Web Audio API custom DSP: Automatic spectral balancer (@NewfangledAudio EQuivocate dynamic EQ?)

Yep, I used triangular-shaped FIR crossover filters to split up the signal to 26 bands equidistantly spaced in Mel scale (f(x) = log2(1+x/700) and g(x) = 700 * (2^x - 1) respectively) and then feed multiple audio data into a single AudioWorkletProcessor that is responsible for spectral balancing effect you can hear

This project: Automatic spectral balancer effect using AudioWorklet (aka. an interactive mockup of Newfangled EQuivocate dynamic EQ feature)
https://codepen.io/TF3RDL/pen/KwKwbRg
https://hydrogenaud.io/index.php/topic,127438.msg1059410.html

Timestamps:
* 0:00 - Intro
* 0:08 - Unprocessed audio
* 0:40 - End of part 1
* 0:53 - Processed audio
* 1:25 - End of part 2 / outro

Visualized using AudioSurf game (bc I'm tired for an eternal wait for foo_enhanced_spectrum_analyzer 2.0.0.0 release in foobar2000)

Song name: HOPEX - Warrior
https://www.youtube.com/watch?v=CpS52QeZkNY

Caveats:
* On the non-technical side, it still change the perceived volume especially on particular segments of extreme bass boosted songs but still, it doesn't affect the overall dynamic range unlike a multiband or spectral compressor
* On the technical side, currently it somehow downmixes the audio to mono when there is more than one input for this AudioWorklet-based effect

Видео Web Audio API custom DSP: Automatic spectral balancer (@NewfangledAudio EQuivocate dynamic EQ?) канала Andri Gunawan