Primitive voice implementation (feature/voice)

[kozabrada123]

(opening this into a pr to track changes, progress, todos and merge conflicts)

This pr (when complete) aims to add voice channel and call support to Chorus via Discord's UDP api.

Current todo:

Features:

• [X] Voice gateway connection and heartbeating
• [X] Voice UDP connection
• [X] Ip Discovery
• [X] Select protocol
• [X] Receiving and deserializing audio
• [x] Audio decryption
• [x] Audio encryption
• [x] Sending and serializing audio
• [x] Public API

Docs:

• [x] Documentation on related structs, methods etc
• [x] Usage examples

Compatibility:

• [x] Gateway WASM combatibility
• [ ] UDP WASM compatibility - the browser cannot do udp.

Code quality:

• [x] Fix or at least clarify FIXMEs

UDP vs Webrtc

User accounts sometimes use an undocumented Webrtc api for voice and video, while bots use the documented UDP api. I've decided to implement the UDP api for now, since Webrtc is much more complex and less documented. We may want to look into Webrtc aswell in the future.

Architecture and public API

My current plan is to have four parts of the impl: the Voice Gateway, voice UDP, Voice Data and Voice Handler

• Voice Handler would handle coordinating and initializing the Gateway and UDP connections,
• Voice Data is a struct, which contains needed shared data, such as the encryption key, ssrc, session id, ip & port, etc..
• Voice Gateway and Voice UDP are the communication channels for Discord's api, using websockets and udp respectively.

The user can then interact with the Voice Gateway to issue gateway commands and receive voice context. The voice UDP handler would expose a slightly higher level api, with listener events for decrypted and deserialized audio as well as a method which serializes, encrypts and sends audio.

[kozabrada123]

Welp these new build errors on wasm are just because I have no compatibility yet

[kozabrada123]

Update: I am now be 100% sure I am properly receiving and decrypting the audio

[kozabrada123]

I've tried to make a very simple app that echoes voice data back, but the sent data never arrives to the original client

[kozabrada123]

Maybe I messed something up in the sending code or Discord might just have a very sophisticated algorithm :tm: that just ignores packets if they have duplicate data

[kozabrada123]

Github is drunk, idk why it wanted to merge w/o any new commits on dev

[kozabrada123]

Update on this, I haven't had the time to work on this for a while, I'll see when I can restart this the next thing I will probably try is encoding and decoding files with a crate

[kozabrada123]

If I can get decoding and encoding working in an example app, I can see whether the ddc client plays my encoded and serialized audio. if not, then I've messed something up / something is still missing.

If it does play, I can ship the app's code as an example and get to merging this!

[kozabrada123]

The transport layer works perfectly, I just can't decode ddc's opus for some reason :/