSource as? MicrophoneSource
micSource.microphone.track[1].volume = 0.5f
micSource.microphone.track[1].append(pcmBufferMusic)

Or even a higher-level API like:

genericStream.audioBgVolume = 0.5f
genericStream.startAudioBg("file.mp3")
genericStream.stopAudioBg()

This would make it much easier to mix background music with the microphone for livestreaming.

Honestly, this feature is quite challenging for me — I’ve been trying for about 3 days but still haven’t managed to get a working result.

What do you think about this kind of API or approach?

Thanks for your time.

[noho501]

Hi @pedroSG94 Could you please help me with supporting this feature, or share an example if available?

[pedroSG94]

Hello,

For now, I did a test of concept that works here:
https://github.com/pedroSG94/RootEncoder/tree/multitrackaudiosource

For testing I used a video with 44100 sample rate and stereo as audio.
You only need compile the app, open rotation example. Select microphone audio source using the menu (I know it is already selected but click the button anyway) and start record/stream.
For now, loop is not supported and need checks to confirm that you are using the same configuration than the original audio source.

This feature don't need add extra code to the library because it is supported using a CustomAudioEffect.

Let me know your tests results and possible improvements.

[noho501]

Thank you so much. I’ll check it and work on improving it.