#recording #sound #pim #audio-devices #akashic

app akasha

Headless recording app, designed to run indefinitely and automatically break microphone stream into segments for ease of handling. Supports WAV and OGG format.

7 releases (stable)

1.2.4 Feb 25, 2023
1.2.3 Feb 23, 2023
1.2.2 Jan 11, 2023
0.1.0 Jan 5, 2023

#207 in Audio

AGPL-3.0-or-later

37KB
760 lines

Akasha

.github/workflows/rust.yml crates.io

Akasha is a headless recording app, designed to split recordings into multiple segments to make them easier to manage.

It's designed for low-resource consumption, making it useful as a background process for e.g. recording meetings, without having to keep OBS up, or write a jank shell utility involving sox or ffmpeg.

It theoretically supports every major platform (though I haven't bothered to test it on non-Linux ones, I have tried to only use platform independent libraries).

It supports autodetecting input devices, manually specifying devices by name, and features a reasonably intuitive command line. To get started, most of the time, you can do:

akasha rec --path-dir ~/MyAudioDirectory/

Where --path-dir is the directory it will write files into.

If you do not specify --path-dir, e.g.:

akasha rec

...it will default to ~/Audio/akasha/, where ~ is your home directory.

Any directories that do not already exist, will be automatically created.

If the default audio device it detects is incorrect, you can override it. First, get a list of all available input devices:

> akasha probe --type input-devices
ALSA lib pcm_oss.c:397:(_snd_pcm_oss_open) Cannot open device /dev/dsp
ALSA lib pcm_oss.c:397:(_snd_pcm_oss_open) Cannot open device /dev/dsp
ALSA lib pcm_route.c:877:(find_matching_chmap) Found no matching channel map
ALSA lib pcm_route.c:877:(find_matching_chmap) Found no matching channel map
ALSA lib pcm_route.c:877:(find_matching_chmap) Found no matching channel map
ALSA lib pcm_route.c:877:(find_matching_chmap) Found no matching channel map
ALSA lib pcm_route.c:877:(find_matching_chmap) Found no matching channel map
ALSA lib pcm_route.c:877:(find_matching_chmap) Found no matching channel map
ALSA lib pcm_dsnoop.c:566:(snd_pcm_dsnoop_open) unable to open slave
ALSA lib pcm_dsnoop.c:566:(snd_pcm_dsnoop_open) unable to open slave
[
    "jack",
    "pipewire",
    "pulse",
    "default",
    "sysdefault:CARD=Generic",
    "front:CARD=Generic,DEV=0",
    "surround40:CARD=Generic,DEV=0",
    "surround51:CARD=Generic,DEV=0",
    "surround71:CARD=Generic,DEV=0",
] 

You'll have to forgive the ALSA vomit -- the cause seems to be CPAL not setting the error handler correctly. Maybe I'll make a PR and fix that at some point.

Anyway, you can now pull audio from that specific device:

akasha rec --device "pulse" --path-dir ~/Audio 

There's also a cute real-time display of volume intensity, that works using SIMD calculations of audio volume via RMS. You can pass the --display flag if you want that.

TODO:

  • Add warnings for if f32 stream is all zeros, so that the user doesn't have to enable --display mode to tell.
  • Refactor error handling logic with snafu.
  • Nicer error messages
  • Eventually refactor out all the stream logic into audio-stream crate, maybe?
  • Default recording path to make more novice-friendly.
  • Maybe fix raw mode not working with logging facilities (/n -> /r/n).
  • Fix tracebacks not working with raw mode (https://werat.dev/blog/pretty-rust-backtraces-in-raw-terminal-mode/)

Dependencies

~13–50MB
~757K SLoC