https://github.com/1j01/precorder
🎙 Record audio from the past ⌚⌛
https://github.com/1j01/precorder
audio audio-recorder audio-recording magic microphone postwhimsy precord precorder psuedo-magic record record-audio recorder recording sox time-travel
Last synced: 3 months ago
JSON representation
🎙 Record audio from the past ⌚⌛
- Host: GitHub
- URL: https://github.com/1j01/precorder
- Owner: 1j01
- Created: 2016-03-08T05:06:24.000Z (about 9 years ago)
- Default Branch: master
- Last Pushed: 2020-08-04T03:44:42.000Z (almost 5 years ago)
- Last Synced: 2025-01-08T08:40:23.569Z (4 months ago)
- Topics: audio, audio-recorder, audio-recording, magic, microphone, postwhimsy, precord, precorder, psuedo-magic, record, record-audio, recorder, recording, sox, time-travel
- Language: CoffeeScript
- Homepage:
- Size: 18.6 KB
- Stars: 3
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Precorder
Psuedo-magically record audio from the past.
When improvising, you can't know when you're *about* to come up with something good.
And you can't always reproduce what you came up with afterwards.
And you shouldn't have to try to.
With Precorder, you don't.
-----
I like to improvise on the piano,
but when I go to record what I just played,
I might recall the "gist" of it,
but I can't remember a lot of the variations
that I came up with originally while experimenting.
At the very least it can have a different "feel" the second time.
And sometimes I can't remember much of it at all.
If I always started a recording before I played,
I would end up saving a ton of long, largely mediocre recordings.
I would think, "well, there's some good stuff in there."
I wouldn't listening back to the recordings and edit each one down to contain only whatever good new stuff was in there.
"If only I could go back in time to record."
With this tool, you still have to hit a button to save a recording,
so you don't end up saving lots of stuff you don't really want or care about,
but you don't have to hit a button to record,
so you can capture creative whimsy, post-whimsy.
\#postwhimsy
## Requirements
* A more-or-less dedicated Raspberry Pi or comparable device
* A more-or-less dedicated USB microphone
(or if you want to use a non-USB microphone on the Pi,
you'll need to buy a USB sound card (or perhaps a sound card "HAT"? (Hardware Attached on Top)),
and you may need to [jump through some serious hoops](http://www.g7smy.co.uk/2013/08/recording-sound-on-the-raspberry-pi/))
* Node.js and SoX (installation instructions below)
* Willingness to use the command line
* Who knows, because I haven't gotten this working yet (at least not on the Pi)
## Setup
I'm not sure this thing works yet, so you probably shouldn't bother with it,
but these are the approximate steps it would take:
* `ssh` into (or open a terminal on) the device you wish to use
* Install a recent version of [Node.js][].
The best way is to use [`nvm`][nvm]; that way you can get the latest versions, and can switch versions easily later.
You can run [their installer][nvm installtion] and then run `nvm install --lts` to get a recent, supported version.
* Install [SoX][] with `sudo apt-get install sox`
* Clone the project with `git clone https://github.com/1j01/precorder.git`
* Enter the project folder, i.e. `cd precorder`
* Run `npm i` to install dependencies
* Run `npm link` to make the `precord` CLI available
* Plug in your USB microphone or sound card
* Find an ALSA device ID such as `hw:0,0`
by running `arecord -l` and looking at the numbers where it says `card N` and `device N`;
the ID should be of the form `hw:card,device` where `card` and `device` are the numbers from that command
(maybe it could need `plughw` in place of `hw`, idk)
* Set some environment variables, using the device ID for `PRECORDER_AUDIODEV`:
e.g. `export PRECORDER_AUDIODEV="hw:0,0" ; export AUDIODRIVER="alsa"`
* Run `npm start`
* Wait for it to say "wrote chunk file"; it might show an error at this point
* If it worked, run `precord 1min` to save an audio file in `data/output/`
(and make sure it's an actual recording you can play)
## TODO
* Get it working!
* Set up `npm start` as a daemon process and allow running at startup; maybe use `npm stop` to stop it
* Make the CLI actually respect the time given
* Make the CLI accept an output file path
* Handle stdout specially: maybe add an alias (`-`); don't output other stuff along with the audio file data
* Polish up and publish to npm
[Node.js]: https://nodejs.org/
[nvm]: https://github.com/creationix/nvm
[nvm installation]: https://github.com/creationix/nvm#installation
[SoX]: http://sox.sourceforge.net/