How To Set Up Audio On Obs Studio

These features require OBS-Studio 21.0 or newer.

---

Tabular array of Contents:

• Overview
  • Use cases
• Setup
• Gaming and Streaming Services Support
  • Uniform Streaming Services
  • Incompatible Services
• Streaming Servers and Players
  • Uniform Streaming Servers
  • Compatible Spider web Players
• Streaming Music
  • Filters and VST
  • High Quality Recording
  • Interfacing with a DAW
• Ambisonics
  • College Gild Ambisonics
• Multilingual Streaming
• Detailed list of surround audio features

---

Overview

OBS-Studio is the offset mainstream streaming software to support surroundings sound streaming and recording.
Traditionally this is a characteristic reserved to professional broadcast appliances.
OBS-Studio can stream and record upwardly to 8 audio channels.
The sound channels can be surround sound channels or more full general multichannel ones.

Utilize cases:

• gaming
• music
• ambisonics
• multilingual streaming

Setup

• Settings > Sound : merely select a channel different from Mono or Stereo in the Channels listing. Click OK to the Warning pop-up and restart OBS.
  

• Settings > Output :

• IMPORTANT Warning: make sure to select the same channel layout as your input (if y'all have a 4.one sound source practise not select 7.1). If you lot don't, channel mixing may (or may not) occur. There is an automated aqueduct rematrixing when either downmixing or upmixing is mandated past a difference in channel layouts between source and output. This aqueduct rematrixing mixes channels in full general. Or it tin remove a channel (ex: 3.i source to 4.0 output removes the LFE channel).

• DOLBY
  Streams can be encoded in ac3 /eac (using Output > Advanced > Custom FFmpeg recording > stream to URL ). But support of the various streaming services or web players has not been tested.
  Alter this wiki if this works.
  Capture of dolby can be tricky if the channels are lumped into two PCM channels; in gild to be decoded correctly and encoded all the channels should be held in unlike PCM channels.

Gaming and Streaming Services Back up

Uniform Streaming Services

Equally of at present, the following services take been tested and are compatible with live surround sound streaming:

• Twitch,
• Mixer (rtmp but NOT with ftl protocol),
• Smashcast ...
• Facebook Live 360 with spacial audio (requires ambisonics capture device)

Incompatible Services

• YouTube Alive (discards channels beyond the showtime two).
• Facebook Live (downmixes all the channels).

Streaming Servers and Players

Apart from streaming to services like Twitch or Facebook Live, you might use your ain streaming server which delivers streams to your website.

Compatible Streaming Servers

The following servers accept been tested and ingest surround sound in rtmp protocol :

• wowza
• nginx with rtmp module

Note however that server-side recordings simply proceed the first two channels although.

Uniform Web Players

• html5 players (tested with hls or mpeg-nuance):
  • videojs,
  • bitmovin,
  • mediaelement,
  • viblast (hls & dash),
  • hls.js

NB: Flash histrion accepts rtmp surround sound in Dolby; information technology might piece of work ==> untested.

Streaming Music

Up to 7.1 surround sound is available. (For more than channels (up to 16.0) check this fork).

Filters and VST

OBS-Studio has born audio filters as well equally VST 2 support. The filters are compatible with surround sound; the VST too if they originally support multichannel. Check the Filters-Guide.

High Quality Recording

Select Custom FFmpeg recording and a PCM format (e.g. pcm_s24le for 24 scrap samples) for uncompressed audio.
For compressed formats, you tin can select likewise libopus and aac which will work very well (target at least 64 kbs per aqueduct).

Interfacing with a DAW

For that you lot volition need some application ensuring the routing.

windows

Reaper :

• SAR (requires having an ASIO driver, for example ASIO4ALL if you lot don't own an ASIO carte du jour).
• rearoute (from Reaper, with upwardly to 256 channels) if using ASIO plugin
• Voicemeeter (Excellent guide here by Jack0r for Voicemeeter Banana)
• Virtual Audio Cable (requires v.4.50 or later).
  
• For routing in Windows, bank check this excellent guide again past Jack0r. This other guide may be of interest too.

MacOS

Tested with Reaper and the following apps ensuring the routing

• Soundflower
• Jack server : cheque this excellent guide.
• Loopback
• untested: iShow U (guide hither).
• untested: WavTap (guide hither).

Linux

No DAW tested. Update the Wiki if you lot tested.

Ambisonics

Although the channel layouts are tagged with position (ii.1 5.1 etc), it is possible to encode your channels for ambisonics use provided your decoding application is setup to supply the ambisonics positions according to aqueduct order.
This is the case for Facebook Live 360 with spacial sound.
To use the latter in Settings > Sound, select : channels > 4.0 ; aac codec is mandatory (if you apply Recording > Custom FFmpeg > to URL > select flv container, with rtmp URL to FB live 360 and aac audio codec + x264 video codec).
When streaming to FB Live 360 the four channel rtmp stream will be interpreted as carrying ambisonics of gild ane .

In order for this feature to work, obviously y'all volition need an ambisonics capture device.
If you have an aac or uncompressed audio recording with 4 ambisonics, you can as well play it past adding a Media Source.
(If it doesn't piece of work, bank check with ffmpeg that it is decoded correctly.)

College Order Ambisonics

No live streaming service supports beyond starting time-order ambisonics.
For recording, though for order 2 or 3, you can apply pkviet's fork.
If you record with libopus (Output > Advanced > Custom FFmpeg recording ) up to 255 channels are available with mkv.
You will have to add the mapping_family=255 option though to FFmpeg audio encoder options in OBS-Studio.

Multilingual Streaming

Multichannel support in OBS allows to stream several languages simultaneously.
This is useful for live translation (public talks ... ).

The mainstream streaming services do not support such a characteristic directly.
In the broadcast industry, one usually uses mpeg-ts multi-track streams instead of a multichannel stream (in single track).

In that location are workarounds though allowing 1 to utilise a unmarried rails multichannel audio.
Here is one requiring :

• nginx with rtmp module;
• ffmpeg scripts which will be exec'd by nginx: ffmpeg will split the audio channels and create mono rtmp streams (as many as there are languages).
• each of these mono rtmp streams tin can and then exist pushed past nginx to a service like FB or YouTube Live.

For nginx setup with rtmp module, check elsewhere. (Make certain to prepare it up with a unmarried worker.)

The setup is the following for environment sound capture with each mono aqueduct conveying a language:
(capture tested: sdi/hdmi decklink cards, reaper, Behringer X32, ASIO sound cards).

• obs with environment sound enabled

• rtmp stream to nginx which is setup with an exec script

                rtmp {          server {         heed 1935;     ping 30s;      notify_method get;        application splitter {       live on;       allow play all;       exec /abode/me/splitter.sh $name;       exec_kill_signal term;       record off;    } }            

  • here nginx executes a ffmpeg script which volition carve up the channels and redirect rtmp mono streams to nginx.

FFmpeg instance script (for two languages and a stereo source):

          #!/bin/bash echo "$(date +"%Y/%m/%d %H:%M:%Due south"): starting" on_die() {   # kill all children   pkill -Kill -P $$ } trap 'on_die' TERM  ffmpeg -i rtmp://IP:port/splitter/ \        -filter_complex \           "[0:a]pan=mono|c0=c0,aresample=async=1000[a0];             [0:a]pan=mono|c0=c1,aresample=async=grand[a1]" \        -map 0:v -copyts -start_at_zero -vcodec copy \          -map [a0] -bsf:a aac_adtstoasc -copyts -start_at_zero -c:a libfdk_aac -ab 64k -ac one \          -f flv rtmp://IP:PORT/app/stream_language1 \        -map 0:five -copyts -start_at_zero -vcodec copy \          -map [a1] -bsf:a aac_adtstoasc -copyts -start_at_zero -c:a libfdk_aac -ab 64k -ac 1 \          -f flv rtmp://IP:PORT/app/stream_language2 & wait        

For more than languages option a respective channel layout and add relevant streams in FFmpeg script. You should also read the ffmpeg filtering guide. Note that for 5.one and seven.1 one aqueduct (the 4th) will exist encoded equally LFE and and then is not useable. For up to 16 channels support, check the pkviet's fork.

Detailed listing of surround audio features

• Recording and Streaming multichannel audio sources (surround audio).

• Compatible streaming services: Twitch, Mixer rtmp (not ftl), smashcast, FB 360 live

• Compatible protocols:

  • rtmp
  • mpeg-ts tcp udp (others untested).

• Streaming servers tested:

  • wowza,
  • nginx-rtmp
    (rtmp with multichannel audio passes through and can be distributed by these servers to another service or cdn supporting environment sound; however, the recording feature of these servers does not work; only the get-go two audio channels are kept).

• html5 players tested and working with live surround:

  • videojs,
  • bitmovin,
  • mediaelement,
  • viblast (hls & nuance),
  • hls.js

• Compatible containers (for recordings):

  • mkv
  • mp4
  • ts
  • flv (others untested).

• Compatible codecs:

  • ffmpeg aac (native encoder, up to 16 channels),
  • libfdk_aac (upwardly to 7.1),
  • cadre audio aac (up to 7.1),
  • opus (libopus encoder, upwards to 255 channels),
  • vorbis (upwards to 7.1),
  • pcm (others untested).

• OS: uniform with win, macOS, linux (alsa, pulse-audio).

• Misc:

  • Higher audio bitrates (up to 1024 kbs) unlocked to accomodate higher number of channels.
  • audio monitoring, audio filters, VST are all working OOB.

Source: https://obsproject.com/wiki/Surround-Sound-Streaming-and-Recording

Share this post