LiveKit Community

logs-from-egress.log

Hi, I upgraded egress v1.7.2 to v1.7.5 and I get below warning. When I downgrade to v1.7.2 again, there is no warning.

Hey <@U05GQ7H856C>, yes! You can start a <https://docs.livekit.io/server/egress/room-composite/|RoomCompositeEgress> with Mux's RTMP ingest endpoint.

We are currently building an adapter so that process is more straight forward.

Can you try with the latest versions? there was a bug at one point where we didn't properly detect that the publisher is still in the room before terminating the recording. I wonder if that's what you are seeing

getting mic code :

```      const stream = await navigator.mediaDevices.getUserMedia({
        audio: {
          sampleRate: 48000,
          deviceId: deviceId,
          echoCancellation: true,
          noiseSuppression: true,
          suppressLocalAudioPlayback: true,
          autoGainControl: false
        }
      })

      micTrack = stream.getAudioTracks()[0]
      await this.room.localParticipant.publishTrack(micTrack, {
        source: Track.Source.Microphone,
        red: true,
        dtx: true,
        audioPreset: AudioPresets.musicHighQualityStereo,
        forceStereo: true,
      })```

getting camera code :

```      const stream = await navigator.mediaDevices.getUserMedia({
        video: {
          facingMode: {ideal: "user"},
          frameRate: {
            ideal: 15,
          },
          width: {
            ideal: 300,
          },
          height: {
            ideal: 300,
          },
          aspectRatio: {ideal: 1}
        }
      })

await this.room.localParticipant.publishTrack(stream.getVideoTracks()[0], {
        source: Track.Source.Camera, videoEncoding: {maxBitrate: 32 * 8 * 1000, maxFramerate: 15}
      })```

getting screen share code :

```      const stream = await navigator.mediaDevices.getDisplayMedia({
        video: {
          frameRate: {ideal: 15},
        }
      })

     await this.room.localParticipant.publishTrack(screenShareTrack, {
        source: Track.Source.ScreenShare,
        videoEncoding: {maxBitrate: 32 * 8 * 1000, maxFramerate: 15},
        screenShareEncoding: {maxBitrate: 32 * 8 * 1000, maxFramerate: 15}
      })```

writing audios/videos to file code using go sdk :

```func NewTrackWriter(track *webrtc.TrackRemote, pliWriter lksdk.PLIWriter, fileName string) (*TrackWriter, error) {
	var (
		sb     *samplebuilder.SampleBuilder
		writer media.Writer
		err    error
	)
	switch {
	case strings.EqualFold(track.Codec().MimeType, "video/h264"):
		sb = samplebuilder.New(maxVideoLate, &amp;codecs.H264Packet{}, track.Codec().ClockRate, samplebuilder.WithPacketDroppedHandler(func() {
			pliWriter(track.SSRC())
		}))
		writer, err = h264writer.New(fileName + ".h264")

	case strings.EqualFold(track.Codec().MimeType, "audio/opus"):
		sb = samplebuilder.New(maxAudioLate, &amp;codecs.OpusPacket{}, 48000)
		writer, err = oggwriter.New(fileName+".ogg", 48000, track.Codec().Channels)

	default:
		return nil, errors.New("unsupported codec type")
	}

	if err != nil {
		return nil, err
	}

	t := &amp;TrackWriter{
		path:   fileName,
		sb:     sb,
		writer: writer,
		track:  track,
		room:   s,
	}
	go t.start()
	return t, nil
}


func (t *TrackWriter) start() {
	for {
		if !t.room.IsRecording {
			break
		}
		pkt, _, err := t.track.ReadRTP()
		if err != nil {
			break
		}
		t.sb.Push(pkt)

		for _, p := range t.sb.PopPackets() {
			t.writer.WriteRTP(p)
		}
	}
}```

<@U01KZ61GV5Z> here is the codes , can you take a look please ?

<@U05BRQ8Q9D1> which cloud provider are you running this on?

I'm not familiar with Oracle Cloud: our helm charts have not been tested to work with it. So you'd likely have to do a bit of troubleshooting and adapting them to make it work

image.png

Hey <@U05GTMD2F3J>, are you using dynacast here? When dynacast is enabled, it would pause all of the streams that are not visible

currently the default component is not tuned for large rooms. we do plan to update this in the near future where it would use <https://docs.livekit.io/client/receive/#selective-subscription|selective subscriptions> instead to page through video tiles. Paging is how folks are typically handling large numbers of participants with LiveKit.

CC: <@U03RM7G8Q6A> <@U02EU2ZPTQU> for visibility

by default the optimizations for this are not enabled in our SDK.
try to pass in `options={{ adaptiveStream: true, dynacast: true }}` on the LiveKitRoom component

you can try to set the publishing codec to AV1, take a look at our JS sample app as an example

Audio loss is usually CPU related, and with 4 cores and 1080p that sounds likely. Right now you can't set default dimensions in the config, but you can use one of the 720p presets in your egress requests

Thanks, may i know where can we set egress preset ? Is it with RoomComoposite.start(). Any resource or documents link on preset setting is highly appreciated.

<@U045CJPACG1>, take a look at the docs here: <https://docs.livekit.io/server/egress/#startroomcompositeegress>

Hi Team,

We have set <https://docs.livekit.io/server/egress/#encodingoptionspreset|EncodingOptionsPreset> to H264_720P_30: 0, but we still observing cracked voice. This was seems to be caused by recent livekit-server latest version moving to v1.4.3 (<https://github.com/livekit/livekit/releases/tag/v1.4.3>) and livekit-egress v  1.6.0 (<https://github.com/livekit/egress/releases/tag/v1.6.0>).

Any idea what would be the cause of this ?

Can you try the latest versions both livekit-server and egress? and what is your CPU utilization while this is observed?

Noted , will try latest version and confirm. Will share the metrics as well.

Screenshot_20230710-153117.jpg

Screenshot_20230710-153114.jpg

Can you tell me which livekit sdk version you are using,
<https://github.com/flutter-webrtc/flutter-webrtc/pull/1349> There is an enumeration fix for bluetooth devices here.
but you need to add
```  &lt;uses-permission android:name="android.permission.BLUETOOTH_CONNECT" /&gt;
   &lt;uses-permission android:name="android.permission.BLUETOOTH_ADMIN" android:maxSdkVersion="30" /&gt;```
In order to get the correct name of the Bluetooth device.

<https://github.com/livekit/client-sdk-flutter/tree/main#android>
``` &lt;uses-permission android:name="android.permission.BLUETOOTH_CONNECT" /&gt;```
You can try to add the permission label above, and use the code here to apply for bluetoothConnect permission, then you should be able to get the correct bluetooth device label.
```import 'package:permission_handler/permission_handler.dart';

Future&lt;void&gt; _checkPremissions() async {
  var status = await Permission.bluetooth.request();
  if (status.isPermanentlyDenied) {
    print('Bluetooth Permission disabled');
  }
  status = await Permission.bluetoothConnect.request();
  if (status.isPermanentlyDenied) {
    print('Bluetooth Connect Permission disabled');
  }
}

void main() async {
  WidgetsFlutterBinding.ensureInitialized();
  await _checkPremissions();
  runApp(MyApp());
}```

Some service on your machine is using the same port `7881`

I checked. but no such service is coming.. I tried restarting my machine as well

For your information, this is how you thread your messages.

If you post in the main thread of the channel, all users of the channels will be notified. If you post in a thread like this, only those who are involved in the conversation will be notified. :slightly_smiling_face:

okok.. got that "will don " repy in thread " now

You can run something like this to see what process is using your ports:
```lsof -Pi | grep LISTEN```

shgupta@INGUWN171177:~$ sudo lsof -i :7881
[sudo] password for shgupta:
COMMAND    PID USER   FD   TYPE DEVICE SIZE/OFF NODE NAME
livekit-s 3673 root   24u  IPv6  49534      0t0  TCP *:7881 (LISTEN)

this is the output i got.. shall i kill this process?

If you want to free the port, you will have to end the process, yes.

i tried..  sudo kill -9 3673 .. but this is not working

How did you install livekit server, how did you start it? Please provide detailed steps of what you did.

i followed this page: <https://docs.livekit.io/getting-started/server-setup/>

used following commands: curl -sSL <https://get.livekit.io> | bash

started with this command:livekit-server --dev

now this livekit process getting start again and again

I don’t know how the Livekit server is started, but it sounds like it runs as a service that restarts automatically.
Can you check if this is the case? You could start by listing all services like this:
```systemctl list-units --type=service --all```

And then you could stop the service with something like:
```sudo systemctl stop &lt;servicename&gt;```

got this output on running the above command:

shgupta@INGUWN171177:/etc$ sudo systemctl status livekit-docker.service
● livekit-docker.service - LiveKit Server Container
     Loaded: loaded (/etc/systemd/system/livekit-docker.service; enabled; vendor preset: enabled)
     Active: failed (Result: exit-code) since Mon 2023-07-10 19:11:35 IST; 1min 1s ago
    Process: 3902 ExecStartPre=/usr/local/bin/docker-compose -f docker-compose.yaml down (code=exited, status=0/SUCCESS)
    Process: 3914 ExecStart=/usr/local/bin/docker-compose -f docker-compose.yaml up (code=exited, status=1/FAILURE)
    Process: 23679 ExecStop=/usr/local/bin/docker-compose -f docker-compose.yaml down (code=exited, status=0/SUCCESS)
   Main PID: 3914 (code=exited, status=1/FAILURE)

Jul 10 19:11:35 INGUWN171177 docker-compose[23679]:  Container livekit-caddy-1  Removing
Jul 10 19:11:35 INGUWN171177 docker-compose[23679]:  Container livekit-livekit-1  Removed
Jul 10 19:11:35 INGUWN171177 docker-compose[23679]:  Container livekit-caddy-1  Removed
Jul 10 19:11:35 INGUWN171177 docker-compose[3914]: Error response from daemon: No such container: ce25b9f20bf8799793dd14da6daabf921f67bf7e85bf70ed7c646f833ea4539a
Jul 10 19:11:35 INGUWN171177 systemd[1]: livekit-docker.service: Main process exited, code=exited, status=1/FAILURE
Jul 10 19:11:35 INGUWN171177 docker-compose[23679]:  Container livekit-redis-1  Stopped
Jul 10 19:11:35 INGUWN171177 docker-compose[23679]:  Container livekit-redis-1  Removing
Jul 10 19:11:35 INGUWN171177 docker-compose[23679]:  Container livekit-redis-1  Removed
Jul 10 19:11:35 INGUWN171177 systemd[1]: livekit-docker.service: Failed with result 'exit-code'.
Jul 10 19:11:35 INGUWN171177 systemd[1]: Stopped LiveKit Server Container.

when i check for livekit server process :

ps -ef|grep livekit
root       16978   16957  0 19:09 ?        00:00:00 /livekit-server --config /etc/livekit.yaml
shgupta    24127    1930  0 19:14 pts/2    00:00:00 grep --color=auto livekit

it seems this process is running again and again

I’m not very familiar with the LiveKit server setup. Let’s wait for some of the folks from the Pacific time zone to come online.
Did you just follow this <https://docs.livekit.io/getting-started/server-setup/|guide> or did you try to install LiveKit server in another way? There seems to be some Docker involved and I’m not sure if this is the case with the default setup from the example.

and the again downloaded the livekit and restart the server

one more query i have .. where i can check the logs of livekit call logs

actuallly I need to understand the call flow of livekit by mapping logs and code ..

You are looking for the LiveKit server logs?

i need to understand the call flows of livekit .. with code mapping

If you run `livekit-server --dev`, the logs will be printed directly to your console.

ohh ok.. Now i see .. i am nt able to connect the call .. getting this: 2023-07-10T19:32:27.916+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}
2023-07-10T19:32:27.955+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}

any idea how to correct it ?? or what wrong i am doing ?

i generated the keys on server  then started the server

then generated the token on client like emntioned on this page: <https://docs.livekit.io/getting-started/cli-setup/>

and then put the generated token in this app in custom section: <https://meet.livekit.io/?tab=custom>

Is your LiveKit server exposed to the Internet or are you just running it locally as described in the guide?

Then you can’t use <https://meet.livekit.io/?tab=custom> because there is no way to connect to your local server.

<ws://localhost:7880/> this is the url i am using

then how can i connect to local server any other client ?

so any other way I can run calls and get logs and understand the flow

I need to understand the flows of these : create conf
join conf
rejoin conf
mute others
mute all
mute conf

One question, what are you trying to accomplish? If you want to do some testing or play around with some client code, you can use the free LiveKit Cloud plan. Then you don’t have to worry about the server setup part.

You get 50GB of free bandwidth and don’t have to enter your credit card information. <https://cloud.livekit.io>

actually i wanna understand the server code flow and call logics

i am mainly need to play around server code not the client code :slightly_smiling_face: need to understand its features

Ok, if you want to deploy LiveKit yourself, this is the docs section: <https://docs.livekit.io/oss/deployment/>

only after this deployment i can run the server locally and egt the logs ??? no other way ??

because i tried this also .. got stucked in multiple issues

No, you can connect to the local server just like any other server. Just replace the URL with your local server’s URL. Like in the manual: <https://docs.livekit.io/getting-started/cli-setup/#simulating-another-publisher>

You can also download the example app and update the `.env` file with the details from your server: <https://github.com/livekit-examples/meet?ref=meet>

./livekit-cli join-room     --url <ws://localhost:7880>     --api-key APIEuhDh8knPZeN --api-secret eTCfXuDVeoU
c6uNtYMVofSQ0Le0A0t1lgeuuQvBpxuMD     --room my-first-room --identity bot-user     --publish-demo
2023/07/10 19:55:58 "msg"="error establishing signal connection" "error"="websocket: bad handshake" "httpResponse"={"Status":"401 Unauthorized","StatusCode":401,"Proto":"HTTP/1.1","ProtoMajor":1,"ProtoMinor":1,"Header":{"Content-Type":["text/plain; charset=utf-8"],"Vary":["Origin"],"Date":["Mon, 10 Jul 2023 14:25:58 GMT"],"Content-Length":["32"]},"Body":{"Reader":{}},"ContentLength":32,"TransferEncoding":[],"Close":false,"Uncompressed":false,"Trailer":{},"Request":{"Method":"GET","URL":"<http://localhost:7880/rtc?protocol=8&amp;sdk=go&amp;version=1.0.13&amp;auto_subscribe=1>","Proto":"HTTP/1.1","ProtoMajor":1,"ProtoMinor":1,"Header":{"Upgrade":["websocket"],"Connection":["Upgrade"],"Sec-WebSocket-Key":["Dx+1v6WhoCyvmYLj9gYFgQ=="],"Sec-WebSocket-Version":["13"],"Authorization":["Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE2ODkwMjA3NTgsImlzcyI6IkFQSUV1aERoOGtuUFplTiIsIm5iZiI6MTY4ODk5OTE1OCwic3ViIjoiYm90LXVzZXIiLCJ2aWRlbyI6eyJyb29tIjoibXktZmlyc3Qtcm9vbSIsInJvb21Kb2luIjp0cnVlfX0.bZ1XsMLES0xkrrbnJsWzj4jJ_mesC54k1ntqcU3N3oE"]},"Body":null,"GetBody":"&lt;unhandled-func&gt;","ContentLength":0,"TransferEncoding":[],"Close":false,"Host":"localhost:7880","Form":{},"PostForm":{},"MultipartForm":null,"Trailer":{},"RemoteAddr":"","RequestURI":"","TLS":null,"Cancel":"&lt;unhandled-chan&gt;","Response":null},"TLS":null}

basicallly we need to start client locally also and not on the internet

you can test your localserver with this app

Follow the getting started here: <https://github.com/livekit/livekit#getting-started> and then pass in the generated access token into the token field in the app and click connect.

2023-07-10T19:55:58.668+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}
2023-07-10T19:55:58.675+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}
2023-07-10T20:03:19.964+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}
2023-07-10T20:03:19.965+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}
2023-07-10T20:03:39.528+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}
2023-07-10T20:03:39.548+0530    WARN    livekit service/auth.go:69      error handling request  {"status": 401, "error": "invalid API key: APIEuhDh8knPZeN"}

Did you follow the instructions step by step?
```livekit-cli create-token \
    --api-key devkey --api-secret secret \
    --join --room my-first-room --identity user1 \
    --valid-for 24h```

earlier i was using my self generatd key and secret

Solved in another thread, it was a custom template not calling `startRecording`

how to revert to 1.4.3 with install script?

<@U05FUDL1QLX> pls thread your message if they are related. There are over 2K folks in this channel. :pray::skin-tone-4:  

You need to generate two tokens for the room. Each participant must have a unique identity when generating the respective tokens.

I don’t think we have official documentation for it yet, but here’s an early release of track processors:

<https://github.com/livekit/track-processors-js|https://github.com/livekit/track-processors-js>

Do we have any UIKit sample code <@U02MAJLGN65>?

here's an example: <https://github.com/livekit/client-example-collection-swift/tree/main/uikit-minimal>

Hey <@U027V3P9UQL> at this time, LK open source does not support server cascading. Could you say more about your use case?

wanted to create a globally distributed system
where multiple sfu's talk to each other and provide lowest latency possible to users

but the crux is lowest possible latency
sfu -&gt; user and vice versa

This is how LiveKit Cloud works, is there a particular reason to not use it?

no no 
i was asking particularly about cascading

in livekit cloud all participants in one room connects to one particular sfu right?

multiple rooms can be distributed across region

No, they all connect to the nearest edge with LiveKit Cloud

<https://blog.livekit.io/scaling-webrtc-with-distributed-mesh/|https://blog.livekit.io/scaling-webrtc-with-distributed-mesh/>

Yes, I believe we use the term “relayed” instead of cascaded

got it
maybe the term is different

ill look into it, if there's any doubt ill ask you again

thanks :raised_hands:

where can i find docs for low apis to relay media from sfu to other?
for something like this?
<@U01KZ0SCDKL>

Hey <@U027V3P9UQL>, we do not offer low-level APIs to control the relaying in the open source SFU. Cloud is a distributed mesh network, the open source SFU is a single-homed SFU.

The part that does the distributed mesh is not open source today. But we do have an enterprise license for companies that are interested in deploying their own Cloud.

got it
thanks for clarifying <@U01KZ61GV5Z> :pray:

Hey, I should be able to release a fix next week :slightly_smiling_face:

Okay! Thanks for the info. I'll look forward to the fix.

As a follow-up, I'm not having success publishing a texture. Publishing mic is great (as is receiving both texture and mic), but the sample code to publish a texture doesn't seem to be working. Is that functionality supposed to be working in the dev preview, or is that also something on its way?

I see. Ok I'll keep fiddling with it, thank you.

Don’t hesitate to send me your code so I can take a look :slightly_smiling_face:

Okay thanks, I'll probably send you some today. I'm going to be spending most of my time integrating LiveKit with Unity, so if there's anything you could use help testing or anything like that, I'd be happy to help out. Would issues on GitHub be helpful?

Yes issues would be very helpful :slightly_smiling_face:
If you’re encountering weird bugs it is also very relevant to send your <https://docs.unity3d.com/Manual/LogFiles.html|Editor.log> with them.

Hey <@U022KB72NFJ> I created a few issues for you. Please let me know if you need any further details. I will be available for any tests you need or assistance I can provide. My company is moving forward with LiveKit and are very interested in it's development!

Hey, thanks a lot <@U057CJT8E3Y>. Looking into it :slightly_smiling_face:

Hey <@U022KB72NFJ>, from your GitHub activity it looks like you're busily working on several other areas of LiveKit! Is there anyone actively working on the Unity SDK?

<@U057CJT8E3Y> Theo does indeed work on a lot of stuff. We are actively working on the Unity SDK, but that SDK is powered by the Rust SDK, which is also under active development (and has users/customers). Rest assured, we haven’t forgotten about the issues you’ve filed. We also happily consider and accept PRs :slightly_smiling_face:

Thanks <@U01KZ0SCDKL> I definitely had the impression no one had even looked at the tickets yet. It's been a few weeks since the initial guestimate for a new release - any chance there's a new guesstimate?

Hey Abhishek, have you checked out the Flutter quickstart?

<https://docs.livekit.io/getting-started/quickstarts/flutter/|https://docs.livekit.io/getting-started/quickstarts/flutter/>

Yes
but i got confused
I have some doubts

Is there a way to create a room using liveclient sdk?

by default, rooms are created automatically when you <https://docs.livekit.io/client/connect/#connecting-to-a-room|connect to it>.

if autoCreate is disabled, you'd need to use one of our server SDKs to create it explicitly

when we connect to the room
what do we have to pass as a url?
 LiveKitClient.connect(url, token,roomOptions: roomOptions)

are you using Cloud? if so it should be on the project dashboard in the top-left corner. it'll be something like `wss://&lt;your-subdomain&gt;.livekit.cloud`

can we set up a meet?
like i have others doubt and it it not convenient for me to write all of them here

<@U05GKPHTJ84> unfortunately the team is pretty swamped, so it’s preferred to discuss these things async. Please share your concerns/questions here and we will respond when we have a chance.

You can try override dependencies in `pubspec.yaml`
<https://dart.dev/tools/pub/dependencies>

```dependency_overrides:
  protobuf ^3.0.0```

ahhha thank you!! I am surprised I couldn't find this online.

Appreciate this :pray:

Fixed by using keys from <https://randomkeygen.com/>

<@U05GNLDLB7W> I don’t know much about Helm charts, but you can generate and API key/secret pair with:
`livekit-server generate-keys`

Hi there, yes, you can publish a camera or screen track directly without going through OBS. What exactly are you trying to do?

I want to create audio-only live-streaming like Twitter Space <https://help.twitter.com/en/using-twitter/spaces>.

Gotcha, yeah this is definitely possible. Here’s an open source project (uses LiveKit) which might help you get started:

<https://github.com/rallydotfm/rally|https://github.com/rallydotfm/rally>

Hi, I have started to mess around with the native Unity SDK. I know it's in early development but I was curious if it is able to publish a data track yet. I couldn't seem to find a relevant function. <@U022KB72NFJ>

<https://www.jetbrains.com/go/|https://www.jetbrains.com/go/>

Those are generated files to satisfy some interface so that they can be used as a mock implementation in tests like you mentioned.

When we  mute any participant : HandleParticipantSignal is the place where code hits the first ?

like how does live kit server recieves info that participant is muted

Hello <@U05FUDL1QLX>, have you had the chance to go through the blog section? There's a lot of great information around the structure which LiveKit has built &amp; put together. <https://blog.livekit.io/> :slightly_smiling_face:

After building what I'm working on I think I will do a video course of Livekit, it's really needed.

is there any video course available ?? can you please share the link for that

this is how i will solve this issue first give all subscribe participant the "_*subscribed only token*_", after which as the publisher you can update a particular user token to allow _*canPublishSources [camera, or mic, etc]*_

so by so doing you have provided priority to your speaker

or another way
during a time a participant is publishing to the room, give him the _*roomAdmin*_ privilege which will make him a _*host,*_  then make other publishers have t _*canPublish, canPublishData*_ privileges, so the Host can then be able to update the token access of the  publishers (speakers)

Thanks a lot! Seems like the way to go here :slightly_smiling_face:

Do you think that this approach is agile enough for a live session? Degrading all tokens privileges every time a varying speaker is publishing?

Yes,

According to the documentation, token can be updated via the client or via the backend

Of course, but the question is whether or not it’s a good practice to change privileges this frequently though :thinking_face:

The Livekit listen for updates, it wont have be a good idea if they didn't introduce that that, for them introducing it in the first place it's a good idea.

Feel free to do that way

during a time a participant is publishing to the room, give him the _*roomAdmin*_ privilege which will make him a _*host,*_  then make sure other participant (subscribers) does not have the _*canPublish, canPublishData*_ privileges, so the Host can then be able to update the token access of the  publishers (speakers)

Can you provide some code instructions on how to give participant the roomAdmin privilege in flutter?

To my best knowledge you need to give the roomAdmin privilege at the backend during time of Jwt Authentication, according to the documentation of Livekit.

What backend language are you using???

so before generating token we have to provide some sort of admin Privilege??

Yes provide it among the objects that will be tokenized

I advice you go through the documentation for a proper understanding

hmm not too familiar with the bluetooth modes. I’m looking into the bluetooth mic bug you linked, but it sounds like the hfp vs a2dp is a different issue?

also, just for reference, what android os version have you seen this on?

The reporter is a team member of me. So its all in the issue reporting. e.g. Android 12

thanks for having a look :slightly_smiling_face:

Regarding the not switching vom hfp to a2dp: Thats something different, yes. I don’t know if this is intended or a bug. Do you think its a bug? We can write an issue as well if you like. We are not sure because it behaves same in iOS.

To the clear: Fixing either one of those two bugs could solve the situation for us. Mainly we want to achieve high quality music (a2dp) while muted OR not publishing track. Which way is recommended?

hmm, I think unpublishing would be the way to go, since muting only mutes the input but is still “recording” for android purposes and prevents the switch over to a2dp. will see about making this experience better

In that case, it would be great that it wouldn’t change the audio input device, right? :slightly_smiling_face:

<@U01NZKTNE1H> we can confirm that on iOS the mic is not switching if no one else is publishing. 

<@U051HC0QG9E> hey I pushed a fix for the bluetooth microphone issue, can you try version 1.2.2-SNAPSHOT?

will need to add `maven { url '<https://s01.oss.sonatype.org/content/repositories/snapshots/>' }`  to your repositories to access the snapshot version

Thanks <@U01NZKTNE1H>, I’ve forwarded it to our Android Lead Dev <@U05H6P8QAM7>

<@U01NZKTNE1H> Thanks for the update, I've tested in our app using the Snapshot version and sadly it's still same behaviour.
Also I pulled the latest main from the repository, trying with the same change  in `callViewModel` with sample-app-compose
in line `169` `localParticipant.setMicrophoneEnabled(false)`
Same behaviour.
I'm testing with An android 12 device and Android 11 device.
The code in the commit `90648076f8d0b490ee3e3789b4482dc3108cf3cd`
is executed, i.e. the `audioManager.setCommunicationDevice(audioDeviceInfo)` in the `AudioSwitchHandler` is executed on the Android 12 device, but still same behaviour.

It happens with older versions too.

Can you let me know in which scenario was the fix working for you?

something that can be related, in the `AbstractAudioSwitch` class
line 167 there's a comment related to webrtc that on Activating device it says
```// Always set mute to false for WebRTC```
I'm not sure if that's related?
If you would like me to continue discussion on the github issue please let me know.

just as a sanity check, is audioDeviceInfo
```info.type == AudioDeviceInfo.TYPE_BLUETOOTH_SCO || info.type == AudioDeviceInfo.TYPE_BLUETOOTH_A2DP```


Yes, the info type for the devices I'm testing with is:  `AudioDeviceInfo.TYPE_BLUETOOTH_SCO`

I mean is that the one that’s being set in the code here? just to be sure

I'm not sure I get you. if you mean in the AudioSwitchHandler&gt; the `when (selectedAudioDevice)` line 87, yes it's that device that is set.

<@U01NZKTNE1H> Do you need to reopen the github issue or do you track somewhere else too?

:slightly_smiling_face: Please let us know if you need more feedback/testing etc. Looking forward to this fix. Its pretty crucial for us as you can imagine. Thanks for your support!

will do. <@U05H6P8QAM7> as a mini update, I *think* I’ve identified the issue to the audio mode; on Android 12 at least (I think may affect android 11 as wel), it seems like MODE_COMMUNICATION just doesn’t support using the bluetooth mic without any corresponding audio output. however, there seems to be an additional issue somewhere compounding this affecting our library specifically

Thanks for the update! Very good find! If you would like me to test some scenario please let me know!

ok, pushed another fix, try the latest snapshot

<@U01NZKTNE1H> Thanks for the fix, I can confirm that it's fixed on the sample app on Android 12 devices and 11 too. However on our app it's working with 12, but not with 11, it's mainly related to that we are trying to use Focus Mode `AUDIOFOCUS_GAIN_TRANSIENT`  on joining room and we try to let media playing in the background to still be audible while in call and the user is muted. so when the user is muted we switch to `audioManager.mode= MODE_NORMAL` which works well on Android 12 after your fix, but not on Android 11.
So the end goal is, if the user is in a call and muted they should be able to hear music and the remote participants clearly at the same time.
Once the user is unmuted, we start bluetooth sco and switch mode to `MODE_IN_COMMUNICATION`  sadly that's not working on android 11 but I'm looking more into it.

sorry, which part is the bug on android 11? the music and remote participants aren’t playing at the same time?

I would say that the bug is fixed ( in regards that if there are no other published audio tracks, the current audio input is taken correctly from the bluetooth device even if I'm the only unmuted user )
The problem with music, is that, on connecting to the room, the AudioSwitchHandler calls audioSwitch?.start() and audioSwitch?.activate().
That let's the `AudioDeviceManager` to call `setAudioFocus`.
If the user is playing music and then joins a room, the app takes the main focus and the playback is stopped.
How we countered this in the app is to call audioSwitch.activate and then audioSwitch.deActivate right after the room is connected. so that the app releases the audio focus and the user can still listen to music. ( that makes the communication mode back to normal, and bluetooth sco is stopped )
and when the user unmutes we set communication mode to IN_COMMUNICATION and start bluetooth sco.
It's a bit hacky but it was working, but now it only works for android 12, not for 11. so it's more of a feature request than a bug, the end goal is:

User when join a room should be able to continue listening to music and other participant as long as the user is muted.
If the user unmutes, then it's okay for the music to be inaudible (due to MODE_IN_COMMUNICATION) and the audio input should still be from bluetooth device.

Gotcha, can you file a separate bug for this?

<@U01NZKTNE1H> Thanks will do, I'm just trying to pinpoint the issue. and then will create the bug. thanks a lot for your support :raised_hands:

<@U01NZKTNE1H> Is there a place where we can ask for feature requests? or it can be done through a github bug? (although it's a feature more than a bug :smiley: )?

File an issue, there should be a feature request template

hi <@U01NZKTNE1H> I've created a feature request regarding our usecase: <https://github.com/livekit/client-sdk-android/issues/252>
To handle this I had to use `NoAudioHandler` and had to write our own device switching implementation. It would be nice to have this as an option though in the sdk.
Thank you!

Sounds good, I can add a flag pretty easily