LiveKit Community

Currently you can enforce media types in your application. I.e. only displaying video publishing UI to users with the right permission.

Eventually we'd make it possible to enforce this from permissions / access token too.

Silent install means, fully automated without manual intervention like entering information at the console.

So in the “Deploy to a VM” section on that page describes the steps to deploy to a VM in the “Startup Script” tab. You can write a shell script to automate those steps. (We can’t write it for you because they are highly dependent on your specific VM, and how/where it’s started and run.)

in case you are talking about the js sdk, we are dependent on iOS Safari supporting this API: <https://developer.mozilla.org/en-US/docs/Web/API/Screen_Capture_API>

So yeah, it’s expected that sharing from the browser won’t work on iOS. However all the other livekit-clients that run on iOS support it (flutter, react-native, swift)

By the way, thank you for this amazing product. :heart:

which sdk are you using?

also, if I may ask, why not just wait until the track would be unmuted and publish at that time?

<@U04DED3V5PT> it's not possible to publish a muted track. WebRTC doesn't support this mode of publishing.

Yeah the current docs are somewhat written for folks familiar with their app development environment — the docs tend to be either conceptual overviews or detailed API reference.

Unfortunately we don’t have Angular components or an Angular quickstart, but the js sdk README may be more helpful: <https://github.com/livekit/client-sdk-js>

Oh, yeah, the `Browser` tab is TypeScript without React. (If you’re not familiar with Typescript, it’s just Javascript with some extra syntax to label the types of variables)

(again, it may be more helpful to look at the JS SDK readme, particularly the section on connecting to a room: <https://github.com/livekit/client-sdk-js#connecting-to-a-room-publish-video--audio>)

Ah. yes, the JS SDK was a bit more understandable for me.

```parentElement.appendChild(element);```
Consider this statement. Is "parentElement" like a shorthand here?

As in, "put your parentElement here."

Yeah so that appendChild stuff will probably be handled by Angular for you

But to answer your question, yeah you could either replace `parentElement` with the element you want to add the track element to, or you could make a variable called `parentElement` containing the element you want

Awesome, thank you, Noah, things are getting much more understandable now.

Hello all ! :)

Is anyone interested in making an open source and more detailed guide on how to setup LiveKit on server (self host)? It seems that current documentation isn't very clear for many people, so this would help others a lot.

If anyone is interested, please drop me a message.

Maybe the comment in the code isn't an accurate description of what it is doing.

I'm trying to understand why livekit consistently acquires a device that isn't the top-listed device, since the device enumeration list order is significant.

For now I can just grab the first item in the list and use its device id to constrain the initial video choices, but I don't entirely understand why it has to be this way.

hi <@U02HS7BES0H>,
a previous approach taken by the js-sdk was to try to filter out the duplicate `default` device.
We removed that filtering logic and are now _also_ returning the default device that chrome displays by appending `(Default)` to the device label.

Does this also happen to you on <http://example.livekit.io|example.livekit.io>?

<@U02EU2ZPTQU> Yes, it's identical behavior.

I haven't dug into the livekit code enough to figure out how it's picking but maybe I'll try to do that next. In this particular case, there's no entry for "default" for video on my system being returned by the browser.

(not sure if that's ordinary or not, but it has nothing to do with livekit in any case)

we don't have documentation for it mainly because there's too much variance for us to support.

Currently we support two paths, the VM deployment generator/docs, kubernetes helm charts. Both involve the docker distribution

It's not difficult to set it up on your own though. You'd just need to run the server with its config file, along with opening up ports documented <https://docs.livekit.io/oss/deployment/ports-firewall/|here>.

Thank you very much for your valuable reply. May I please ask two questions?

1. Which path do you use for the cloud service?
2. Isn't a non-docker solution supposed to be faster/more efficient one? If so, would you consider adding it in the future?
Thank you very much for your time!

We run Kubernetes for Cloud. No, there aren't any visible overhead with Docker.

oh okay, I understand. Apart from this, I have 2 last questions:

1. Can we replace caddy with nginx for https?
2. Is caddy/nginx reverse proxy GRPC required for the apps?

sure.. there's nothing specific to Caddy that we need. As mentioned before, we do not use GRPC.

what's your recommended webserber for LiveKit? :slightly_smiling_face:

As <@U02EU2ZPTQU> wrote you yesterday, it seems that safari does not support screen sharing. So I guess the best bet is to make a flutter application in order to be able to do that.

Yes, that would be a good option. Im am just wondering whether there is already something that can used out-of-the-box. Maybe someone did it already.

<@U04D2CC8TTL> it’s not the simplest example (we will be adding more of these in the coming months), but this covers all the things you mentioned:

<https://github.com/livekit/client-example-swift|https://github.com/livekit/client-example-swift>

Hello guys,

I am using livekit react components and i am creating my own stage renderer. I need to use `room` object with in different level of components. To make it easier to use it i tried to save it in to both `context` and  Redux. But i couldn’t make it work for some reason. On redux i got this kind of error   `Type 'Room' is not assignable to type 'WritableDraft&lt;Room&gt;'` .

As far as i understand, it cannot be saved on Redux store because room object is constantly updating i assume. Is anyone of you came upon this problem. If so how did you manage to store it and updated proparly. Because right now i have to pass it to every component as prop which is getting uglier every second.

Any little help would be great! Thank you in advance :slightly_smiling_face:

As the instance below, when doing the ICE negotiation, there are both TCP

selectcandidate.PNG

If udp is not succeed or not stable, tcp will be chosen 

From the packet capture,  I don’t even see udp is tried. How ice conclusion that udp is not stable or not succeed?

It should try udp if it is a first connection for a client , what’s you network config and how do you capture the packets

This is the code snippet for reference.

```function handleTrackSubscribed(
  track: RemoteTrack,
  publication: RemoteTrackPublication,
  participant: RemoteParticipant,
) {
  if (track.kind === Track.Kind.Video || track.kind === Track.Kind.Audio) {
    // attach it to a new HTMLVideoElement or HTMLAudioElement
    const element = track.attach();
    parentElement.appendChild(element);
  }
}```
I can get the audio and video feed from my webcam, I just don't understand how to display it on my webpage.

Akhil, take a look at this <https://github.com/livekit/client-sdk-js/tree/main/example|simple example> to get a better idea of how it works.

As a beginner, I would recommend playing with changing the sample project first to understand the <https://docs.livekit.io/concepts/api-primitives/|concepts> and <https://docs.livekit.io/client/events/|events>.

Thank you, David. Yeah, I've been doing exactly that if not for the heaps of errors I've received while creating an Angular component and pasting the example code into it. Most were jQuery related. It also seems like example depends on the src folder. I believe that's what the
```npm install livekit-client --save```
command is for.

I cloned, did an "npm install" and then did a "yarn run" on sample. The example worked! Happy to see some progress. The yarn command is new to me. I'll look further into my Angular errors and come back if I don't have any luck with fixing them.

Hello <@U04F5J001LG>, could you explain a bit more what you are looking for exactly. I'm not sure if you are aware of their cloud service, which has an easy to use, user interface. If you wish to have a glance: <https://livekit.io/cloud|https://livekit.io/cloud>

Any update? I found Audiosession function but it does nothing. I want to mute whole room at once.

have you tried setting the volume of the audio track? <https://docs.livekit.io/client-sdk-js/classes/RemoteAudioTrack.html#setVolume>

```const muteAudioOutput = () =&gt; {
    setMuted(prev =&gt; !prev);
    room?.participants.forEach(participant =&gt; {
      participant.setVolume(0);
    });```
  };

I did this and still there's no change in volume

<@U01NZKTNE1H> go sdk, nodejs sdk. it doesn't matter

you could use track egress to export the video and then use a separate tool to take screenshots from that at an interval, but we don’t currently have any way to save screenshots

I understand. I will try this. I would like you to consider adding the screenshot feature in the future. thank you.

P.S. : If you try to turn on the tracks after the start of recording, in the case when the tracks were muted at the time the recording started, you can get an error:
```{"nodeID": "NE_LXRg6ZRck3jM", "handlerID": "EGH_tsh3jpY69yq9", "egressID": "EG_Ycf9mbvBrPQn", "element": "GstRtpVP8Depay", "message": "Input buffers need to have RTP caps set on them. This is usually achieved by setting the 'caps' property of the upstream source element (often udpsrc or appsrc), or by putting a capsfilter element before the depayloader and setting the 'caps' property on that. Also see <http://cgit.freedesktop.org/gstreamer/gst-plugins-good/tree/gst/rtp/README>", "error": "No RTP format was negotiated."}

<http://github.com/livekit/egress/pkg/pipeline.(*Pipeline).handleError|github.com/livekit/egress/pkg/pipeline.(*Pipeline).handleError>

/workspace/pkg/pipeline/pipeline.go:799

<http://github.com/livekit/egress/pkg/pipeline.(*Pipeline).messageWatch|github.com/livekit/egress/pkg/pipeline.(*Pipeline).messageWatch>

/workspace/pkg/pipeline/pipeline.go:300

<http://github.com/tinyzimmer/go-gst/gst.goBusFunc|github.com/tinyzimmer/go-gst/gst.goBusFunc>

/root/go/pkg/mod/github.com/livekit/go-gst@v0.0.0-20220603230042-cef031256427/gst/cgo_exports.go:128

_cgoexp_752bbda710ba_goBusFunc

_cgo_gotypes.go:17363

runtime.cgocallbackg1

/usr/lib/go-1.18/src/runtime/cgocall.go:314

runtime.cgocallbackg

/usr/lib/go-1.18/src/runtime/cgocall.go:233

runtime.cgocallback

/usr/lib/go-1.18/src/runtime/asm_arm64.s:1181

runtime.systemstack_switch

/usr/lib/go-1.18/src/runtime/asm_arm64.s:194

runtime.cgocall

/usr/lib/go-1.18/src/runtime/cgocall.go:167

<http://github.com/tinyzimmer/go-glib/glib._Cfunc_g_main_loop_run|github.com/tinyzimmer/go-glib/glib._Cfunc_g_main_loop_run>

_cgo_gotypes.go:2822

<http://github.com/tinyzimmer/go-glib/glib.(*MainLoop).Run.func1|github.com/tinyzimmer/go-glib/glib.(*MainLoop).Run.func1>

/root/go/pkg/mod/github.com/tinyzimmer/go-glib@v0.0.25/glib/gmainloop.go:53

<http://github.com/tinyzimmer/go-glib/glib.(*MainLoop).Run|github.com/tinyzimmer/go-glib/glib.(*MainLoop).Run>

/root/go/pkg/mod/github.com/tinyzimmer/go-glib@v0.0.25/glib/gmainloop.go:53

<http://github.com/livekit/egress/pkg/pipeline.(*Pipeline).Run|github.com/livekit/egress/pkg/pipeline.(*Pipeline).Run>

/workspace/pkg/pipeline/pipeline.go:214

<http://github.com/livekit/egress/pkg/service.(*Handler).Run.func2|github.com/livekit/egress/pkg/service.(*Handler).Run.func2>

/workspace/pkg/service/handler.go:61

2022-12-26T11:09:54.148Z DEBUG egress pipeline/pipeline.go:641 pipeline stopped {"nodeID": "NE_LXRg6ZRck3jM", "handlerID": "EGH_tsh3jpY69yq9", "egressID": "EG_Ycf9mbvBrPQn"}

2022-12-26T11:09:54.151Z DEBUG egress pipeline/pipeline.go:731 removing temporary directory {"nodeID": "NE_LXRg6ZRck3jM", "handlerID": "EGH_tsh3jpY69yq9", "egressID": "EG_Ycf9mbvBrPQn", "path": "/out-m/EG_Ycf9mbvBrPQn/"}

2022-12-26T11:09:54.163Z WARN egress service/handler.go:121 egress failed {"nodeID": "NE_LXRg6ZRck3jM", "handlerID": "EGH_tsh3jpY69yq9", "egressID": "EG_Ycf9mbvBrPQn", "egressID": "EG_Ycf9mbvBrPQn", "request_type": "track_composite", "output_type": "file", "error": "No RTP format was negotiated."}```


what is most unpleasant, a frequent case: at the time the recording started, the participant was with the microphone turned off, and he will not be recorded until he turns on the microphone

it would seem that way: `"request_type": "track_composite"`

<@U04GQNX15PW> I believe you also filed <https://github.com/livekit/egress/issues/203|this issue> on GH?

Yes, that’s trackCompositeEgress
 and yes, that’s me, <@U01KZ61GV5Z>

I would be happy if you share any thoughts on this
:disappointed_relieved:

Denis, the team will follow up on the GH issue when they have a chance.

It's not clear what you are trying to achieve. Can you be more descriptive?

I need from localParticipant.publishTrack function returned RemoteTrackPublication id.

it's the same TrackID. and it does get returned in the `LocalTrackPublication`

image.png

Yes it's same id. Video track is not published.

Good morning people, how are you? Has anyone here already had the experience of uploading the livekit on their own server in a scalable way? In case someone has already run the livekit on their own server for more than 1000 people in streaming?

Ahhhhhhh Android device compatibility :confounded: 

<@U01NZKTNE1H> when you’re online I think this one’s yours

Hey <@U049R9F507Q>, would you be able to get logs from an affected device with debug logs turned on?

<https://docs.livekit.io/client-sdk-js/functions/setLogLevel.html|https://docs.livekit.io/client-sdk-js/functions/setLogLevel.html>

Yes, certainly, will do. Thanks so much. To be sure I understand correctly, that would be done the way it’s done here? <https://github.com/livekit/client-sdk-react-native/blob/main/example/index.tsx#L11>

Hey <@U01NZKTNE1H> finally managed to get my hands on a replicatable device (Redmi/Android 10). I ran on my local machine in dev mode. Here’s a gist with logs: <https://gist.github.com/sruli/1d4429acfc8e99947f006a62196ae9b3>

<@U049R9F507Q> for these devices, are the camera streams viewable locally?

weird. would it be possible to get server logs during when this happens?

We’re on Livekit cloud. Would it help if I gave you the session id?

yeah, that’ll work, estimated timestamp as well

Here’s an example:
session id: RM_ndRDjMCCSXrR
timestamp: 12/27/22 08:51 PM GMT + 2 (10:51 am Pacific)

In that example, there are 3 participants. One of them has 2 published tracks (camera + mic). The other 2 only published mic.

Here’s one more I just did locally:
Session id: RM_oKWtiudzbmTu
Timestamp: 12/27/22 9:43 PM GMT +2 (11:43 am Pacific)

On participant has 2 published tracks (video + mic) and the other has 1 published track (just mic)

can you test trying turning off simulcast with
```new Room({
        publishDefaults: { simulcast: false },
 })```


nice! hopefully that fixes for the other devices as well

simulcast is still in progress for react-native; I’ll make a note in our readme

it’s getting late here. I’ll get out a new build tomorrow. I’d imagine this should do the trick :pray:

<@U01NZKTNE1H> would simulcast need to be `false` for all participants in the room? Or can it be `false` for Android users and `true` for iOS? Just thinking, in terms of how to test/release this, if we need to make sure all users have the same app version.

It just needs to be false for the affected devices

Hm what is your goal here, I mean, why is it important to reduce this time?

(Not asking out of skepticism, just want to understand your goals to better help)

<@U02MC5V3C2H> I need to identify the participant who did not exit the meeting properly (by closing the application completely, disconnecting the internet or running out of battery, etc.) in a short time. We need to check this participant in our own API.

<@U01KZ61GV5Z> self-deploying on kubernetes

And I tried to unpublish/publish audio track 3 times, But it's not getting Remote Track published event

<@U01KZ61GV5Z> Can you help for this issue?

This issue is the urgent for me. Can you help how to fix this issue?

hi <@U04DED3V5PT>
please provide some reproduction, best in the form of a working example app or at least an extensive code snippet.
Without some more context it’s hard to say what’s triggering your issue.

```  public async stopAudioTrack() {
    const audioTrackPublication = this.participant?.getTrack(Track.Source.Microphone);

    if (audioTrackPublication &amp;&amp; audioTrackPublication.audioTrack) {
      await this.participant?.unpublishTrack(audioTrackPublication.audioTrack.mediaStreamTrack);
      this.audioPublication = undefined;
    }
  }```
This is unpublish current audio track. After call this, I tried to publish new track.
And I tried this way for 2 times

```await this.stopAudioTrack();
await this.participant?.publishTrack(audioTrack, {
            name: generateRandomString(6),
            source: Track.Source.Microphone,
          });```

thanks, can you also provide the parts where you listen to the publish/unpublish events and how the unpublish/publish sequence get’s triggered?

After connect to the room, there is no issue. But when I try to publish/unpublish, it's not getting the event sometimes

thanks for the info, are you using livekit cloud or a self-deployed instance?

if the issue is urgent for you, would it work, as a workaround, to mute the track instead?

When the user enable/disable noise reduction, It's trying to unpublish/publish the audio track.

So at that time, it's broken to get the event track published after get track unpublished

aha! In that case the API you want to use is `localAudioTrack.restartTrack({noiseSuppression: true/false})`
there’s no need to unpublish/republish the track.

Sorry. we attached our own audio track with noise reduction

So I used `replaceTrack`. But `replaceTrack` has the issue with noise reduction

is this an external noise reduction that you’re applying?

what kind of issues are you seeing with the `replaceTrack` approach?

So I hope you help me why it's not getting `TrackPublished` event sometimes when the user try to publish/unpublish

<@U04DED3V5PT> please remember that this is a public support channel where folks are helping you at no cost.

Some etiquette that's worth called out: please avoid tagging individuals unless you are already in a conversation with them. and do not expect immediate help even if the issue is urgent to you. 

thank you for understanding

another tip: you are more likely to get a high quality answer if you are making it easy for others to understand what you mean. I do not recommend posting screenshots with scribbles or ChatGPT dialogue.. we will not be able to read them

<@U04DED3V5PT> this PR fixes the timing issues of unpublish/publish <https://github.com/livekit/client-sdk-js/pull/541>

this is expected, the track will only (reliably) be present, once `TrackSubscribed` has fired!

Hey <@U03MFML4QH1>, this certainly sounds useful! Though it's currently not on our radar. Would this be something that you can implement by yourself? i.e. by placing VideoView inside of another View, and implementing gesture recognizers on the container itself?

You should be able to transform the underlying VideoView with Zoom/etc.

but if you feel that this would be better implemented within the SDK itself, feel free to open an issue (and PR as well if you are up for it)

cc <@U02MC5V3C2H> re: request for tutorial

:wave:  regarding egressClient initialization, take a look at the other examples <https://docs.livekit.io/server/egress/room-composite/#starting-a-roomcomposite|here>. Noted that it's not obvious when you are just looking at TrackEgress. We'll make improvements.

Thanks for your time.
For now I’m thinking of implementing webhooks to get the trackIds, and use TrackEgress to send the tracks via websockets, is this the correct approach?

Yup, that seems right. Or you could have your clients call a method on your server after `publishTrack`.

hey <@U03T51VH7JS>, just a friendly reminder to pls thread your messages :pray:

From the log, it seems to be related to the camera, and an exception was thrown when setting the frame rate, which may need to be debugged with webrtc native code, because the exception did not occur on the flutter side
```Crashed Thread:        1  Dispatch queue: org.webrtc.RTCDispatcherCaptureSession```

When I have sometime to compile framework in debug mode and get exact error, I will ping you back with exact line of code.

I this crashing on a camera track? or is it a screen share?

Got it. Which type of camera are you using?

I test it on two virtual cameras, OBS and fine cam and iphone continuity camera

all works fine on my intel based machine btw

ah, virtual cameras. does it work with a standard camera on device?