A reminder to my future self (time shifted and added to my original self)

Over a year ago I was probably at the top of my game in terms of audio k-nowledge. I had gotten through the weed-out and fluff courses and started taking the harder-yet-interesting audio and signal processing classes. Anyway, I remember thinking that I should post a picture as a reminder to my future self that at one point I used to know stuff. What you see above is a block diagram describing a mathematical algorithm for an audio process, called an all-pass filter. An all-pass filter, when applied to a sound, will not make an audible difference. In other words, you listen to it before and after you apply the filter, and it sounds the same.

What's cool about it is that it works "behind the scenes". It changes the time information of the sound for certain frequencies. By pushing and pulling the time information for the sound and adding it to the original sound, certain frequencies will amplify or get cancelled out. Kind of like how sometimes when waves in the ocean hit each other, depending on when whey crash into each other the wave will get twice as big or will cancel itself out. The below diagram takes the sound, changes some of the time information for some of the frequencies, then adds it back to itself.

You end up getting a boost or reduction in sound at a given frequency:

It's like magic. Crazy stuff. Shout out to partner-in-crime, Jason Laska who is holding it down at Rice university, inventing crazy stuff that will soon make him famous.

( categories: )