Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: Olaf – Acoustic Fingerprinting on the ESP32 and in the Browser (0110.be)
166 points by joren- 9 months ago | hide | past | favorite | 30 comments

ESP32 is a source of infinite frivolity. My latest project does exactly nothing, except when certain door is opened. Then it logs into Wifi and continues to do nothing. Except now my desktop computer can ping that address and inform me the bloody door is open.

What’s funny is that it’s still probably cheaper to do it this way than with wireless reed switches, which often cost more but are much simpler??

Maybe for other people, not for me.

My free time is very expensive.

An ESP32 costs nothing compared to the cost of my free time, and comes with wifi and bluetooth, which makes many things simpler.

Other solutions whose hardware costs less than an ESP32 cost me a lot of time, and end up being infinitely more expensive than the ESP32 approach.

Which power source do you use for this project?

Tiny Lithium. A diode drops the voltage from 4.2 to 3.5 . Mechanical switch connects the power when the door is opened.

I see a very zealous and illegal, albeit useful application: I could hide an ESP32 with a relay inside of a circuit breaker and install it in my neighbor's breaker box, cutting their power every time they start blasting obnoxious disco music at 3 AM.

ahhh, evilMe likes this. likes this a lot!! except, you better train it well to differentiate the specific music you don't like lest your own partying have negative consequences. (which don't all evilMe type plans usually fail from some whimsical reason?)

You’d probably make a lot of money if you had something that could tell what music is obnoxious (and what is not)

All music is at 3AM. We're halfway there!

Wo-ah! Livin' on a prayer!

Would be cool to do that a rave.

I'm constantly impressed by the ESP32 and have dabbled in a few project ideas.

You could even use them to talk to other ESP32 devices (in other dresses!) using WiFi/UDP, Bluetooth LE or Espressif's own 'ESP-Now' mesh networking (which recently blew my mind).

I've done a few RGB LED Matrix based ESP32 projects now. This project has given me a few ideas.


He should get in touch with Disney’s merchandising department and send them a prototype. Nothing would stop them from just copying the idea without compensation, but doesn’t seem like much downside to trying since he made it for fun for his daughter.

Or, because they're Disney, will respond to the author with a Cease and Desist order because of some sort of copyright violation.

being Disney, they would sue him, regardless what

Yeah, Disney sued my Dad, a small jewelry store owner, for having a charm, which he purchased from a supplier, that looked too much like Bambi. It was just a little silver deer charm he bought from a wholesale supplier. They sent someone to his store, and sued him over one charm.

I am so happy for your daughter. Such great memories she will have of her father. Thanks for sharing, this gave me an idea...

How well would this work towards voice activated things, i.e. "Hey Olaf, make it rain?" which turns the sprinklers on.

These type of algorithms are only able to match audio when the audio was added to an index beforehand. So it only capable to match pre-recorded audio. A live rendition of a text or song would not work since it would not match the index.

That would be the domain of mycroft.ai

ESP32 has a variant which is compatible with ESP-ADF (audio development framework)

I haven't checked them in a while but they could've do on-device activation and text conversion before sending to Alexa API.

The problems were the docs are mostly chineese and its in early stages.

It would work if you can sing out the phrase ;). Basically a real life version of a musical.

What is the demo supposed to do when it positively recognizes the song?

Definitely didn't just sing Let It Go :) Awesome stuff!

Have you definitely not just sung it in ancient greek yet?


Very cool! What's the license for Olaf?

Don't forget about patents, from their README:

> Please be aware of the patents US7627477 B2 and US6990453 and perhaps others. They describe techniques used in algorithms implemented within Olaf. These patents limit the use of Olaf under various conditions and for several regions. Please make sure to consult your intellectual property rights specialist if you are in doubt about these restrictions. If these restrictions apply, please respect the patent holders rights. The main aim of Olaf is to serve as a learning platform on efficient (embedded) acoustic fingerprinting algorithms.

Very cool project, thanks for sharing.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact