How to include a short digital message in a sound recording

Question

I wish to superimpose a short digital message, less than a kB, at an above audible frequency, but still low enough to be transmitted over FM radio stations, and then detect and decode that signal.

What software tools can I use to modulate the message to audio, and how can I go about detecting it. The latter seems simple, a high pass filter, but what software can do that?

ADDED: I want to use the audio signal as the carrier of the digital signal, not a radio transmissions FM carrier signal. This signal should be of such a natutre that it should also be able to be directly transmitted by audio cable.

I only use FM radio as an example medium that the message must survive. I don't want the radio stations to be involved; only the producer of the carrier sound track and the receiver of the message.

Answer 1

You could use a software based RDS radio encoder.

Here is an example for Windows: Airomate RDS Encoder.

Airomate makes it possible to use a 192 kHz sampling rate capable soundcard to generate a high quality stereo MPX signal with full featured RDS/RBDS for FM radio transmitters. Airomate gives your radio station the professional look and sound.

By using a media player on the same computer, the audio-signal will be edited fully digital, so there is no loss in quality produced by A/D converters. The RDS/stereo signal will be calculated digital, so that a perfect RDS/stereo signal will be generated and a good stereo separation will be reached. ( >50 dB)

Answer 2

Since you don't want to use RDS (although it is the most appropriate tool for FM transmissions, but of course wouldn't adapt to other transmission methods), you may be able to use techniques established for audio watermarking. There is spread spectrum watermarking (SSW), which is very robust.

Search also for "audio steganography" and "realtime audio steganography". There's a potentially informative paper here that covers a number of realtime techniques, although you will have to search around to find a copy of it.

There is also steghide, which is an open source (GPL) command line tool. You may be able to adapt the algorithm there for detection in streams, although I don't know what algorithm that uses and I suspect it won't survive the trip.

Many image watermarking algorithms could theoretically be adapted for audio as well, although they may be more obvious in audio form than in image form.

A lot of the simple algorithms operate on digital data and assume no or little loss; so if you find one you'll have to test to see if the encoded data survives a trip through audio processing equipment, a transmitter, air, and a receiver. Algorithms like spread spectrum watermarking are the best bet to survive all the abuse. Start with Google searches for this and this, and related.

There is a library here (via https://stackoverflow.com/a/21269707/616460) as well; might be worth taking a look at.

There is a collection of source code for various watermarking algorithms here (presumably including SSW). Might be geared towards images.

There is a paper specifically about SSW on audio data, Spread-Spectrum Watermarking of Audio Signals.

There is a related IEEE paper, Data embedding in audio signals.

---

Edit: Sorry, I'm used to talking to programmers, I just realized you were looking for tools to do this.

There is a free SSW audio watermarking tool from Microsoft. I have no idea what it is capable of but probably worth checking out. See also this search.

Still, use some of the papers and search terms listed above as starting points for finding tools. I looked around a bit for hardware-based solutions but couldn't find any after a cursory search. There might be something.

Answer 3

What you are trying to do is encode data within the FM band. This is called Radio Data System (or widely known as RDS).

You will need to look for a RDS encoder - you can get hardware and software ones - the software one will require additional hardware (from a quick search)...

Answer 4

There is technology called Radio Data System in place for that and the standard is already defined. From wikipedia:

Radio Data System (RDS) is a communications protocol standard for embedding small amounts of digital information in conventional FM radio broadcasts. RDS standardizes several types of information transmitted, including time, station identification and programme information.

And here is some tech data you can be interested in:

Both [RBDS and RDS standards] carry data at 1,187.5 bits per second on a 57-kHz subcarrier, so there are exactly 48 cycles of subcarrier during every data bit. The RBDS/RDS subcarrier was set to the third harmonic of the 19-kHz FM stereo pilot tone to minimize interference and intermodulation between the data signal, the stereo pilot and the 38-kHz DSB-SC stereo difference signal. (The stereo difference signal extends up to 38 kHz + 15 kHz = 53 kHz, leaving 4 kHz for the lower sideband of the RDS signal.)

Answer 5

I wish to superimpose a short digital message, less than a kB, at an above audible frequency, but still low enough to be transmitted over FM radio stations,

I'm no audio engineer, but ...

"Above audible frequency" will be > 20,000Hz. Anything common and user-accessible designed to transmit/receive audio in physical, analog, hearable form will cut off at 20,000Hz and probably lower. Of course digital mediums and protocols work at much higher frequencies, but none of those are directly connected to a speaker or an audio output, they are transceived to "physical" or analog 20-20KHz audio on both ends to drive a speaker, etc. So I don't think what you want to do is possible, though again I am not sure.

And of course, the higher frequency you go, even if audible, the higher the chance it will be attenuated - if the actual sound medium is an unknown to you. Probably the best "cover-all-cases" frequency range would be those supported by analog POTS, which is only 300 to 3400Hz range. International shortwave's limit is 5000Hz. FM radio's range is up to 15KHz.

If you accept defeat/compromise at this point and want to transmit data over human-hearable audio technology, you have DMTF and of course the plethora of modem standards. A Linux program I've never tried that can decode DMTF and others is multimon. Then there was this project which was a Linux modem implemented in software, that would take incoming streaming audio and decode various modem standards, and obviously encode them too, but I don't think this was updated since 2000.