<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>MIDI on lavafroth</title><link>https://lavafroth.is-a.dev/tags/midi/</link><description>Recent content in MIDI on lavafroth</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Mon, 08 Jun 2026 19:52:23 +0530</lastBuildDate><atom:link href="https://lavafroth.is-a.dev/tags/midi/index.xml" rel="self" type="application/rss+xml"/><item><title>Becoming All the Instruments</title><link>https://lavafroth.is-a.dev/post/becoming-all-the-instruments/</link><pubDate>Mon, 08 Jun 2026 19:52:23 +0530</pubDate><guid>https://lavafroth.is-a.dev/post/becoming-all-the-instruments/</guid><description>&lt;p&gt;I&amp;rsquo;ve been working on a tool using very simple math
to convert humming or human whistles into musical MIDI notes.&lt;/p&gt;
&lt;p&gt;Here&amp;rsquo;s how the idea morphed into its current form.&lt;/p&gt;
&lt;p&gt;First, I tried using Fourier transform over chunks of the audio, which is known
as the STFT. I wanted to get close to straight lines for each humming pattern
or whistle. To isolate these lines, we could chunk the audio by note boundaries.
I wanted to something simple that the user can supply, preferably without using
any neural networks or fancy math.&lt;/p&gt;</description></item></channel></rss>