Unable to work on the gain of the microphone in a native iOs app (using swift)

[Philippe Bourez]

Hello guys,

I'm actually working on an native iOs application using webrtc. The sound is live streamed to a Ruby on Rails server. It's all working great, but I get quite quickly a Larsen problem.

I would like to work on the "volume" / gain of the iPhone microphone, but I find nothing helpful on the internet. Does someone here already worked with the gain of the microphone?

I have taken inspiration of the PhoneRTC project (https://github.com/alongubkin/phonertc).

Here is where I create my stream, and add it to my peerConnection. 

     var peerConnection: RTCPeerConnection?
     var pcObserver: RTCPeerConnectionObserver!
     var queuedRemoteCandidates: [RTCICECandidate]?
     var stream: RTCMediaStream?
     var streams: StreamsConfig
     var localAudioTrack: RTCAudioTrack?

     func createOrUpdateStream() {
        if self.stream != nil {
            self.peerConnection!.removeStream(self.stream)
            self.stream = nil
        }
        self.stream = peerConnectionFactory!.mediaStreamWithLabel("ARDAMS")

        if self.streams.audio {
            if self.localAudioTrack == nil {
                self.initLocalAudioTrack()
            }
            self.stream!.addAudioTrack(self.localAudioTrack!)
        }

        self.peerConnection!.addStream(self.stream)
    }

    func initLocalAudioTrack() {
        self.localAudioTrack = peerConnectionFactory!.audioTrackWithID("ARDAMSa0")
    }

I don't know if it's possible in a way or another to get the RTCAudioTrack microphone input device, and then change it's gain? If it is possible, I sadly didn't find where.

Thanks for taking time reading my problem,

Philippe

[Philippe Bourez]

I'm now trying to modify webRTC source got from "http://webrtc.googlecode.com/svn/trunk" thanks to this blog (http://ninjanetic.com/how-to-get-started-with-webrtc-and-ios-without-wasting-10-hours-of-your-life/)

Is there any way to work on the audio_device.mm class found in the source code of webRTC, in order to change the input sound?

[Lars Bork]

Hey Philippe,

have you been able to find a solution for this problem? I am trying to do something similar right now and would love to hear you opinion!

Thanks a lot and have a nice day!

[Michaël Albert]

Hi Lars,

I've done such a thing in modifying the file audio_device_ios.mm.

AudioDeviceIOS::RecordProcessImpl(AudioUnitRenderActionFlags *ioActionFlags,

                                      const AudioTimeStamp *inTimeStamp,

                                      uint32_t inBusNumber,

                                      uint32_t inNumberFrames) {

    // Setup some basic stuff

    // Use temp buffer not to lock up recording buffer more than necessary

    // todo: Make dataTmp a member variable with static size that holds

    //       max possible frames?

    int16_t* dataTmp = new int16_t[inNumberFrames];

    memset(dataTmp, 0, 2*inNumberFrames);

    AudioBufferList abList;

    abList.mNumberBuffers = 1;

    //abList.mBuffers[0].mData = dataTmp;

    // HERE - insert gain modificator

    abList.mBuffers[0].mData = dataTmp * 0.3;

    abList.mBuffers[0].mDataByteSize = 2*inNumberFrames;  // 2 bytes/sample

    abList.mBuffers[0].mNumberChannels = 1;

    // Get data from mic

    OSStatus res = AudioUnitRender(_auVoiceProcessing,

                                   ioActionFlags, inTimeStamp,

                                   inBusNumber, inNumberFrames, &abList);

    if (res != 0) {

        WEBRTC_TRACE(kTraceWarning, kTraceAudioDevice, _id,

                     "  Error getting rec data, error = %d", res);

        if (_recWarning > 0) {

            WEBRTC_TRACE(kTraceWarning, kTraceAudioDevice, _id,

                         "  Pending rec warning exists");

        }

        _recWarning = 1;

        delete [] dataTmp;

        return 0;

    }

As you can see you have just to multiply the value in the buffer. you can do a binary division but this code is working.

It is an older version of WebRTC, check with the latest version.

Let me know if it helped you.

Cheers,

Michael