The International Obfuscated C Code Contest
A 27th IOCCC Winner
The code for this entry can be found in prog.c
./prog < pi.wav # [https://en.wikipedia.org/wiki/867-5309/Jenny] ./prog 867-5309 > jenny.wav # Use an audio player to play jenny.wav
A cross platform utility for both encoding and decoding tones.
This program encodes and decodes dual-tone multiple frequency signals used in telephone systems, more commonly known as Touch-Tones.
If the program is executed with no arguments, it will read a WAV file from standard input, decode the touch-tones in the file, and output the corresponding digits to standard output. This program only supports WAV files that have exactly 16 bits per sample, but it allows any sample rate and any number of audio channels.
$ ./prog < pi.wav 31415926
If the program is executed with a command-line argument, it will generate the tones corresponding to the specified characters, writing them to standard output as a WAV file.
$ ./prog 867-5309 | aplay
gcc -pedantic -Wall -Wextrathere are no compiler warnings.
memcpyso there are no strict-aliasing violations.
CHAR_BIT == 8
doubleis an IEEE 754 binary64 floating-point type
sizeof(double) == 8
aais both a variable and a macro for error handling, and
memcpyis both a library function and a goto-label.
fmacro is used to unroll some of the loops.
l. This array is usually accessed at index
zplus a constant. Accesses of this type have been rewritten from, e.g.
[l+z]ends up occurring quite often, so I created the macro
aas a shorthand.
Each tone in a DTMF signal is a combination of two frequencies. The lower frequency determines which row the digit is in, and the higher frequency determines the column. There are 4 possible row frequencies and 4 possible column frequencies.
|1209 Hz||1336 Hz||1477 Hz||1633 Hz|
|697 Hz||1||2 abc||3 def||A|
|770 Hz||4 ghi||5 jkl||6 mno||B|
|852 Hz||7 pqrs||8 tuv||9 wxyz||C|
This program determines which frequencies are present in the input by passing it through a set of 8 virtual resonators, each tuned to one of the frequencies used for DTMF. These can be implemented using only basic arithmetic, and they are able to ‘select’ a specific frequency from the input sound, blocking all other frequencies from reaching the output. The loudness of the sound output from each resonator can therefore be used as a measure of how much of a particular frequency was present in the input. The program then decides, over the course of one tone, which row frequency and which column frequency were most prominent.
Tones are generated using those same resonators by providing an impulse as input. An impulse can be thought of as consisting of sine-waves at all frequencies, and when it is put through one of the resonators, the result is a single pure sine wave.
© Copyright 1984-2020,
Leo Broukhis, Simon Cooper, Landon Curt Noll
- All rights reserved