archive-voice

[carmine timber]

Try to update to latest dev build .

and then check

[wise anchor]

ight

works now ty

[carmine timber]

np

[lost lynx]

FYI @vocal valley I succeeded in using the adapter, and using this new voiceConnection and play audio succesfully from my system. When I deployed to my host, it still gave an error, but this time I was able to track it to the UDP connection.

So, this means my Hosting provided might be blocking the UDP connection. I'll try to contact them about that...

But ti also means that in the regular Voice module of discord.js (not the new @discord/voice, which is OK), UDP error are not trapped by try/catches.

Anyway, thanks guys for your help! Hopefully my Hosting provided will be receptive... !

[vocal valley]

no problem!

Good luck with getting that resolved

[wise anchor]

how can i bassboost a song?

[vocal valley]

outside the scope of this lib, presumably you could do it with ffmpeg

[wise anchor]

ight

[last mist]

Mind if we have a (hopefully) short discussion on how to install the basics and get an audio file to play? I had it working in v12 obviously it is different now. I see the pinned messages. I'm looking at the preview-595 guide as I type. Should I type the npm install line with "libsodium-wrappers" it specifically says "for example" which sounded like I could change that to sodium for "(best performance)". Doing that didn't work the install barked and logged something I had never seen before re: please set your MS VS version before you run install. I see the tip in the doc about "if you are struggling" is this what is meant? and am I correct in assuming that one types in --vs2015 regardless of what version of VS one is using even if it is VS Code?

Oh, the log file BTW mentions "npm config set msvs_version 2015" as a way to set the version do we believe that is more or less equivalent to the TIP instruction?

[steel crescent]

Don't bother installing sodium on windows.

[last mist]

okay thanks

I tried the line mentioned in the log, it didn't help as the installer now wanted a copy of VS2015 to be around

libsodium-wrappers worked

the generateDependencyReport report looks reasonable

I seem to have tweetnacl: 1.0.3 automatically because I didn't ask for it

[brazen crag]

hey all!
i'd like to achieve this kind of flow - a bot which will react on a command (for example !play last-show) which will output a button to the user with a PLAY label on it, and once the user clicks - it will stream the audio from the latest mixcloud's show from my user.
Getting the last show from mixcloud's api is quite straightforward.. i wonder if there's an option to stream its' audio via discord's audio channel or somehow embed mixcloud's widget in the channel itself?

[vocal valley]

might have been included with your discord.js install 👍

[last mist]

thanks for that heads up... wasn't worried about having extra mostly about not having enough 🙂

[last mist]

a few hours later... no errors for the moment and no sounds ... sure is fun 😫

[vocal valley]

are you running the example?

[last mist]

if you mean the music-bot example then no.

[vocal valley]

ah ok

if you want, you can send the code you're using and i can take a look tomorrow to see what the issue might be

[last mist]

that's awfully nice of you... I'm resting and then I'll take another look or two. I've looked at that music-bot code but frankly as an example it is over engineered. Typescript, babel, jest, prettier, linter, some streaming stuff. That might be example two but not the hello world of audio.

I'm putting in more logging to try to determine where the issue may be. I don't see the connection.on events firing for instance.

[vocal valley]

the basic example might be easier to try and aid with debugging

[last mist]

I also search for messages here... one error informed me I didn't have the latest version of v13

[vocal valley]

ah yeah, that's come up a few times today

[last mist]

Yup 🙂

I wish I could do it but if the things people encounter could be set into a "you're probably going to do this" page it would go a long way.

I was also hoping that someone who distilled it down to "play a single .mp3 file" might post their experiences.

Ah I seem to have overlooked the basic example... what you posted now makes sense... I think I can paste that somewhere into my app without a lot of issues, thanks.

[vocal valley]

no problem! hopefully it works well for you

[last mist]

works well? Hopefully it "beeps" 🙂

thanks

[grim fulcrum]

This should work:

const { Client, Intents } = require('discord.js');
const { joinVoiceChannel, createAudioResource, createAudioPlayer, StreamType } = require('@discordjs/voice');

const client = new Client({ intents: [Intents.FLAGS.GUILDS, Intents.FLAGS.GUILD_MESSAGES, Intents.FLAGS.GUILD_VOICE_STATES] });

const player = createAudioPlayer();

client.on('message', async (message) => {
  if (message.content === '!play') {
    const channel = message.member.voice.channel;
    if (!channel) return message.reply('You must be in the voice channel.');
    const resource = createAudioResource('https://www.soundhelix.com/examples/mp3/SoundHelix-Song-1.mp3', { inputType: StreamType.Arbitrary, });

    const connection = joinVoiceChannel({
      channelId: channel.id,
      guildId: channel.guild.id,
      adapterCreator: channel.guild.voiceAdapterCreator,
    });

    player.play(resource);
    connection.subscribe(player);
    await message.reply('Playing now!');
  }
});
void client.login('token here');

[last mist]

now that is an audio "hello world"...

gracias

[last mist]

Yahoo that plays... much better when I type the command in my server, in my bot 🙂

Nice music

I can adapt it from here, thanks everyone

[grim fulcrum]

🎉

[last mist]

I've been having great success and migrated my little sound demo from v12 to v13 and added that SoundHelix song to the options. Meanwhile it seems that one can only set the volume if the inlineVolume option is set to true. I gather from the docs that this plays through the ffmpeg service which (I'm guessing) is the only one that supports volume, is that correct? The notes about improved performance concerns me a bit has this always been the case with ffmpeg? Volume control seems to be pretty much necessary doesn't it?

"There are only two kinds of programming languages: those people always bitch about and those nobody uses." -- Bjarne Stroustrup

[carmine timber]

@brazen crag You can use this example

[vocal valley]

Hi Tom. You can only set the volume if inline volume is enabled that is correct. It's not necessarily because of FFmpeg, but it basically requires us to decode the audio stream into raw audio within Node.js somewhere in the pipeline which isn't the best for performance (although not horrendous) - this is why the feature is opt-in 👍

[brazen crag]

Not sure, since mixcloud doesn't have mp3 endpoints which i can use directly

[carmine timber]

Give me a link of mixcloud

example of mixcloud endpoint

The audio streams are not available through the Mixcloud API. There are two reasons for this.

Firstly, we need to know what has been listened to so that we can report usage, pay royalties and provide features such as 'Suggested Shows'.Secondly, Mixcloud needs to pay the bills! We can't give away the audio for free outside of mixcloud.com simply because it costs us to host and stream the files and pay royalties.

@brazen crag Mixcloud never streams audio

So there is no possible way to stream audio through their api's

[brazen crag]

And what about their embeddable widgets which can be freely used on html pages? Is it somehow possible to embed them graphically on discord's interface?

[carmine timber]

interesting question, let me double check 👀

I can't find any way to use that embeddable widgets as Discord don't allow to attach HTML frames in their embeds

[brazen crag]

ok buddy! thanks for research and answers ❤️

[spare wave]

Looking at the example, feel like over engineered @.@

[carmine timber]

👀

[spare wave]

I meant the typescript example, well at least after toasting my brain, able to "downgrade" it to js and make it modular

[carmine timber]

I know, there is only typescripts examples only 🙂

[carmine timber]

But you could have used your brain here also : https://deploy-preview-595--discordjs-guide.netlify.app/voice/

[vocal valley]

There is also the basic example as a nicer introduction

[carmine timber]

@vocal valley Is there anything in player like ispaused() like in v12 voice ??

[spare wave]

is ok, manage to settle it

[vocal valley]

player.state.status === AudioPlayerStatus.Paused || player.state.status === AudioPlayerStatus.AutoPaused

[carmine timber]

Oh I forgot about that LOL

[spare wave]

I wont say it is 100% my code though, since I didn't understand the track and subscription, I shamelessly copied it and import it to my code

[vocal valley]

I don't think there is anything wrong with that,

[carmine timber]

subscription just creates a new Player and attaches one queue to voice connection.
While tracks collects data using ytdl-core like song.title, song.url etc.

[vocal valley]

The new voice library has a lot of low-level stuff and it makes sense to use something that works and handles edge cases for you

[carmine timber]

New Voice API is more flexible than old one. You can literally handle any event with connection and audio player separately.

Not place to ask this here , go to #archive-djs-v12-deprecated

[rich flicker]

oh, sorry

[spare wave]

For music bot, what is the recommended RAM size?

I don't plan to make it public though

[stark drum]

It will depend how many simultaneous players you'll have

[lime gale]

Error: Cannot perform IP discovery - socket closed
    at Socket.<anonymous> (D:\Coding\DisTube\DisTube\node_modules\@discordjs\voice\dist\networking\VoiceUDPSocket.js:120:52)
    at Object.onceWrapper (events.js:482:28)
    at Socket.emit (events.js:388:22)
    at socketCloseNT (dgram.js:748:8)
    at processTicksAndRejections (internal/process/task_queues.js:81:21)

My connection emitted this, should I ignore it since my connection is not disconnected nor destroyed? It's disconnected and reconnect the vc after that

[vocal valley]

yep you can ignore it 👍

[frosty dawn]

                secretbox.methods.close(opusPacket, connectionData.nonceBuffer, secretKey),
                                  ^

TypeError: secretbox.methods.close is not a function

any solution what's happening here?

[subtle granite]

Is this for music like rythm?

[lime gale]

Install sodium / libsodium-wrappers

[frosty dawn]

why djs voice needs encrypiton now?

it worked btw

[lime gale]

It's in the requirement of the new package

[carmine timber]

Actually @discordjs/voice has some issues with tweetnacl(default with djs), so please install libsodium-wrappers or sodium

[frosty dawn]

sodium wrappers worked

[carmine timber]

You can get a comparable sound quality and performance with this new voice api as compared to rythm and many more music bots

https://deploy-preview-595--discordjs-guide.netlify.app/voice/

@frosty dawn Here it is

[vocal valley]

it always has, it's just that discord.js included tweetnacl for other purposes, so the voice api in there could use that as a fallback

[subtle granite]

How can I use it to get the good sound quality?

[vocal valley]

as long as you provide it a good quality audio stream, it should be able to play it back reliably

it's much less choppy compared to discord.js v12's voice

[frosty dawn]

VoiceConnectionStatus.Ready !=="Ready"

will this work?

[vocal valley]

what are you trying to do

[carmine timber]

maybe if statement i guess

[vocal valley]

the string values for the VoiceConnectionStatus are here if this is what you were wondering, https://discordjs.github.io/voice/enums/voiceconnectionstatus.html

[frosty dawn]

check for ready state

[carmine timber]

yes

[vocal valley]

you should be doing something like:

voiceConnection.state.status === VoiceConnectionStatus.Ready

to check for the ready state

if you want to listen for it like an event,

const myConnection = createVoiceConnection(...);
await entersState(myConnection, VoiceConnectionStatus.Ready, 30_000);
// ready within 30 secs, otherwise it would have thrown

or

// one option
voiceConnection.on(VoiceConnectionStatus.Ready, () => console.log('Ready'));

// another option
voiceConnection.on('stateChange', (oldState, newState) => {
  if (oldState.status !== VoiceConnectionStatus.Ready && newState.status === VoiceConnectionStatus.Ready) {
    console.log('Ready');
  }
});
// both are equivalent

[lime gale]

If someone disconnects the bot, the reason of VoiceConnectionDisconnectReason should be EndpointRemoved or WebSocketClose?

[vocal valley]

hmmm i have not actually tested that

I think it would be WebSocketClose

[lime gale]

Hmm... I thought it must be EndpointRemoved. Thanks :v

[vocal valley]

np, can confirm it is WebSocketClose, with code 4014

However I made a mistake in the typings that makes it hard to extract the closeCode, will fix that later today

[spare snow]

How can I get all audio players?

[carmine timber]

in a connection or in all connections ??

[vocal valley]

you're expected to track them yourself

if you have all your voice connections, you can find your audio players that have at least one subscriber

[spare snow]

all of them

[spare snow]

How do I get all connections then? because getVoiceConnection asks for a guild ID

[carmine timber]

I think this might work :

var guild_array = client.guilds.cache.map(guild => guild.id)
guild_array.forEach((guild) => {
  var connection = getVoiceConnection(guild)
  if(connection) {
   //RUN some Code
}
})

[last mist]

So time to move onto YouTube videos... these used to play in v12 but I've tried any number of twists on the code that plays audio files (both from my drive and from a url) and while I see the video image displayed in Discord message the video does not play automatically. Do these operate similar to audio files?

[carmine timber]

Show me code where you played local file

[last mist]

it is spread out a tiny bit but let me see about gathering the parts. In a nutshell what I've tried was requesting a YouTube video and at the last moment playing a file on my PC. This worked which tells me all the "setup" seems to be present. I then pass the url of the video to the createAudioResource function and the video does not play.

' const resource = createAudioResource(soundFile, { inputType: StreamType.Arbitrary, inlineVolume: true });

audioPlayer.play(resource);

'

how much code do you need this is the resource and playing in the audioPlayer

[carmine timber]

SoundFile is a path ??

[carmine timber]

Only this much

[last mist]

soundFile is the url or the path depending

[carmine timber]

Can you show me where you declared SoundFile ?

[last mist]

if I set it to the URL of an audio file it plays, if I set it to a path on my drive it plays

soundFile is string property

[carmine timber]

So what is not playing ??

[last mist]

the YouTube video

maybe there is an "actually start it" option?

[carmine timber]

You need a module to play youtube Video

Download ytdl-core module

[last mist]

done already

and ffmpeg

[carmine timber]

Can you download ytdl-core-discord ??

[last mist]

this was running in v12 maybe there is an updated version of those modules?

bet you not

oh I thought you asked if I did

of course I can

[carmine timber]

So do it

[vocal valley]

by URL of the video, do you mean the URL of the videos page on YouTube?

[last mist]

Yes both the actual string and the url that came back from a call to ytdl.getInfo(args[1]);

[vocal valley]

Hmmm ok, I'm not sure if you can play that directly as a string

[carmine timber]

const ytdl = require('ytdl-core-discord')
const resource = createAudioResource(await ytdl(url), { inputType: StreamType.Opus, inlineVolume: true });

audioPlayer.play(resource);

Then try this

[last mist]

that will probably work, one moment

[vocal valley]

also as a rule of thumb now, I don't recommend ytdl-core or ytdl-core-discord for serious music bots because of how error-prone they are to streaming audio (a lot of the time the stream aborts)

if you're making a proof of concept they're fine, but otherwise I'd recommend youtube-dl (the music bot example uses this)

[carmine timber]

One thing, if we use ytdl-core-discord, do we need to specify stream type to OPUS or WebmOpus

??

[vocal valley]

Opus

[carmine timber]

Ok

[last mist]

Ray Parker Jr. is playing on my server!!!

Ghostbusters

[vocal valley]

🎉

[carmine timber]

🎉

[last mist]

I'll look at youtube-dl but this works for now

thank you, thank you

Uh youtube-dl looks like a library specifically for downloading files is that the case? Isn't that overkill and potentially a lot of overhead? I'll read more but it already looks like it runs on your server so I'd have the Windows version during testing and a Linux install in production. That seems like a lot of work. Thanks for the pointer though.

[vocal valley]

No problem, it can actually be used to stream the video files too rather than download

[last mist]

okay

[spare snow]

What happens when maxMissedFrames is reached?

and is it possible to set the bitrate?

[spare wave]

Anything I can do about this? Or that's that

I got face slapped by the error aborted few times

[spare wave]

Nvm, just saw the new example, and youtube-dl used, maight try later

[vocal valley]

the player stops

[vocal valley]

resource.encoder?.setBitrate(48000)

[carmine timber]

@vocal valley The ytdl-core-discord gives us a readable stream, then why we need to pass it through CreateAudioResource ?? Can we pass that directly to player.play ??

[vocal valley]

Nope, a player only plays audio resources. Creating a resource from an Opus stream doesn't pass the stream anywhere else though, it just wraps the stream as it is, so performance isn't made any worse

[carmine timber]

I did a research, please let me know your thoughts : https://github.com/discordjs/voice/discussions/123

[vocal valley]

i will take a look later today

although i am fairly sure if it is not being destroyed, then it is due to ytdl-core

🔔 v0.4.0 has been released https://github.com/discordjs/voice/releases/tag/v0.4.0 🔔

[carmine timber]

I also agree to this

[mental patio]

in docs i saw <VoiceConnection>.subscribe() is let the AudioPlayer can play the stream. but in the djs/voice's "Hello World" i saw it did player.play() first then do connection.subscribe(). Why?

[vocal valley]

to be honest, it does not matter too much about which way around you do it

doing it a little earlier than .subscribe() means that your audio stream starts to be downloaded faster, so there's less time to wait before actually playing

if you do it too far in advance, depending on where the stream is coming from, it might have disconnected or exited

[mental patio]

so in that case it is acceptable

[vocal valley]

yep, that particularly audio source is fine to buffer in advance

and also we're assuming the user will run the example quickly after starting the bot

[mental patio]

is there any event emitted when the resource is finish playing?

[vocal valley]

the player transitions to Idle when it finishes playing

[carmine timber]

use this :

player.once(AudioPlayerStatus.Idle, () =>{
     //Player Ended
    })

[vocal valley]

player.on('stateChange', (oldState, newState) => {
  if (oldState.status !== AudioPlayerStatus.Idle && newState.status === AudioPlayerStatus.Idle) {
    console.log(`Finished playing ${oldState.resource.metadata.title}`);
  }
});

(since Idle is the only state of an audio player that doesn't have a resource, the above works nicely)

[mental patio]

oof okay i think i got it

so both this code works

[carmine timber]

🎉

Yes

[mental patio]

oof nice thx for helping

[carmine timber]

np

[carmine timber]

I have tested my 2nd code on a loop of 15 minutes and I didn't face any memory leak issues 🙂

And the music that I play now (2nd code) is soo good and without any disconnects (Tested on local PC)

[mental patio]

wait so the creatAudioResource() can have a youtube link as parameter

[carmine timber]

No

One thing, I want to ask , I can get yt-stream link in the format

container: 'webm',
codecs: 'opus',

So Can I use StreamType.WebmOpus while creating audio source in this case ??

[vocal valley]

yep, you should be able to in this case

wait

[carmine timber]

It is a link and not a file

[vocal valley]

if you are passing a link, it will get passed to ffmpeg and turn into an Ogg opus file automatically

[carmine timber]

Thanks

[vocal valley]

Np, it will be a little bit worse performance-wise than compared to using youtube-dl

Since it converts the stream for no reason

[carmine timber]

I got a very good performance with it with no disconnect issue

[vocal valley]

If it works then it's good I guess, even in that case the performance penalty shouldn't be too much since it's just changing the container of the file for you

[mental patio]

oh i see that we can got the audio file link we ytdl

[carmine timber]

Yes 🙂

[carmine timber]

But in my case, I think I got more performance with StreamType.Opus than StreamType.WebmOpus. With webmopus, stream lags a little but with opus, there was no lag at all

[vocal valley]

If you are using a link, the stream type is ignored

[carmine timber]

LOL

[vocal valley]

since create resource function will automatically use the correct format to understand what ffmpeg gives back 😅

[carmine timber]

Have you updated this on npm ??

[vocal valley]

It should be released, yep

[carmine timber]

npm i @discordjs/voice is installing 0.3.1

Not 0.4.0

But I managed to install with npm i @discordjs/voice@latest

[hazy copper]

have all these random crashes been taken care of? i just updated and it looks like im not getting them anymore?

[carmine timber]

What type of crashes ??

[hazy copper]

like the thing would stop working on ytdl

nvm, it still gets them

[carmine timber]

ytdl is broken, so don't ask for help here

[vocal valley]

yea this is a ytdl-core issue, not a /voice issue

The music bot example has been updated to use youtube-dl instead, which doesn't have the "random crashes"

[spare wave]

Any idea or direction if I wanted to do repeat function/command

[vocal valley]

to replay a song?

[spare wave]

yeah

[vocal valley]

make a new audio resource from the same song, and then play that again

[spare wave]

roger, will look into the documentation, btw any plan making the documentation more friendly like discordjs?

[vocal valley]

at the moment i can't do much about the API documentation (we are planning on upgrading our doc generator to work with typescript), but I am planning on writing more overview documentation, e.g. like https://deploy-preview-595--discordjs-guide.netlify.app/voice/ and adding more examples

[spare wave]

https://github.com/Ayeven/discord-musicbot I just published the initial commit yesterday, not sure can contribute as example or not

[spare snow]

Hey again @vocal valley, I just found out that, for some reason, the memory leak doesn't happen on windows but does on linux. I have tested it multiple times with same code and nodejs version on 2 windows machines and 2 linux machines and I wish if anyone can reproduce this

[vocal valley]

is this when using ytdl-core?

[spare snow]

yep

[vocal valley]

could you try the latest example that uses youtube-dl instead?

[spare snow]

I tried it on linux and it was leaking as well, but slowly because I didn't pass a highWaterMark

[vocal valley]

🤔 thats interesting

[spare snow]

on windows I can literally see the stream getting garbage collected

[vocal valley]

are you running the example as it is? or did you make changes

[spare snow]

as it is

[vocal valley]

and how do you know that it is leaking exactly? does the youtube-dl process stay alive after the song has finished playing?

[spare snow]

I was monitoring process.memoryUsage, on windows, when I skip a song it goes down to the idle usage but on linux it just keeps piling up

[vocal valley]

ok after you skip a song, if you wait a minute or two and then check memory usage again, does it go down?

e.g. for me, i have ~60mb to start,

playing a song takes it to about 70mb

skipping takes it to about 80mb but then it drops back down to 70mb

and then when it finishes, after a few seconds it goes back to ~60mb

[steel crescent]

Damn you just took the example, compiled it to JS, and slapped it into a repo :D
I'm not sure what the difference here is, compared to just looking into the example folder. You didn't even bother double checking removing the safety checks the typescript compiler adds.

[spare wave]

Yeah, the subscription and track is from example

[spare snow]

let me try this, I was trying with a lot of songs queued up so when I skip it plays the next one but I instantly check the memory and I could see it resetting to 60-70. I will try finishing the queue now

[vocal valley]

ok 👍

sometimes it just takes time for GC to kick in

[spare snow]

It never does

[vocal valley]

if after skipping/stopping and some time it drops back to 60/70, it means that there isnt a memory leak

could you try this:

$ node --expose-gc yourbot.js

and then whenever a song finishes, can you run global.gc()

if this reduces your memory usage, it means it's not a memory leak

[steel crescent]

a memory leak means you constantly gain

its as simple as that, stop checking your memory all the time lmao

if you go from 50 to 100 to 80 to 120 to 50 to 80 to 100

thats not a memory leak

You also need to test a considerably longer time for this to establish to be a memory leak

just so we are on the same page for example

this isnt a leak

[spare snow]

That IS what I'm saying, on linux it never does down, only gaining

[steel crescent]

Over which period of time?

And yeah thats fine, I just want us to be on the same page. because people claim something to be a memory leak rather fast around here

and its increasingly annoying to deal with that when they instantly jump to conclusions

[spare snow]

I'm just saying that is behaves differently on windows, when I skip a stream it instantly gets gc and never does on linux

[radiant ocean]

how long is never

[spare snow]

days

[steel crescent]

is this on a rather big public bot

or some test bot on 1 guild

[spare snow]

it doesnt reset when I stop the player, only when I override the stream

[spare snow]

both

[vocal valley]

what do you mean by override the stream

[carmine timber]

I think, my code will fix his errors 🤔

[vocal valley]

maybe but he says he is also getting this issue with youtube-dl

[spare snow]

this, aka skipping

[carmine timber]

As I ran that code over 15 minutes and memory was still at 60

[vocal valley]

ok, so when you do this, memory goes back down? but not when you call .stop()?

[spare snow]

yup, It only frees when another stream replaces it

[vocal valley]

so if a song naturally comes to an end, and then you play another song after, the memory from the first song isnt freed?

but if you enqueue a song while one is playing, and then skip, it does get freed?

[spare snow]

it is freed

I thinks it's a @discordjs/opus issue but not sure

[vocal valley]

🤔

[steel crescent]

You aren't making sense

[vocal valley]

so under which circumstance is it not freed

[spare snow]

It's never freed unless I play another song, which resets the memory and goes back up at a certain limit (120mb for me)

[vocal valley]

that doesn't sound like a memory leak then

since it does go back down

[carmine timber]

Yup

That's not a memory leak

[spare snow]

yeah that's what I am saying! on linux it's piling up

just try to reproduce this

[vocal valley]

so on windows, this happens

[carmine timber]

So you mean ytdl has no garbage remover right .

[vocal valley]

and on linux it keeps going higher and higher and doesnt come down at all?

[spare snow]

exactly

[vocal valley]

ok, you're using the youtube-dl example right?

[spare snow]

ytdl-core-discord

[carmine timber]

Oh LOL

XD

[vocal valley]

ok,

one sec let me do something

[spare snow]

with highWaterMark: 1 << 25 to make the leak clearer

[carmine timber]

We were thinking about youtube-dl whole time and he is using ytdl-core-discord 😂

[steel crescent]

we just need to stop recommending this lib

[vocal valley]

using highWaterMark is not a good solution

ok so

[spare snow]

without it the stream aborts early and the player stops

[vocal valley]

can you go to node_modules/ytdl-core/lib/index.js, and around line 180, change stream._destroy = () => ... to this:

  stream._destroy = () => {
    console.log('destroying stream');
    stream.destroyed = true;
    req.destroy();
    req.end();
  };

[spare wave]

using the youtube-dl it is within expected memory usage, on raspi pi 4 8GB

[vocal valley]

and then try running your bot again, when you skip or whatever on linux, it should say "destroying stream"

[vocal valley]

it works but only because it essentially downloads up to 33MB of the video

so it's not really streaming, more of "download this entire video into memory as quick as you can and then stream that"

the problem is when you have longer videos, e.g. that are more than 33MB

[spare snow]

it's same thing 🙄 the issue is likely with opus

[carmine timber]

hydrabolt, did you checked my discussion yet ??

[vocal valley]

highWaterMark is more of a bandaid fix, rather than a fix of something wrong in ytdl-core's reconnect logic i assume

[vocal valley]

not yet

[vocal valley]

i am sure the opus library does not have mem leaks lol

[spare snow]

what if I stream 10 hours long videos

[vocal valley]

can you try this first

that will help us narrow down where the error is

[vocal valley]

thats when the highWaterMark fix will probably no longer work

[carmine timber]

ytdl-core never ends stream when we use skip command, it still plays that stream in background

[vocal valley]

i just want to be sure that is what is happening,

[carmine timber]

ok

[steel crescent]

how are you so confident in this

have you tried with opusccript

the opus library doesnt even allocate memory, so how would it keep something in memory

lol

[carmine timber]

lol

[spare snow]

IDK, just the 2 different behaviors made me think that, since it needs to be compiled or something

[steel crescent]

so you dont know

just say that instead of pointing the finger somewhere else

you should try youtube-dl or opusscript

[carmine timber]

He could even manage that without youtube-dl and opusscript also

[vocal valley]

but please also check this first @spare snow 😩

[spare snow]

I'm on it

@vocal valley good news, on windows it does free the memory when I stop the player, just needed some time. (I just found out about the windows thing don't roast me)
but still, on linux it never does

[vocal valley]

ok but on linux, does it print out the destroyed thing

[spare snow]

yeah it does

I just stopped the player and I'll wait to see if it gets gc, it's sitting at 150mb already

I hate my ocd, I just could ignore it..

[vocal valley]

can you also try this

[spare snow]

nothing got collected

[vocal valley]

how high can you get the memory to go on linux, could you test that

[spare wave]

about this memory thing, is this only affect ytdl-core? Because using youtube-dl, an it behave like it supposed to be

memory().heapUsed btw

[vocal valley]

i believe so? but even then i'm not sure if i can reproduce it with ytdl-core myself

[spare wave]

My short test earlier. The first song is quite long, 32min

[wary forge]

Hello, sorry for the dumb question, if I'm using lavalink, do I still need to install this package or its dependencies?

[carmine timber]

Lavalink and this voice api are different

If you have lavalink, then don't install this

[wary forge]

the lavalink takes care of it on its own right

[carmine timber]

yes

[wary forge]

i see thanks

[carmine timber]

using youtube-dl or ytdl-core ??

[spare wave]

youtube-dl

[spare snow]

it keeps adding up with each new player

[carmine timber]

I think, I had a fix with ytdl-core itself, but I don't know whether hydrabolt will accept that or not

[spare snow]

260mb for 3 players and so on..

[spare wave]

ytdl-core always faceslap me with econReset or something those line

[carmine timber]

So it is due to youtube-dl memory only I guess

[carmine timber]

I have a better fix, no need to use code like this ytdl(url)

wanna try ??

[spare wave]

sure, though it will be later or tmrw

half gaming, hafl discord atm

[carmine timber]

Ok wait

function filter(format) {
    var results = []
    format.forEach((ele) => {
        if(ele.codecs === 'opus' && ele.container === 'webm' && ele.audioSampleRate === '48000'){
            results.push(ele)
        }
    })
    return results
}

//Below Code : Under message event
    let connection = await join_vc(message);
    const info = await ytdl.getInfo(<your - url>)
    let stream = filter(info.formats)[0].url
    var source = createAudioResource(stream, { inputType: StreamType.Opus , inlineVolume : true })
    var player = createAudioPlayer()
    player.play(source)
    connection.subscribe(player)

@spare wave

[spare snow]

also each player adds more memory with each new song btw

[carmine timber]

Can you try above code ??

and then tell me if memory still goes up or not

[vocal valley]

What node.js version are you running

[carmine timber]

@vocal valley 1 more issue with youtube-dl

[spare snow]

14.17.0 and 15.6.0, same results

[vocal valley]

This seems fine to me

[spare snow]

I'll give it a shot

[vocal valley]

Yep, either try Youtube-dl example or the example above

[spare wave]

well, yeah, it climb up because the song is long

[vocal valley]

and then, once a song ends or you've skipped a few songs

Can you see how many ffmpeg (for the example @carmine timber posted) processes are running, or how many youtube-dl (for example on repo) processes stay running

[spare snow]

I already tried this, no youtube-dl processes stay running after the song is skipped/end, and it didn't spawn any ffmpeg cuz i'm using opus type

[vocal valley]

hm that is good, it would suggest there isn't actually a memory leak

[carmine timber]

So this means that youtube-dl requires high usage of memory ??

[vocal valley]

No

If the process was alive after the song finished, that would suggest there's a memory leak

[spare snow]

Well, Windows says otherwise.
But if it's working as expected I'll just move on

[steel crescent]

just dont use ytdl-core

[vocal valley]

just see if your process runs out of memory after long usage

if it does, then you probably do have a memory leak

It also could just be d.js cache build up

a lot of factors here

[carmine timber]

Are you irritated with ytdl-core ??

[stark drum]

Windows and Linux don't manage memory the same way
You will always encounter differences between the two systems (even for how node allocates/free memory itself)

As long as it does not grab more and more memory over time it's fine
Make your bot in C if memory usage is a critical point

[vocal valley]

Idk about crawl but I am for sure

the developer of it has tried their best to fix it for discord.js so props to them

[steel crescent]

its time to start writing voicelink

[vocal valley]

But it still doesn't work great with d.js and people attribute their ytdl errors to discord.js

[carmine timber]

But is there any fault, the way I am streaming content ??

[vocal valley]

I think it's ok to use FFmpeg? but honestly I just trust youtube-dl to download the video cause that's why it exists

[carmine timber]

Oh

[vocal valley]

maybe there is some reconnection logic youtube-dl has that ffmpeg doesn't

[carmine timber]

But you can add reconnect args in ffmpeg also

[grim fulcrum]

I don't understand what it is, but there is no sound.

[carmine timber]

New Voice API or old ??

[grim fulcrum]

New

[carmine timber]

Code plz

[grim fulcrum]

Example

[carmine timber]

Maybe it is something related to your internet connection problems

[stark drum]

Likely to be binary data, so not designed to be printed by definition which explains those characters
Show code

[carmine timber]

It just failed to connect to your voice channel

[grim fulcrum]

No, it's connected to a voice channel.

[stark drum]

Why would it print what it seems to be audio data then, also you'd get an error from discord.js at some point too

[vocal valley]

Can you send the full output of the console to us

[carmine timber]

@vocal valley Can You add some ffmpeg args in this : https://github.com/discordjs/voice/blob/main/src/audio/TransformerGraph.ts

-reconnect 1 -reconnect_streamed 1 -reconnect_delay_max 5

It would be great

These are reconnect args which allow ffmpeg to reconnect once disconnected by itself

Just tell me if you will add or not

[vocal valley]

I think I will

[carmine timber]

Thanks

[grim fulcrum]

That's all I could find.

.......��6����ވ;^E_���yؙ_ǝF���>    �5�Ǵ����n�:��r�7⟴�չt�#�F��������L,�J))��=�@��Fq
                                       ��������1�G�?3�����j��Db�WW2#l�-pRN�̬{%'c�ue�<�1`Ƿd��xy�����ű�LV/y��-�V������h
                                    cH�+ä����@�������T�������h������u��'ġ��|���
    at makeError (/Users/penguingl/Documents/example/node_modules/execa/lib/error.js:60:11)
    at handlePromise (/Users/penguingl/Documents/example/node_modules/execa/index.js:118:26)
    at runMicrotasks (<anonymous>)
    at processTicksAndRejections (internal/process/task_queues.js:93:5) {
  shortMessage: 'Command failed with exit code 1: /Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl https://www.youtube.com/watch?v=VBKNoLcj8jA -o - -q true -f bestaudio[ext=webm+acodec=opus+asr=48000]/bestaudio -r 100K',
  command: '/Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl https://www.youtube.com/watch?v=VBKNoLcj8jA -o - -q true -f bestaudio[ext=webm+acodec=opus+asr=48000]/bestaudio -r 100K',
  escapedCommand: '"/Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl" "https://www.youtube.com/watch?v=VBKNoLcj8jA" -o - -q true -f "bestaudio[ext=webm+acodec=opus+asr=48000]/bestaudio" -r 100K',
  exitCode: 1,
  signal: undefined,
  signalDescription: undefined,
  stdout: `\x1AEߣ�B��\x01B��\x01B�\x04B�\bB��webmB��\x04B��.......
  
      tts: false,
      nonce: undefined,
      embeds: undefined,
      components: undefined,
      username: undefined,
      avatar_url: undefined,
      allowed_mentions: undefined,
      flags: 0,
      message_reference: [Object],
      attachments: undefined
    },
    files: []
  }
}

[vocal valley]

huh

could you try running your bot and piping it's output to a file?

node bot.js > debuglog.txt

and then once the error occurs, send the debuglog.txt? annoyingly the top of the error is cut off

but doing this allows us to capture the whole thing

[vocal valley]

that is so odd 🤔

ill take a look at the youtube dl npm wrapper we use maybe there is an issue there

what node.js version are you using?

and are you on windows?

@grim fulcrum what do you get if you run this in a terminal:

/Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl --version

[grim fulcrum]

node 16.3.0
mac

[vocal valley]

ok, can you try this:

[grim fulcrum]

2021.06.06

[vocal valley]

/Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl https://www.youtube.com/watch?v=VBKNoLcj8jA -o - -q true -f bestaudio[ext=webm+acodec=opus+asr=48000]/bestaudio -r 100K

what happens if you run that?

[grim fulcrum]

...����(̳�i����U�`.6�3��V�
ERROR: u'true' is not a valid URL. Set --default-search "ytsearch" (or run  youtube-dl "ytsearch:true" ) to search YouTube
�       ճ��[���Th��v��...

[vocal valley]

ok, what about this:

/Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl https://www.youtube.com/watch?v=VBKNoLcj8jA -o - -q true -f "bestaudio[ext=webm+acodec=opus+asr=48000]/bestaudio" -r 100K

it might still give you nonsense in the console

but if there's no error it's good

[grim fulcrum]

same error

ERROR: u'true' is not a valid URL. Set --default-search "ytsearch" (or run youtube-dl "ytsearch:true" ) to search YouTube

[vocal valley]

🤔

oh

/Users/penguingl/Documents/example/node_modules/youtube-dl-exec/bin/youtube-dl "https://www.youtube.com/watch?v=VBKNoLcj8jA" -o - -q true -f "bestaudio[ext=webm+acodec=opus+asr=48000]/bestaudio" -r 100K

now?

[steel crescent]

what are you passing true here for?

isnt -q enough to activate silent mode

[vocal valley]

ohh shit

[grim fulcrum]

ERROR: u'true' is not a valid URL. Set --default-search "ytsearch" (or run  youtube-dl "ytsearch:true" ) to search YouTube

[vocal valley]

thats a youtube-dl-exec bug crawl is right

on linux it doesnt pass the true 🤔

[steel crescent]

i mean it might not be wrong

I just thought its unneeded

[vocal valley]

you are right i think

[carmine timber]

But he is on windows right ??

[vocal valley]

on linux, running youtube-dl-exec with options {q: true} just yields -q iirc

on windows it must be doing -q true

[grim fulcrum]

Mac

[carmine timber]

Oh

[steel crescent]

hmm

[vocal valley]

@grim fulcrum try this:

change line 52 in examples/music-bot/src/music/track.ts:

- q: true,
+ q: '',

[steel crescent]

yeah I tried it on debian

removing the true seems to work

[vocal valley]

thats so odd

it converts the args object into a list in JS, i thought that'd be platform-independent

[grim fulcrum]

Works

[steel crescent]

might just be a quirk in the youtube-dl args parser

not expecting a value for -q so it poops itself

[spare wave]

on ubuntu, q:true all working good

[vocal valley]

🤨

guess i'll go with empty string

works in both

[steel crescent]

I only pasted the command from above, didn't try the actual bot

so theres that

[spare wave]

actual bot here

[spare wave]

what q parameter do though? have not read the youtube-dl docs

[vocal valley]

means "quiet", stops youtube-dl printing unnecessary info

[spare wave]

changing q:'' work too for ubuntu

[vocal valley]

@carmine timber i decided against adding reconnect options to ffmpeg at least for now since they might not be the optimal values for each user

[carmine timber]

discord py Voice API has those args in ffmpeg. They will never affect and stream quality or playback. They are used just for reconnecting stream and nothing else

[viscid sable]

should i only have discordjs/voice or do i also need to install discord.js for it?

[vocal valley]

i understand that, but even if we look at https://github.com/discordjs/voice/issues/117, it seems that reconnect_delay_max=5 was not enough for this user here

[viscid sable]

(currently getting the error that the module dosen't exists)

[vocal valley]

i can't plan for every user's setup, instead they should do this themselves

[vocal valley]

the /voice library doesn't have to be used with discord.js, but in all our examples, we use discord.js

[viscid sable]

i somehow always get the error Error: Cannot find module '/home/container/node_modules/@discordjs/voice/dist/index.js'. when i try to start it

[vocal valley]

did you run npm run build?

wait

did you install the library from npm?

[viscid sable]

yes

[vocal valley]

can you do ls /home/container/node_modules/@discordjs/voice and send the output here?

[viscid sable]

only this is in the folder

[steel crescent]

did u run npm build before the publish hydra

[vocal valley]

🤔

i did

i am sure i did

[viscid sable]

on my local pc i have the folder node_modules but nothing more

[vocal valley]

when i run npm i @discordjs/voice i get the dist folder

so im not really sure what's going on here

[viscid sable]

this is only what i get when i run it in the folder

and npm says it installed it without any errors

[carmine timber]

Just increasing the delay to 300 and at least add in the files.

[vocal valley]

but that is a subjective change

[carmine timber]

300 is sweet spot for most users

[vocal valley]

some people might just want the resource to die after a few seconds of delay

not all users will think 5 minutes is acceptable

[carmine timber]

Can you add something like max_reconnect_delay option in createAudioResource ??

Default to be 5

[steel crescent]

just do it yourself lol

[carmine timber]

Then that's ok for all users

[vocal valley]

i'll consider it but that is the same reason i removed seek from createAudioResource

because it only applies if you use ffmpeg

createAudioResource is not guaranteed to use ffmpeg

[carmine timber]

Oh

[steel crescent]

id hate to have an option forced onto me

especially because I never had any voice connection problems ever lol

[vocal valley]

the ffmpeg option we have so far does the bare basics of just turning the stream into an opus stream, that's all

[steel crescent]

so it would be totally useless for me

[viscid sable]

after doing npm remove discordjs/voice and redownloading it i now got the dist folder but somehow not on the host machine

[steel crescent]

bruh

[vocal valley]

sounds like a cursed setup

[steel crescent]

that sounds like you install from github

[vocal valley]

oh yeah

[steel crescent]

npm i @discordjs/voice

the @ is important

[viscid sable]

its what i did

[carmine timber]

@vocal valley Did you check out my research ??

[steel crescent]

lol

[carmine timber]

XD

[viscid sable]

i removed it and redownloaded it with npm

[carmine timber]

Try typing this npm i @discordjs/voice@latest

[viscid sable]

now i got a dist folder on the host machine

[carmine timber]

Try typing npm config list

and show me result

[viscid sable]

i sadly can't run commands on the host machine, its just allows me to connect with FTP and install npm packages

(as its a docker container runned by Pterodactyl)

[carmine timber]

So where do you want to install voice module ??

[viscid sable]

i was trying to install it on the host machine but only gave me the error in the reply

[carmine timber]

Do you have npm in windows ??

[viscid sable]

it seems like this fixed it for me as it dosen't say that it cant find the module anymore

[carmine timber]

🎉

[viscid sable]

now i only need to rework some stuff because the example is in ts and im using js

[carmine timber]

Can you create something like extra_options option in CreateAudioResource ? If ffmpeg is used then, extra args will be used otherwise no use of that

ExtraOptions will be just an extra argument that users want to supplement in ffmpeg.

Just like you supplemented extra args while not touching basic args for ffmpeg : https://github.com/discordjs/voice/issues/117#issuecomment-860081673

[chrome sundial]

can i just tsignore

or it will make me some problems in future

[vocal valley]

can you update your @discordjs/version

should be fixed on 0.4.0

[vocal valley]

but that leads to inconsistency

e.g. people will then think the stream they pass will have reconnect set for them

if people want to use custom FFmpeg arguments, they should just spawn their own FFmpeg instance, it's not too hard

and at least then they are being explicit about using FFmpeg

[chrome sundial]

thanks i updated it

[bold cairn]

Where’s the documentation for “new-voice”?

[rustic temple]

https://discordjs.github.io/voice/

[mental adder]

Anyone can help me?

[rustic temple]

Are you sure it'll be executed in a guild?

And not a DM

[mental adder]

The code

[rustic temple]

Mmh you should make them required, not optional I guess

[carmine timber]

Change this to something like :

joinVoiceChannel({
        channelId: voiceChannel.id,
        guildId: guild.id,
        adapterCreator : guild.voiceAdapterCreator,
        selfDeaf : true
})

[rustic temple]

They're using typescript, it's often unhappy when params are optional and you try to use them where they're not optional

[proud spindle]

will joinVoiceChannel also support stage channels?

[carmine timber]

yes

[vocal valley]

It already does @proud spindle

[proud spindle]

probably just me needing an update, thanks!

oh wait, I meant the createDiscordJSAdapter in the example

[vocal valley]

Ohh

The type needs updating to include stagechannel, but besides that it will work

[proud spindle]

Great, thanks!

[covert sky]

hey

please client for voice code

[proud spindle]

if you mean examples there are some at https://github.com/discordjs/voice/tree/main/examples

[covert sky]

no no i mean client like this'

const Discord = require('discord.js');
const client = new Discord.Client();

but for voice

[rustic temple]

I don't think client changes

[proud spindle]

No, you use your existing client from djs with the new voice package. Apart from some methods, nothing really requires changes

[covert sky]

am trying to join my bot an voice channel

[grim fulcrum]

https://deploy-preview-595--discordjs-guide.netlify.app/voice/voice-connections.html

[covert sky]

thx broo

thats the client true?

[grim fulcrum]

This library does not have a client.

[covert sky]

[hallow pumice]

Any good tutorials or sources on how to use this thing? I don't fkin understand it... My smol brain

[icy maple]

There’s the guide linked above and examples linked in the pin

[hallow pumice]

examples are in ts... And the guide just straight up sucks, or I'm stupid

Probably the second one

[icy maple]

Just ignore the types and it becomes js

[hallow pumice]

When you ignore the types it becomes just script... Ok ok jk, I'll try looking at the examples

[hallow pumice]

const { NoSubscriberBehavior, StreamType, createAudioPlayer, createAudioResource, entersState, AudioPlayerStatus, VoiceConnectionStatus, joinVoiceChannel } = require('@discordjs/voice'); When using this, this doesn't work const player = createAudioPlayer(); It says it's not a function and I fully understand that it's not. I think the problem is that in the example it's import, because of ts, and here it's different. How to change it?

[wind echo]

Is there any way to invite bot to voice channel on v13? Previous it works like message.member.channel.join() but on v13 it doesn't. I'm not ready yet to use new voice library.

[steel crescent]

Nope

V13 already removed all that code

[wind echo]

oh damn

okey thanks to answer

[grim fulcrum]

Do you have @discordjs/opus and libsodium-wrappers installed?

[covert sky]

const Discord = require('discord.js');
const client = new Discord.Client();
(message.content.startsWith('!join'))
voiceChannel.join()
  .then(connection => console.log('Connected!'))
  .catch(console.error);```

no respone

whats the problem?

[upper storm]

Voicechannel ist Not defined

[covert sky]

how??

[upper storm]

const Discord = require('discord.js');
const client = new Discord.Client();

if (message.content.startsWith('!join')) {
let voiceChannel = client.channels.cache.get("VC ID");

voiceChannel.join()
  .then(connection => console.log('Connected!'))
  .catch(console.error);
}

Also you forgot an if

[covert sky]

wow

[upper storm]

And

[covert sky]

then?

[summer needle]

So, I tried looking at the docs about AudioResource, but I am still confused.

const resource = voice.createAudioResource(play);

play is a ReadableStream, but It throws this error:
"stream.on is not a function"

[upper storm]

You forgot the client.on("message“) @covert sky

[covert sky]

what if client.on("message“) not found@upper storm

[upper storm]

Uff

[covert sky]

[upper storm]

const Discord = require('discord.js');
const client = new Discord.Client();

client.on("message", async message => {
if (message.content.startsWith('!join')) {
let voiceChannel = client.channels.cache.get("VC ID");

voiceChannel.join()
  .then(connection => console.log('Connected!'))
  .catch(console.error);
   }
});

Her you go

Read the docs bro

[steel crescent]

This doesn't work in v13 anymore.

🤔

[upper storm]

V13 is Released?

[steel crescent]

no but people in here are using it

its the "new" voice channel

not #archive-djs-v12-voice-deprecated

[upper storm]

Yeah Ik say this to him 😂

[summer needle]

#archive-voice message

[grim fulcrum]

Where do you get play from?

[covert sky]

no respone again

[covert sky]

okay

[summer needle]

It is a ReadableStream from a buffer

[grim fulcrum]

@discordjs/opus and libsodium-wrappers installed?

[summer needle]

@grim fulcrum yes

Yeah, i dont know. This is weird

[carmine timber]

Can you show me what is play defined ??

[carmine timber]

Your readable Stream doesn't have events support, so it is throwing this error

[summer needle]

Oh I'm just using Readable.from(buffer), how would I give it events support?

[carmine timber]

What exact code are you using ??

show me

[summer needle]

In not on my PC right now, I'll show you later.

[carmine timber]

To create a proper readable stream, you need a code like fs.createReadStream('<your file>')

[carmine timber]

Ok np

[summer needle]

It has to be a buffer though, not a file

[carmine timber]

You can use buffer also

I just gave an example

[carmine timber]

Refer to this : https://nodejs.org/api/fs.html#fs_fs_createreadstream_path_options

[carmine timber]

Can you show me node_modules/@discordjs/voice/ folder ??

[mental adder]

Can someone help me? I uploaded my bot to my host, but it doesn't play there, on my computer it does

[carmine timber]

Any error on host ??

[mental adder]

Nop

[carmine timber]

Can you run this on your host ??

const { generateDependencyReport } = require('@discordjs/voice');

console.log(generateDependencyReport());

[mental adder]

It looks like this on the voice channel, but no audio

[carmine timber]

Ok

[carmine timber]

Run this

[mental adder]

Wait a minute, I'm using the docker, I'll publish it real quick here and I'll send it

[carmine timber]

Ok

[mental adder]

wtf, My dependencies updated and now it's giving an error here

[chrome sundial]

What d.js version are you in

[mental adder]

^13.0.0-dev.1ac9a2eb5bfef6d009de7c54d3fac5e3d4a5afdc

[carmine timber]

Type this on host, npm i discord.js@dev@latest

[mental adder]

But now is on my pc the error

[carmine timber]

Type this in pc also

[mental adder]

ok

wait

@carmine timber same error

[carmine timber]

Can you show me full code of join vc ??

[mental adder]

async join (): Promise<VoiceConnection> {
    const connection = await joinVoiceChannel({
      channelId: this.voiceChannel.id,
      guildId: this.guild.id,
      selfDeaf: true,
      adapterCreator: this.guild.voiceAdapterCreator
    })

    await entersState(connection, VoiceConnectionStatus.Ready, 30e3)

    return connection
  }```

[carmine timber]

Does code run ??

[mental adder]

I can't even compile or anything, forever on this error

[carmine timber]

Can you console.log(this.guild) ??

[chrome sundial]

That's weird can you try to use npm i in latest discord.js commit

[carmine timber]

He tried it

[carmine timber]

read above chats

then comment

[mental adder]

@carmine timber

[carmine timber]

is there anything like voiceAdaptorCreator

??

[mental adder]

On the guild obj? Nop

[carmine timber]

Can you type npm list @discordjs/voice

??

[mental adder]

same on npm

[carmine timber]

Ok

[carmine timber]

Can you reinstall discord.js dev ??

[mental adder]

ok, wait

I'll clear the yarn cache first

[carmine timber]

Ok

[carmine timber]

Try reinstalling it with npm

[mental adder]

ok

Okay, maybe I'm unlucky, add an error and double up

[carmine timber]

Try restarting VS

[mental adder]

Ok

[carmine timber]

Fixed ??

[mental adder]

Yes

now i'm going to test on my host

[carmine timber]

😄

[mental adder]

with this

[carmine timber]

Yes

It will tell me about your dependencies that you have on your host

[mental adder]

put it in the bot's ready event, is there a problem?

[carmine timber]

Go ahead

np

[carmine timber]

Your audioplayer is having issues while sending packets to voice channel

[wise anchor]

and how do i fix?

[carmine timber]

Host somewhere else

[wise anchor]

im not hosting it anywhere

only on my pc

[carmine timber]

I know

This is your internet issue

[wise anchor]

no? i have really good internet

[carmine timber]

Then try restarting the bot

[wise anchor]

ight i restarted the bot

now he joins

then leaves

no error

[carmine timber]

Joining code ??

[carmine timber]

In joinVoiceChannel, add debug : true
and then after that add a line queue.connection.on('debug', (data) => {console.log(debug)})

[wise anchor]

okay

[mental adder]

@carmine timber ```

Core Dependencies

• @discordjs/voice: 0.4.0

• prism-media: 1.3.1

Opus Libraries

• @discordjs/opus: 0.5.3

• opusscript: not found

Encryption Libraries

• sodium: not found

• libsodium-wrappers: 0.7.9

• tweetnacl: not found

FFmpeg

• version: 4.3.2-static https://johnvansickle.com/ffmpeg/

• libopus: yes

--------------------------------------------------```

[carmine timber]

You don't have tweetnacl. That's interesting 👀 , since djs uses tweetnacl by default

[mental adder]

Does it have something to do with my Dockerfile?

[carmine timber]

NO

[mental adder]

I'm using node:14-alpine

[carmine timber]

Try to play song again

[mental adder]

Same thing, it enters and stays like that, but no audio

WAIT

WTF

I write in pt

[carmine timber]

What exactly are you trying to play ??

[wise anchor]

it printed a lot of json stuff

[mental adder]

is a webradio link here in Brazil, can I send it here?

[carmine timber]

Show me

[carmine timber]

Is that link of mp3 or live audio ??

[mental adder]

Live radio

Can I send it for you to see?

[carmine timber]

DM me the link

[mental adder]

I changed the stremtype so it's all right, the bot on my computer plays normally

[carmine timber]

Show me playing code also

[mental adder]

async play (url: string): Promise<AudioPlayer> {
    const player = createAudioPlayer()
    const resource = createAudioResource(url, {
      inputType: StreamType.WebmOpus
    })

    player.play(resource)

    await entersState(player, AudioPlayerStatus.Playing, 30e3)

    return player
  }```

[carmine timber]

Try to change streamtype to opus and then check

[mental adder]

ok, wait

[carmine timber]

I can't see any disconnect logs

[mental adder]

StreamType.Opus, right?

[wise anchor]

it disconnects if there is no song playing

that's in my code

like if there is no song playing it will disconnect

and if i remove the disconnect part it will just play nothing

[carmine timber]

yes

[carmine timber]

Do this error still comes ??

[wise anchor]

nope

no error apart from the debug part

[carmine timber]

So what's your issue now ??

[wise anchor]

it doesn't play anything

[carmine timber]

Remove debug code BTW

[wise anchor]

ight

[mental adder]

In my pc works, i'm sending to host

[carmine timber]

Ok

[carmine timber]

What is your issue now ??

[wise anchor]

.

[carmine timber]

Show me playing code

[wise anchor]

wait i think i know what is the problem

[carmine timber]

What is it ??

[carmine timber]

First of all, ytdl has memory leak issues and secondly, I will choose ytdl-core-discord at any point of time

ytdl-core has many issues

[wise anchor]

i use discord-ytdl-core

[mental adder]

@carmine timber nothing

[wise anchor]

the filters are the problem i think

[carmine timber]

That module also uses ytdl-core only with some extra ffmpeg args.
Try to use ytdl-core-discord

[mental adder]

What if I install this tweetnacl when I build it?

[carmine timber]

Try it

[carmine timber]

Did it work ??

[mental adder]

It's building

[carmine timber]

Ok

[carmine timber]

Just asking, why do you prefer ts over js ??

[mental adder]

it's because i'm learning to program with ts, for me it was a mess, but now it's starting to fit

@carmine timber ```JS

Core Dependencies

• @discordjs/voice: 0.4.0

• prism-media: 1.3.1

Opus Libraries

• @discordjs/opus: 0.5.3

• opusscript: not found

Encryption Libraries

• sodium: not found

• libsodium-wrappers: 0.7.9

• tweetnacl: 1.0.3

FFmpeg

• version: 4.3.2-static https://johnvansickle.com/ffmpeg/

• libopus: yes

--------------------------------------------------```

In my host

and again, nothing

[carmine timber]

Host is which version of linux ?? like ubuntu, desbian, centos ??

[mental adder]

Alpine

[wise anchor]

does ytdl-core-discord have ffmpeg args

[carmine timber]

No you need to separately pipe ffmpeg args

[mental adder]

@carmine timber I'll try without Alpine

[carmine timber]

Can you install ffmpeg separately in alpine

??

[mental adder]

oof, idk

[carmine timber]

Try running sudo apk add ffmpeg

[carmine timber]

Then for testing, try running ffmpeg -h

[mental adder]

ok

[carmine timber]

It worked ??

[mental adder]

Yes

i tested now

without alpine works

[carmine timber]

Oh

That means alpine is your problem

[subtle granite]

Is it possible to use this module on v12?

[carmine timber]

yes

[mental adder]

I've been trying to solve this for hours, thank you very much @carmine timber

[carmine timber]

😄

[carmine timber]

You need adapter file if you want to use it in v12 : https://github.com/amishshah/discord-radio-bot/blob/feat/upgrade/src/adapter/adapter.js

[carmine timber]

This is not place to ask, but the error says that you are converting a null to a string

Which is not possible

Show me what is there in play.js at line 55

[carmine timber]

are you using any message collector ??

Yes that's what's causing you the issue

Can you show me the message collector code

??

Where is the message(used in client.player.play) defined

??

[carmine timber]

You should have used message.content, I guess

You are playing a message class

Not sending any string

[wise anchor]

wait lemme show you my play function

it uses the message so it gets the message guild's queue

[carmine timber]

ok

[carmine timber]

ADS

WTF

@keen plaza Ban him

For me personally, I can't find any errors. Maybe ask more knowledgeable person

[wise anchor]

ight

thx for the help tho

[carmine timber]

😄

[bold cairn]

Can new-voice listen to more than 1 user at a time?

[carmine timber]

I think you can

[steel crescent]

Voice receive is not implemented yet

[carmine timber]

It is there in New Voice API

[steel crescent]

No

[carmine timber]

https://discordjs.github.io/voice/classes/voicereceiver.html

https://github.com/discordjs/voice/tree/main/src/receive

Receiving files

[steel crescent]

That’s far from done

[carmine timber]

I know that

But it is present

That's what I am saying

It is not complete but it is present

[steel crescent]

Doesn’t really matter if you can’t use it

[vocal valley]

its present but awful

[carmine timber]

Did you checked my research ??

[vocal valley]

I did, it is quite similar to ytdl-core-discord but there are a few pain points

• some videos may not have an opus 48000 format and will have to be transcoded (not handled by your example code)
• { inputType: ... } is ignored, so I would remove inputType

really it does the same thing as youtube-dl, except it has no fallback in case there is no opus/48000Hz format

[carmine timber]

Are you going to remove inputType from createAudioResource ??

[vocal valley]

i guess i can update the typings to show that inputType is not allowed with a string-based resource

[wind echo]

Can I play m3u8 format audio?

[vocal valley]

yes, via ffmpeg

all you'll need to do is install ffmpeg and it should just work

[wind echo]

oh good

[carmine timber]

Oh I see

[carmine timber]

BTW, you are also defining that you need 48kHz opus encoded sound in youtube-dl

[vocal valley]

yep but there is also /bestaudio fallback

in case that the opus format doesnt exist

[carmine timber]

Oh

What is function of entersState ??

[hallow pumice]

I have opus, but don't have libsodium-wrappers

[hallow pumice]

Installed it, same thing continues

[grim fulcrum]

Reinstall @discordjs/voice

[carmine timber]

You don't have discord voice module

Try typing this npm i @discordjs/voice@latest

[jolly mason]

is there a way to go forward/backwards in audio while its being played

like skip some seconds

?

[carmine timber]

Then you need to pass a ffmpeg instance to audio resource with seek options

[wise anchor]

hello, i keep getting this error when I try to play music FFmpeg/avconv not found!

[rustic temple]

Are you on windows?

[wise anchor]

Yes

[rustic temple]

Do npm i ffmpeg-static

[wise anchor]

Did that

Nothing changed

[rustic temple]

Mmh

Idk

Are you using the new voice module

[wise anchor]

Yes

[wise anchor]

i think i know the problem

[vocal valley]

how does your npm get so corrupted

[wise anchor]

anyway i fixed it by clearing cache

but still

ffmpeg/avconv not found

[vocal valley]

can you use generateDependencyReport and send the output?

example of using it is in the guide in the pinned messages

[wise anchor]

this guide? https://deploy-preview-595--discordjs-guide.netlify.app/voice/

@vocal valley

[vocal valley]

yep

idk what to say to be honest, if it's not finding ffmpeg it must just not be installed in the right place

where are you installing it

[wise anchor]

i installed it using npm install ffmpeg

[vocal valley]

is it in the example directory?

[wise anchor]

no...?

i used npm install so i assume in

node_modules

[vocal valley]

but where did you run the command

[wise anchor]

in my bot folder

[carmine timber]

You need to set environment path on windows

[vocal valley]

npm i ffmpeg-static, right

[carmine timber]

Without that nothing is possible

[vocal valley]

ok, in your bot folder do this in a terminal:

[wise anchor]

but it worked like a few hours ago

[vocal valley]

node
> require('ffmpeg-static');

what gets printed

[carmine timber]

Type ffmpeg in cmd ??

[carmine timber]

Type ffmpeg in cmd

[vocal valley]

how odd 🤔

[carmine timber]

See not set

[vocal valley]

but thats not the issue

if ffmpeg-static is installed, it should get resolved

[wise anchor]

but why did it work like 5 hours ago 🤔

[carmine timber]

That's the issue

[vocal valley]

in node_modules/prism-media, is there another node_modules folder?

[wise anchor]

lemme check

[vocal valley]

🤔

im really not sure why it's not resolving, all prism does is require('ffmpeg-static')

so if you can require it from your bot's root directory, it should work in prism too

[carmine timber]

Go to this PC and then right click --> Properties -->Advanced System settings --> Enviournmental Variables --> Path --> click on NEW and add the path where you have ffmpeg.exe

[wise anchor]

i tried requiring media-prism using cmd and it says module not found

wait

[carmine timber]

it is prism-media

not media-prism

[carmine timber]

@wise anchor Try this

[wise anchor]

oh

[wise anchor]

that works

[wise anchor]

ok

[vocal valley]

that will fix it yes, but it still doesnt explain why the module isnt getting imported in node.js

[carmine timber]

That's an issue with module, as I have to separately install ffmpeg in Linux and add path in Windows

[wise anchor]

umm where is the ffmpeg.exe if i installed it using npm

[carmine timber]

Go to node_modules/ffmpeg-static/

[wise anchor]

oh

now?

[carmine timber]

WTF

is this your ffmpeg-static

??

[wise anchor]

yes

wait

i ran node install.js

now it appeared

[carmine timber]

Yes

[wise anchor]

now it detects it

[carmine timber]

Now try restarting bot

[wise anchor]

ight

[jolly mason]

How to know if what the player is playing has ended or not?

is it done by player.on(AudioPlayerStatus.Idle, () => {}

[carmine timber]

yes player.on(AudioPlayerStatus.Idle, () => {})

[wise anchor]

how can i get the time that passed since it was playing

like let's say i have a song of 3 mins

i want to get the current time that it is at

[carmine timber]

https://discordjs.github.io/voice/classes/audioresource.html#playbackduration

[wise anchor]

ty

[steel crescent]

no, javascript is for losers

just remove the types and its javascript

not that hard

[fervent bloom]

Ummmm
So
When I ask it to join the stage
it joins as a listener
The bot
How do I make it join as a speaker?

[pulsar ibex]

https://github.com/discordjs/voice/tree/2b8a45687dab99c7f89291668d59d3d3a4df109a/examples/radio-bot

[fervent bloom]

C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\networking\Networking.js:413
                secretbox.methods.close(opusPacket, connectionData.nonceBuffer, secretKey),
                                  ^

TypeError: secretbox.methods.close is not a function
    at Networking.encryptOpusPacket (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\networking\Networking.js:413:35)
    at Networking.createAudioPacket (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\networking\Networking.js:397:53)
    at Networking.prepareAudioPacket (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\networking\Networking.js:321:37)
    at VoiceConnection.prepareAudioPacket (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\VoiceConnection.js:282:33)
    at C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\audio\AudioPlayer.js:426:54
    at Array.forEach (<anonymous>)
    at AudioPlayer._preparePacket (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\audio\AudioPlayer.js:426:19)
    at AudioPlayer._stepPrepare (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\audio\AudioPlayer.js:398:22)
    at prepareNextAudioFrame (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\DataStore.js:71:31)
    at audioCycleStep (C:\Users\HP\Documents\GitHub\Sharkymusic\node_modules\@discordjs\voice\dist\DataStore.js:56:5)

What is this error?

I tried to make it join a stage but it flipped with this message

Plus it joined the listeners

How do I make it join speakers?

[carmine timber]

Can you install libsodium-wrappers ??

There has been some issues with tweetnacl

[fervent bloom]

Oh jesus

How do I install it?

[carmine timber]

I think this error got patched in latest PR

[carmine timber]

Can you update voice module ??

to 0.4.0

[fervent bloom]

This is the latest version of discord/voice, I literally installed it minutes ago