Code Log: In which I explore how to make sounds
What is a code log? It is sort of like live coding, but badly done, and all in text.
I’ve drawn a heap of pixels to the screen over the years, but I haven’t made computers bleep or bloop much. Here are some notes on some quick audio explorations.
Git gets grumpy
First, a quick aside on a silly error I ran into trying to clone a git repo.
firstname.lastname@example.org: Permission denied (publickey). fatal: Could not read from remote repository. Please make sure you have the correct access rights
Despite having everything properly configured in my
~/.ssh/config git wasn’t playing nice with Codeberg. To remedy I ran:
my_awesome_key was the name of my private key.
This ensures that your ssh-agent is aware of the specific key you are trying to use.
My goal is to either emit 1 or more tones or to generate some sort of white noise. Ideally I would like the running program to play the audio directly, but I’d also settle for the program writing a file that can then play the audio.
The silliest/simplest way that I know of to make my computer bloop is to send the
\a escape sequence to stdout.
\a is the escape sequence for alert, so, at least on macOS, it triggers the OS-level alert sound.
echo -e '\a'
This, however, is lame and feels like cheating.
I know that in order to make my computer bleep and bloop I will need access to its audio APIs. Some quick searching surfaced a variety of ways to do this — the ones I’m most familiar with leverage game engines, tools like löve2d. Those also sort of feel like cheating. So, I continue onward!
A brief sampling of some stuff found in quick research mode
After doing a quick
pip3 install simpleaudio I got the sample to play what sounds like a piano!
import simpleaudio.functionchecks as fc fc.LeftRightCheck.run()
This is very promising!
Next up I tried to run one of
simpleaudio’s more involved examples that leverages numpy, a dependency I know nothing about 🤷…and that threw the following error:
Traceback (most recent call last): File "/Users/eli/Desktop/noise.py", line 12, in <module> t = np.linspace(0, T, T * sample_rate, False) File "<__array_function__ internals>", line 5, in linspace File "/usr/local/lib/python3.9/site-packages/numpy/core/function_base.py", line 120, in linspace num = operator.index(num) TypeError: 'float' object cannot be interpreted as an integer
I MUST REMAIN FOCUSED ON THE BLOOPS! Gonna nope out of this!
WWW Audio API
If I resemble any sort of developer it is a web developer; the web calls to me!
Researching the Web Audio API brought me to this excellent demo. An excellent demo that didn’t work in my default browser!? This was devastating and led me to realize (perhaps (read as “most assuredly”) falsely?) that Web Audio API feels like a cheat because it isn’t my computer beeping, it is my computer as dictated by my browser…this is an absurd reason for giving up on it, but it was my reason nonetheless.
Next I took a brief excursion into ASM, thinking that if I want my computer to really bleep and bloop I gotta do it in assembly…but…wooof. It’d be rad., but perhaps not for this code log.
Diversions and sleepiness set in
With the wanderings into ASM complete I was getting sleepy — desperate times and all that — my measure became desperate.
Rather than play the tone directly what about writing it to a file and playing that back? Here is a post about generating a
.wav file in C, and another post from the same source about doing it in Python.
Promising, but a bit complex.
#programming, but I know I can generate a very simple
.wav file with
ffmpeg -f lavfi -i "sine=frequency=1000:duration=5" beeeeep.wav
Or, a lower tone
ffmpeg -f lavfi -i "sine=frequency=800:duration=5" bloooop.wav
I didn’t 100%, or even really 60% accomplish exactly what I set out to initially, but for my inaugural code log I’m willing to call this a success.
Some other programs I’d like to explore more related to computer audio:
EDIT: Don’t miss the follow up post!