3

I'm currently making a Discord bot that has a "User Recording" feature, and I was in a VC testing it, and kind of noticed that the output .pcm file was ~20GB after 13 minutes.

        this.voiceChannel = await message.member.voice.channel.join()
        this.reciever = this.voiceChannel.receiver
        this.voiceChannel.on('debug', (debug) => {
            let packet = JSON.parse(debug.slice(8))
            console.log(packet.op)

            if(!packet.d || packet.d && packet.d.speaking != 1) return;
            let user = this.client.users.resolve(packet.d.user_id)
            if(packet.d.speaking) {
                let userStream = this.reciever.createStream(user, {mode: 'pcm', end: 'manual'})
                let writeStream = require('fs').createWriteStream('./recording.pcm', {})
                this.us = userStream
                this.ws = writeStream

                this.us.on("data", (chunk) =>{
                    console.log(chunk)
                    this.us.pipe(this.ws)
                })
                this.ws.on("pipe", console.log)
            }
        })

Is there any way possible to compress the .pcm file from.. I dunno, 20GB down to 5-10 MB? This seems odd as each Buffer that comes from Discord.js is a whopping 4000 bytes (4KB) (This also had my Disk capped out at 100% and writing at 60MB/s)

Mozza
  • 149
  • 2
  • 11
  • Does this answer your question? https://stackoverflow.com/a/43829712/13176517 – Syntle May 09 '20 at 21:40
  • @Syntle It wasn't putting a file into a zip/compressed folder. It was more of compressing the data on receive, and putting the compressed data in the file itself. None of the less, that link did help me dig deeper into modules of the same, and helped out. – Mozza May 11 '20 at 12:59

1 Answers1

4

I'm going to answer my own question, but I believe I know what I did wrong.

this.us.on("data", (chunk) =>{
    console.log(chunk)
    this.us.pipe(this.ws)
})

I was incorrect in that section, because I was sending data whenever data was received, not once, but twice. I also used the zlib module (https://npmjs.org/package/zlib) and that helped even more with Data / Voice Compression. 

this.us.on("data", (chunk) => {
    let slowBuf = this.zlib.deflate(chunk, (er, res) => {
        console.log(res)
        this.ws.write(res)
    })
})

That worked like a charm and is now writing data at 600KB per 2 minutes or so of testing.

Mozza
  • 149
  • 2
  • 11