Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Ggwave: Message people over sound (github.com/ggerganov)
136 points by BasedDebianUsr on May 1, 2023 | hide | past | favorite | 46 comments


This (FSK + Reed Solomon for data over sound) is exactly what my team and I built at Chirp in 2012. We had some success as an app/SDK. If you Google around you can find some implementation details.

The hard/er stuff was making it nice for humans, robust to echo/noise, work in ultrasound, and functional on low-spec devices

https://web.archive.org/web/20131010174943/http://chirp.io/

I'm very glad to see people still thinking about it!


I remember Chirp! It was a big influence for the work we did at the time for a project called "Add-on" which we tried to pitch to TV advertisers at the time.

The idea was to embed sounds during ad-breaks that people could opt-in too.

Didn't go anywhere, but our initial PoC code is still around https://github.com/pebblecode/adOn-soundlib


Fellow chirp-er here, we used it at school to chit chat in a group of friends. The fact that it was 'broadcasted' in some sense was an interesting dynamic. Thanks for the fun times!


It would be cool if someone could take this and change the alphabet a little bit to be more pleasing for humans as well. If there was something like this but that sounded like Star Wars droids for example, then that would be very cool.

In fact, I can totally imagine a world where droids speak to each other through the air in a way that well-trained humans could make out - maybe they're even mandated to do so as part of their protocol - but then they have like a data com for binary data.


Related:

Show HN: GGWave – Data over Sound for Microcontrollers - https://news.ycombinator.com/item?id=31987674 - July 2022 (1 comment)

Show HN: I made a tool to communicate data using the PC speaker - https://news.ycombinator.com/item?id=26892930 - April 2021 (71 comments)

Ggwave: Tiny Data-over-Sound Library - https://news.ycombinator.com/item?id=26119109 - Feb 2021 (27 comments)

Show HN: Waver – Messaging Through Sound - https://news.ycombinator.com/item?id=25761010 - Jan 2021 (11 comments)


I shipped a small ultrasound prototype library to the Uber rider and driver apps. A 4 digit code would get sent to the driver app, and the rider app, and one side would chirp while the other would listen and confirm it. Ostensibly, this meant you were in the right vehicle. It was a cool toy but I left it at that. IIUC Uber put more work into it and made it robust enough to ship in prod. Shoutout to OC.


Normally when I get into the wrong car it inevitably is a non taxi and they ask wtf I think I am doing faster than phones could fail to do the ultrasonic handshake and let me know I had made a mistake.


It's good for the driver to confirm they have the right riders. And for riders to know that the driver is not just pretending to be an Uber driver in order to drive away and rob the rider.


If only we had human-readable identifiers on each and every car... Maybe we could put these identifiers on the outside surface of the car, so that the user can verify their identity without getting into the car and risk getting kidnapped.

Any VCs here want to fund this idea??


They’d need to be regulated by some central governing agency though…


We'll put it on The Blockchain!


And then some US states could decide whether to require placement of the Blockchain ID on the front bumper.

(Ah yes, states, the laboratories of democracy in our federalist system.)

> The phrase "laboratories of democracy" is often attributed to Justice Louis D. Brandeis, who served on the United States Supreme Court from 1916 to 1939. In a dissenting opinion in the case of New State Ice Co. v. Liebmann (1932), Brandeis wrote:

> "It is one of the happy incidents of the federal system that a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country."

(the previous two paragraphs via ChatGPT 3.5., verified by me)

I would like to add the preceding sentence in the above quote from Brandeis’ dissent because it gives more context:

> “To stay experimentation in things social and economic is a grave responsibility. Denial of the right to experiment may be fraught with serious consequences to the nation.”


Yes. New Jersey has embarked on the courageous step of allowing shell companies to issue temporary tags, allowing criminals, cops and others to evade tolls, automated enforcement and pervasive LPR surveillance in New York.


One day I was walking down the boulevard (in a tie-dye rainbow shirt - perhaps this had something to do with it) and I only had two blocks to go to the post office. Homeboy comes to a stop in an old and busted car, and he hollers out "Hey, I'm ... like, a Uber... you wanna ride somewhere?" and I had a lasting creepy feeling that that would happen in broad daylight. I mean, I'm sure homeboy just wanted me to hand over a few bucks in cash for an easy ride, but imagine the suckers and predators that could be brought together by fake Uber.


Depends on how busy things are where you are, and also how popular silver Priuses are there.


> one side would chirp while the other would listen and confirm it. Ostensibly, this meant you were in the right vehicle.

In Finland this is solved by giving the person reserving the taxi a two-digit code that you repeat to the driver to confirm that you were the requester. It works great unless you're so loaded you can't remember a two-digit number.


This is how HomePod pairing works, too!


One of Greg's many excellent works. https://news.ycombinator.com/from?site=github.com/ggerganov

Seen today, the Whisper.cpp 1.4.0 release. https://news.ycombinator.com/item?id=35765882


Haven’t people been messaging other people over sound for tens of thousands of years?


I don't think that's quite right.

Homo Sapiens first emerged 300,000 years ago, and since other animals communicate via sound, it's at least that long. If you want to define "people" as encompassing all of the genus Homo though, then you're looking at more like 3.3 million years.

Just depends how you want to define "people", really.


Doesn't that mean what they said is still right?


What insight are you seeking to learn or share here?

Do you not see nor understand the differences between animal vocalizations, human speech, and this?

I hope the above isn’t just a throwaway or snarky remark. That’s why I’m asking these questions.


I once wrote a code for transferring data from one computer to another using computer using VGA cable just for fun. That was a great learning opportunity for me to understand network protocols and error correction. I did not go through the code to understand the kind of protocol this piece is using but I see that it feels like a fun project. I am tempted to do this using "light"


Lots of examples being added in the comments, so I'll mention Google Tone. https://chrome.google.com/webstore/detail/google-tone/nnckeh...


I am probably missing the point of this. Didn’t we have stuff like this back in the 80’s with phone couplers and modems?


An acoustic coupler is made for coupling a modem to a phone microphone. A modem is made for harnessing the phone line directly. Both are for creating long running connections. This is made for short messages sent thru the air. The most obvious novel use for this would would be IoT without a network.


The term "air-gapped system" has just lost its meaning.


this is so cool but the headline, "message people over sound," made me feel like I was living in a world where verbal speech was a lost art (perhaps like cursive handwriting?)

Anyway, not sure where I'm going with this post, other than to say OP has big 2023 vibes


Compare Asimov's story "The Feeling of Power":

https://en.wikipedia.org/wiki/The_Feeling_of_Power#Plot_summ...


Interesting association, thanks!

We're pretty close to that these days -- I am old and grew up in analog times, and used to be pretty handy at long division. Had to do a bill the other day, and for old times' sake, I busted out the pencil.

It did not go well.

But in cheerier news, I've successfully integrated copilot into my neovim install!


Might write a pitch about reimagining and disrupting verbal communication by inventing talking.


[pulls down medical chart of human larynx]

This organ is capable of high-bandwidth sonic transmission. By modulating the flow of air over the meat-baffles, it is capable of producing a wide range of distinct


This is a old and very known technology: https://www.google.com/search?q=data+over+sound&oq=data+over...

I once worked with a company that have a whole business around this technology (for some reason their website isn't working): https://il.linkedin.com/company/sonarax


I was sure this was a thing on a browser not long ago ... 2015 chrome is what I was thinking of - a severless send to phone method I guess https://www.independent.co.uk/tech/google-tone-chrome-extens...


Yeah, it's called a dial up modem ;-)


How does this stack up against apt install minimodem? That has been my go-to sound data transfer method. Is this faster aka more robust or just a reimplementation?


Rattlegram is an app i've used with Ham for some time to send audio based messages over analog audio. It also works well in air-gapped noisy rooms too, due to the excellent efficiency of COFDM (coded orthogonal frequency division multiplexing). Of course not many situations are actually like that, so more of a fun experiment, but the Ham use is actually pretty useful.


Anyone has any idea of the bandwidth you can achieve with that ? Could I wrap this in a TLS socket and stream a video to friends nearby ?


Finally, the droids can talk to each other!


Could this enable embedded as metadata in audio files? Ie: could this be similar to what Spotify might use to fingerprint the music they distribute?



I don't mean to come across as insulting, rather I am genuinely curious to understand why this was done? what was the inspiration?


> why this was done?

Because they could.


I do this all the time!!!

I yell out "Hey, What's up?"

And the person who hears me responds!!!

Maybe I should patent that!!!


I mean seriously, when you start talking about "Messaging people over sound", you really need to get up from the computer and go outside...


now all those beeps and tweets in the star wars films make sense!




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

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

Search: