https://github.com/ennuicastr/ennuicastr

Ennuicastr is a system for recording multiple users distributed across the world in a well-synchronized way, without significant loss, over the web.
https://github.com/ennuicastr/ennuicastr

audio webrtc

Last synced: 3 months ago
JSON representation

Ennuicastr is a system for recording multiple users distributed across the world in a well-synchronized way, without significant loss, over the web.

• Host: GitHub
• URL: https://github.com/ennuicastr/ennuicastr
• Owner: ennuicastr
• Created: 2019-08-21T20:02:48.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2024-05-13T19:49:21.000Z (9 months ago)
• Last Synced: 2024-05-21T20:14:29.263Z (9 months ago)
• Topics: audio, webrtc
• Language: TypeScript
• Homepage: https://ecastr.com/
• Size: 3.39 MB
• Stars: 69
• Watchers: 5
• Forks: 5
• Open Issues: 13
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
Ennuicastr is a system for recording multiple users distributed across the

world in a well-synchronized way, without significant loss, over the web.

There are a few problems that arise with distributed recording:

* Voice-over-IP solutions are the natural way to distribute voice

  communication, but all sacrifice completeness for performance. That is, all

  suffer loss in a way that should not be necessary to record. Ennuicastr uses

  existing voice-over-IP technology (namely, WebRTC) for the actual live

  communication component.

* Recording each user's audio separately usually requires that each user

  perform their own recording manually, and then send the audio to a central

  source. Aside from causing synchronization nightmares (see the next bullet

  point), this demands a fair amount of expertise from each user, and is very

  fragile: If even one user's audio doesn't work out, usually the only viable

  backup is a single recording of all users. Ennuicastr resolves this by

  packaging recording into a simple web-app, and sending the recorded audio

  live to a central server.

* Distributed users have different clocks. Even with initial synchronization,

  the clocks have different crystals, and so will drift naturally from each

  other over time. The traditional technique of having every user record

  separately, even when no one makes any mistakes, still creates headaches for

  editing, as the tracks will drift out of sync. Ennuicastr resolves both of

  these problems by having the client software synchronize its time with the

  server *continuously*, and timestamp *every* frame of audio data. Server-side

  software can then resolve the timestamped audio frames into continuous

  streams which are correclty in sync, by removing or adding silence as

  necessary.

* Most communication software does considerable processing, to make

  communication more pleasant, but for editing a recording, this preprocessing

  is at best unnecessary, and at worst counterproductive. Ennuicastr resolves

  this by... well... not doing it. Recorded audio is as raw as possible. It can

  even be recorded in lossless FLAC if you so choose.

Note that this repository includes only client-side software; the server side

is in ennuicastr-server. A small example server is provided in this repository

for testing.

Ennuicastr is licensed under the ISC license, and can be used for more-or-less

any purpose so long as you retain the copyright notices.

Ennuicastr has a number of dependencies each under their own compatible

license. See the directories vad and libav for more information on them.

No documentation is provided for running your own instances of Ennuicastr or

for its protocol.