[HN Gopher] Show HN: Waver - Messaging Through Sound
___________________________________________________________________
Show HN: Waver - Messaging Through Sound
Author : ggerganov
Score : 37 points
Date : 2021-01-13 12:04 UTC (10 hours ago)
(HTM) web link (github.com)
(TXT) w3m dump (github.com)
| cgsmith wrote:
| This could have multiple uses. One that I can think of is to have
| a profile of what a normal sounding manufacturing floor is or
| lathe, mill, any machine that operates. If the sound profile
| changes you might be able to distinguish what component is
| failing on that machinery. This could possibly be extended to
| listening to organisms or ecosystems like the human biome to
| identify problems.
|
| I know other tech might exist in this but I find the possible
| uses for this fascinating.
| codebeaker wrote:
| There's quite some research, and even some products (I mentored
| a company about five years ago doing such a thing, but I even
| forgot their name, and don't remember what their outcome was)
| about this topic.
|
| - https://www.hindawi.com/journals/js/2018/4105208/
|
| If memory serves, you don't even need to have permanently
| installed sensors, you can do it with a portable stethoscope
| kind of contraption which can detect anomalies and help to
| schedule preventative maintenance.
| skibz wrote:
| This is a super impressive collection of work. Well done.
|
| The only other tool that I've ever seen similar to what you've
| built is libquiet. Are you familiar with it? If so, could you
| provide a comparison?
| ggerganov wrote:
| Thanks. Yes - I am familiar with libquiet. It is much more
| advanced and mature compared to what I am doing. I haven't used
| it for development, but have played with it's online examples.
| I think my library is sort of a poor man's solution to what
| libquiet offers, with the advantage of being a very small and
| simple piece of software.
| ziffusion wrote:
| You mean speaking?
| dehumanise wrote:
| i knew it sounded familiar
| ggerganov wrote:
| Heh, kind of - but in a weird sounding language :)
| ggerganov wrote:
| Hello HN,
|
| This is a hobby project that I have been working on and off for
| the past few years. It started with a simple PoC [0] for sharing
| files through WebRTC using sound signaling to exchange the SDP
| between peers and I recently decided to extract the data-over-
| sound part into a standalone library [1] so I can use it in
| different projects. The current post is one such project: "Waver"
| - a simple application with the primary purpose to showcase the
| capabilities of my data-over-sound library. The main
| functionality of the app is to exchange short text messages
| through sound with nearby devices. All you need to use it is to
| enable the device's speakers and microphone. The app also
| includes an early prototype for sharing files with sound
| signaling. The app runs on iOS, Android, Linux, Mac OS and Web -
| I put quite some effort to make it easy to try out :)
|
| The implemented data-over-sound protocol is quite simple. It is
| essentially a 6-tone FSK modulation scheme with special markers
| to mark the start and end of the transmission. The bandwidth is
| quite low (8-16 B/s), but I think the transmission is reliable.
| I've been testing it with a few devices across the room and it
| seems robust enough. The ultrasound protocols especially are very
| robust to surrounding noise.
|
| An additional challenge I had to solve while working on the data-
| over-sound protocol was to make the encoded data sound good to
| the human ear. Although the current sound is not the most
| pleasant thing to hear (especially the markers), I think it
| turned out relatively OK :-)
|
| There are a few reasons about why I decided to implement this FSK
| approach instead of using an existing and well-known modulation
| scheme used by modems:
|
| - My DSP knowledge is fairly limited
|
| - Most modulations typically sound like noise (not very pleasant
| to the ear)
|
| - Reliable air-gapped transmission seems to be much more
| difficult to achieve compared to over-the-wire
|
| - But mainly because I find it more fun to implement something
| from scratch
|
| There are similar solutions available on the market, but I feel
| they are difficult to try out. I wanted to make something free
| and very easy to test and see how it works. The data-over-sound
| library [1] still needs some work, but it's getting there. I am
| hoping to find some fun and potentially useful applications of
| this, so I am looking for ideas. Also, any feedback about the
| performance is very much appreciated.
|
| If you decide to try out the Waver app, make sure to watch the 2
| min video in the link [2] for a quick tutorial. Also, there is a
| small easter egg at the end of the video :-)
|
| [0] - https://github.com/ggerganov/wave-share
|
| [1] - https://github.com/ggerganov/ggwave
|
| [2] - https://www.youtube.com/watch?v=Zcgf77T71QM
| sidpatil wrote:
| Cool product! This is essentially a modern take on the
| radioteletype [0], which is much older than dial-up modem
| technology. You might find it interesting to study RTTY
| modulation schemes if you haven't already done so. (If you do
| end up implementing RTTY modulation schemes, it could make the
| product more popular among amateur radio enthusiasts, who use
| similar software already [1].)
|
| [0] https://en.wikipedia.org/wiki/Radioteletype
|
| [1] https://en.wikipedia.org/wiki/Fldigi
| ggerganov wrote:
| Thanks - I should definitely take a more detailed look at
| existing RTTY modulation. I admit I haven't looked too much
| into it. I know I am probably "rediscovering the hot water"
| for most part :)
___________________________________________________________________
(page generated 2021-01-13 23:03 UTC)