UPDATE : Source code is moved to a git repository, http://github.com/dilawar/sound . The project README file has the updated instructions. Instructions here may be outdated. This post is for demo purpose only.
This work is derivative of Dr. Taylor’s work http://people.msoe.edu/~taylor/examples/wav.htm . There was a bug in his implementation which is removed from this implementation.
We give here c++ class to read WAV file and store its content in an array. We also store these values in a ASCII file and plot it using
gnuplot to check whether we get it right.
praat can be used to confirm whether our output is right or not. Google to read about the WAV format. We only support non-compressed WAV file.
WAV File Format
We do not support compressed WAV file. Format which is programmed here is available here https://ccrma.stanford.edu/courses/422/projects/WaveFormat/.
The Makefile project can be found at github repository http://github.com/dilawar/sound . After compilation (using
make all), you’ll get the binary
parseSpeech. Use this binary
./parseSpeech -f . If your wav file is long (more than 3-4 seconds) the txt file generated by this program
speechData.txt will be very lengthy. Use can use
gnuplot to draw the waveform. We are sorry for not giving an extensive documentation on this.