Audio Quality Issues on Android Devices with LiveKit

[JahagirDS0]

Hi @davidliu

I’ve recently integrated the LiveKit SDK into my video call application, and I’m now looking to implement call recording for both participants.

Could you please confirm if this can be handled at the app level, or if it must be done on the server side using the Egress API, as mentioned in the documentation?

Looking forward to your guidance on this.

Thanks,

[JahagirDS0]

Hi @davidliu
I've successfully implemented call recording using the Egress API on the server side.

However, during testing across different countries, we’ve noticed some audio issues on Android devices. Specifically:

There’s a slight echo on certain sentences, especially at lower volumes.

On some Android devices, we're also hearing unwanted noise artifacts (e.g., “chik-chik” sounds) during the call.

Do you have any recommendations or best practices for optimizing audio quality across various Android devices when using LiveKit?

Thanks,

Here is my code
;
import {useEffect} from 'react';
import {
AudioSession,
LiveKitRoom,
useTracks,
TrackReferenceOrPlaceholder,
VideoTrack,
isTrackReference,
registerGlobals,
AndroidAudioTypePresets,
useRoomContext,
useVisualStableUpdate,
useParticipants,
useLocalParticipant,
useRemoteParticipants,
useIOSAudioManagement,
useIsMuted,
} from '@livekit/react-native';
import {Track} from 'livekit-client';import { KrispNoiseFilter } from '@livekit/react-native-krisp-noise-filter';

const wsURL = '********************************';

export default function LivekitJoin() {
const token = useSelector(
(state: RootState) => state.userReducer.liveKitToken,
);
console.log('LivekitJoin token', token);
useEffect(() => {
let start = async () => {
// Configure audio session
AudioSession.configureAudio({
android: {
audioTypeOptions: AndroidAudioTypePresets.communication,
},
});
await AudioSession.startAudioSession();
};

start();
return () => {
  AudioSession.stopAudioSession();
};

}, []);

return (
<LiveKitRoom
serverUrl={wsURL}
token={token}
connect={true}
options={{
// Use screen pixel density to handle screens with differing densities.
adaptiveStream: {pixelDensity: 'screen'},
}}
audio={true}
video={true}>


);
}

const RoomView = () => {
const token = ''
const room = useRoomContext();
const remote = useRemoteParticipants();
const {localParticipant, isMicrophoneEnabled,microphoneTrack} = useLocalParticipant();
const krisp = React.useMemo(() => KrispNoiseFilter(), []);

useEffect(() => {
const localAudioTrack = microphoneTrack?.audioTrack;
if (!localAudioTrack) {
return;
}
console.log("localAudioTrack?.setProcessor" , localAudioTrack?.setProcessor)
localAudioTrack?.setProcessor(krisp);
}, [microphoneTrack, krisp]);

useEffect(() => {
let connect = async () => {
if (!token) {
console.warn('LiveKit token is undefined, skipping connect.');
return;
}
await room.connect(wsURL, token, {});
};
connect();
return () => {
room.disconnect();
};
}, [ token, room]);

useIOSAudioManagement(room, true);

const tracks = useTracks(
[{source: Track.Source.Camera, withPlaceholder: true}],
{onlySubscribed: false},
);
const stableTracks = useVisualStableUpdate(tracks, 5);

const renderLocalTrack = () => {
let localParticipant = stableTracks.find(
item => item.participant.identity === room.localParticipant.identity,
);
if (isTrackReference(localParticipant)) {
return (



);
}
return ;
};

const remoteTracks = () => {
let remotetracks = stableTracks.find(item =>
remote.find(
remoteItem => remoteItem.identity === item.participant.identity,
),
);
if (isTrackReference(remotetracks))
return (



);
return ;
};

return (

{remoteTracks()}
{renderLocalTrack()}

);
};

[JahagirDS0]

@davidliu
Please let me know.

What is the issue with this code?

[JahagirDS0]

@davidliu
Can you please update me on this?
Getting a very strong echo in Samsung's devices