I'm wondering if there is any protocol which would allow to encode data through audio in a human tolerable kind of way ?

An example would be the LTC protocol which is not at all human tolerable (my left ear remember) but it encode timecode information in an audio signal for synchronisation purposes.

The application could be to transmit an URL through a radio show or a podcast without having to click the description.

Of course it would require a dedicated decoder on the listener device (like for QR code which is mainly integrated in the device camera these days).


Another application would be to carry information along the signal like watermark do for a video.

  • 2
    What would be the benefit? Radio stations can already transmit a large amount of non-music data alongside the audio. DSL used to do something similar by using the analog audio bandwidth above that of the usable human voice range, using a splitter. That won't work for digital because of Nyquist/Shannon & because you can't lop the top off music like you can the human voice.
    – Tetsujin
    Commented Apr 18, 2022 at 8:54
  • The idea wouldn't be to hide the audio signal encoding data. QR code are visible to the human eyes. The benefit would be that the audio would self contains the data. I'm not sure if there is industrial application to this but at least on an artistic point of view it could be nice to have a double content to an audio signal (one esthetic one and another informational one). Like QR code art. Commented Apr 19, 2022 at 13:13
  • 1
    In some TV channels I hear a few tones right when the commercials begin. Is it DTMF? OTOH, just going with the question, QR code allow the transfer of information from one device (or medium) to another by optical means. In audio the equivalent could be, for example, your phone listening for a URL "played" on your car's radio... Is that what you mean?
    – Jahaziel
    Commented Apr 20, 2022 at 0:16
  • @Jahaziel yes. Like a QR code I could accept a certain level of degradation. Commented Apr 22, 2022 at 14:29
  • Would kind of be / is morse code, just even simpler, wouldn't it?! Furthermore you could always build a software that takes tone/tone area A for 'Black' and B (or a pause) for 'white' and simply transmits a standard qr code (whatever it contains) that way. But i guess you're asking for a standard with already existing ''decryption'' apps like common qr scanners for visual qr codes?! Commented May 1, 2022 at 21:02

3 Answers 3


Dual-tone multi-frequency signaling (DTMF) seems to be the obvious answer, because it also has two dimensions (two sets of tones).

To make the tones more tolerable, you could move them to a lower frequency.

To transmit an URL you would need to furter encode your data, like e.g. with an evolved form of something like a BASE64 encoding.

However, while all this would work, it sounds quite involved and probably is not the solution to your actual problem anyway.


A similar idea (sometimes called “data over sound”) was the Chirp technology. A mobile app was provided which could send and receive data via audio (speaker/microphone). With a data rate significantly slower[1] than bluetooth etc., the use case indeed seems to be in a similar area as QR codes. Human tolerability isn’t explicitly mentioned, but it should be part of the specs given the intended use. It uses sound frequencies between 20 Hz and 20 kHz which is the typically cited audible band. Chirp itself seems to have been discontinued, but there is an open source (MIT) project that implements a similar (not compatible) protocol called “audio-barcode” (I’m not affiliated in any way).

[1]: “about 20 bits/second up to 1K or 2K bits/second, depending on ambient noise and other factors” according to Jim Turley: “Chirp Does Data Over Sound”. 2019. EEJournal.


This type of technology is used for audience measurement of TV channels by Médiamétrie. https://www.mediametrie.fr/en/tv-audience-measurement-tagging-signal-used-channels-key-link-chain

Médiamétrie has used watermarking as a television audience measurement technology since 2008. This technology involves the insertion of an inaudible tag – a tattoo – into the audio stream broadcast by television channels; this tag carries a unique identifier specific to each channel, as well as the time and date of the programme broadcast live or viewed as catch-up.

The tag is recognised by two types of audience meters. The first - a fixed audience meter - is a tablet installed in panellist households beside each measured television set. The second is a small smart meter worn by the panellists as a wrist bracelet or belt clip whilst they are at home, away from home or on the move. This equipment allows Médiamétrie to identify which channel the panellist is watching - key data for calculating audience figures. Médiamétrie is recognised worldwide for its innovations in audience measurement.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.