Based on 15 visible X reactions from 17 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Many other interesting applications: - Spatial audio in crowded public spaces - "hearing aids" - Design objective for quieter machines - Acoustic levitation?? Generally speaking, I'm excited to master sound waves as a civilization https://x.com/mathemagic1an/status/2076319950458024166/video/1
Interactive exhibit and code here: https://jayhack.github.io/noise-cancellation/ If you are interested in building this - please reach out, I would love to fund a team to pursue this 🚀
@mathemagic1an people have worked on this for the last 40 years. guessing you didn't run this by any acousticians prior to posting? 🤣
@mathemagic1an Well that's completely insane, do you think it's feasible from a cost perspective?
Many other interesting applications: - Spatial audio in crowded public spaces - "hearing aids" - Design objective for quieter machines - Acoustic levitation?? Generally speaking, I'm excited to master sound waves as a civilization https://x.com/mathemagic1an/status/2076319950458024166/video/1
Interactive exhibit and code here: https://jayhack.github.io/noise-cancellation/ If you are interested in building this - please reach out, I would love to fund a team to pursue this 🚀
@mathemagic1an people have worked on this for the last 40 years. guessing you didn't run this by any acousticians prior to posting? 🤣
@mathemagic1an Read Schopenhauer on noise. Great idea.
@mathemagic1an hell yeah bro made the code of silence
What if we could cancel out urban noise pollution? Place a set of speakers that act like noise-cancelling airpods for all pedestrians simultaneously You can actually do this - and it paves the way for an "acoustics foundation model" Simulation code + proposal 👇 https://x.com/mathemagic1an/status/2076319574388269564/video/1
Real world speakers are discrete - perfect cancellation w/ finite speakers is impossible You can do surprising well with discrete speakers, however, if you focus on minimizing sound only *around humans* Simulation below. See the "shadow" following the human around. https://x.com/mathemagic1an/status/2076319708870307982/video/1
Concretely: Consume live video and initial set of frequency response measurements for a scene Infer realtime transfer matrix for sound waves Place N mics, M speakers, infer audio sources (cars) + ports (humans), and run a quick optimization for live noise cancellation https://x.com/mathemagic1an/status/2076319870795555309/photo/1
This also extends to scenes with obstacles You can estimate relevant parameters - the "transfer matrix" - via a small set of localized sound observations In practice: sufficient to get a guy to walk around the construction site with his iphone out https://x.com/mathemagic1an/status/2076319841322201499/video/1
The Kirchoff-Helmholz integral theorem says perfect noise cancellation is in theory possible You can eliminate an arbitrary wave source - but it requires a continuous "shell" of cancellation speakers https://x.com/mathemagic1an/status/2076319641329336790/video/1
This works well even for multiple pedestrians You can run a fast, exact solver for N speakers and M pedestrians, given their position (And pedestrian tracking is now a solved problem 🙌) https://x.com/mathemagic1an/status/2076319774066516424/video/1
Based on 15 visible X reactions from 17 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@mathemagic1an hell yeah bro made the code of silence
What if we could cancel out urban noise pollution? Place a set of speakers that act like noise-cancelling airpods for all pedestrians simultaneously You can actually do this - and it paves the way for an "acoustics foundation model" Simulation code + proposal 👇 https://x.com/mathemagic1an/status/2076319574388269564/video/1
Real world speakers are discrete - perfect cancellation w/ finite speakers is impossible You can do surprising well with discrete speakers, however, if you focus on minimizing sound only *around humans* Simulation below. See the "shadow" following the human around. https://x.com/mathemagic1an/status/2076319708870307982/video/1
Concretely: Consume live video and initial set of frequency response measurements for a scene Infer realtime transfer matrix for sound waves Place N mics, M speakers, infer audio sources (cars) + ports (humans), and run a quick optimization for live noise cancellation https://x.com/mathemagic1an/status/2076319870795555309/photo/1
This also extends to scenes with obstacles You can estimate relevant parameters - the "transfer matrix" - via a small set of localized sound observations In practice: sufficient to get a guy to walk around the construction site with his iphone out https://x.com/mathemagic1an/status/2076319841322201499/video/1
The Kirchoff-Helmholz integral theorem says perfect noise cancellation is in theory possible You can eliminate an arbitrary wave source - but it requires a continuous "shell" of cancellation speakers https://x.com/mathemagic1an/status/2076319641329336790/video/1