Web Audio Weekly


Hello! Welcome to the April edition of Web Audio Weekly. This issue is kindly sponsered by the folks at Nada Brahma. They're looking for a Web Audio developer so please check out their post below if you're looking for a new gig. If you'd like to advertise in next month's newsletter, please get in touch. Stay safe and healthy, friends.

-- Chris



Typebeat is a hybrid live performance and live coding instrument that uses the computer keyboard as an instrument. The author Kofi explains:

Typebeat turns keystrokes into music production commands. It's a fast, hands-on workflow for sampling, synthesizing, and sequencing sound.

Combining a classy two-tone design with some innovative features, it's really fun to play. Kofi digs into the background and inspiration for Typebeat in the accompanying blog post.

Ableton's Learning Synths adds exporting, recording and more


Released in 2019, Ableton's Learning Synths project remains one of the most polished and compelling uses of Web Audio in an educational context. While we probably won't hear about what they're working on next until it arrives, it's exiting to see that Learning Synths continues to receive updates. It now features recording (up to 60s of audio for you to drop into your favourite Web Audio groovebox for example), export to Ableton Live and the customary dark mode. A good excuse to revist, or to experience for the first time.

Nada Brahma need a Web Audio developer


Nada Brahma Ltd. is searching for developers! We are building a rhythm sampler WebApp with predefined rhythmical patterns arranged in a matrix. The ideal candidate has experience with web audio and knows about the latest developments and libraries. Knowledge about xenharmonics/microtonality is a big advantage. You will be part of our Development Team and report to our CTO. The position can be full- or part-time for a fixed period, with the possibility to join the Company. Salary to be negotiated. Interested? Email julene@nadadigital.com .

Open-Unmix demo, umx.js


Open-Unmix uses a trained deep neural network to perform source separation on mixed audio files. It can separate a recording into vocals, drums and bass tracks, for example. I noticed recently that the reference implementation includes a version ported to JavaScript using tenserflow.js which can be used to test the models on shorter files before commiting to getting the full pipeline working.

Coincident Spectra


The twelve-note musical scale we are familar with from a piano keyboard is a compromise between the "natural" tuning of vibrating strings and the ability to play in any key. Brian Ginsburg's synth (inspired by the work of William Sethares) allows you to explore what happens when you break away from that convention:

Coincident Spectra is an additive synthesizer that generates sounds from 16 sine wave 'partials' that can be switched between harmonics and spectra. Coincident spectra reduce sensory dissonance and beating between the timbre and tuning, rendering chords more in tune.

Build a DAW plugin in JavaScript


An introduction to building a VST plugin for any DAW (digital audio workstation) using JavaScript and the Elementary library.

Wide-band WebSDR


Although this site was first launched in 2008, and I'm therefore a bit behind the times, it's too good to pass up. Enjoy listening to static noise while controlling a network-connected software-defined radio receiver at the University of Twente. All the fun of Ham Radio without installing a massive antenna on your roof.

What's this?

Web Audio Weekly is a newsletter about audio on the web platform, curated by Chris Lowis. Check out the archives for issues you may have missed and let your friends know they can subscribe by forwarding them this newsletter.


Buy Me A Coffee