From dd3fc2142076062683f5d691303a1c7763b887ca Mon Sep 17 00:00:00 2001 From: rmens Date: Sun, 12 Nov 2023 00:23:51 +0100 Subject: [PATCH] Update README.md --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 152420d..3d176c9 100644 --- a/README.md +++ b/README.md @@ -1,7 +1,7 @@ # rpi-audio-encoder This repository contains the audio streaming software for [ZuidWest FM](https://www.zuidwestfm.nl/) in the Netherlands. It uses a Rapsberry Pi 4 and a [HiFiBerry Digi+ I/O](https://www.hifiberry.com/shop/boards/hifiberry-digi-io/) as audio input. As encoder ffmpeg is used, which is combined with Supervisor to manage the process via a webinterface. It sends audio to an Icecast2 or SRT server. -This encoder resides in the studio and is connected to an Optimod. It can stream to any Icecast or SRT server. Our server software to complete the audio stack can be found in [this respository](https://github.com/oszuidwest/liquidsoap-ubuntu). +This encoder resides in the studio and is connected to the digital output of an Orban Optimod. It can stream to any Icecast or SRT server. Our server software to complete the audio stack can be found in [this respository](https://github.com/oszuidwest/liquidsoap-ubuntu). @@ -14,7 +14,7 @@ This encoder resides in the studio and is connected to an Optimod. It can stream ⚠️ There is a [major problem with kernel verions 6.x and recording from a HifiBerry with FFmpeg](https://github.com/raspberrypi/linux/issues/5709). Linux versions that ship with kernel 6.0 or newer do not work. We strongly recommend using Ubuntu 22.04 LTS with kernel 5.15. Raspbian 11 or 12 ship kernel 6.x and do not work. ⚠️ # How to configure the audio processor -- Connect the digital output of the audio processor to the input of the HiFiBerry +- Connect the digital output of the audio processor to the input of the HiFiBerry. - Ensure the processor is sending out 48khz 16-bits audio. The HiFiBerry can't resample. This is hardcoded. - If possible, configure the digital output to send SPDIF data. AES/EBU could work, but is not 100% the same standard. @@ -23,16 +23,16 @@ _This is an example for an Orban Optimod:_ # Audio encoding presets -There are a couple of audio encoding configurations: +These audio encoding presets are inclued. They are limited to what Icecast supports: - `mp2` sends MPEG-1 Audio Layer II audio on 384 kbit/s. This is considered the gold standard for compressed broadcast audio. - `mp3` sends MPEG-1 Audio Layer III audio on 320 kbit/s. This is the highest quality mp3 possible. - `ogg/vorbis` sends OGG Vorbis audio on 500 kbit/s. This is the highest quality ogg/vorbis possible. - `ogg/flac` sends FLAC audio in an OGG wrapper on ~1200 kbit/s. This is the highest possible uncompressed audio. ### SRT support -Besides Icecast we now support SRT for streaming. It's still in beta since we didn't do an endurance test yet. A working SRT implementation of the server software [can be found here](https://github.com/oszuidwest/liquidsoap-ubuntu/tree/srt). +Besides Icecast we also support SRT for streaming. It's stable but we didn't do an endurance test yet. A working SRT implementation of the server software [can be found here](https://github.com/oszuidwest/liquidsoap-ubuntu/). In the future Icecast support might be deprecated, but we are giving SRT a bit more time to prove itself. -For more about SRT: +More information about SRT: - SRT overview: https://datatracker.ietf.org/meeting/107/materials/slides-107-dispatch-srt-overview-01 - SRT deployment guide: https://www.vmix.com/download/srt_alliance_deployment_guide.pdf - SRT 101 video: https://www.youtube.com/watch?v=e5YLItNG3lA