Re: Measure Frequency
- --- In email@example.com, "Richard Russell" <yahoo@...>
> --- In firstname.lastname@example.org, "rod_bird001" wrote:
> > Keep track of the directional changes and hold an array to record
> > the "frequency" ie time to change direction, and "amplitude" ie
> > how far did it move up or down. Keep track of a second or third
> > level minimum and maximum to record the underlying tones over and
> > above the harmonics.
> In practice it's very hard to determine anything other than the
> fundamental frequency that way. Instead, one conventionally
> performs a Fourier Transform to convert the audio waveform (time
> domain) to its spectrum (frequency domain). Having done that,
> discovering what frequencies are present is simply a case of
> for peaks in the spectrum.the
> At first glance performing a Fourier Transform sounds very
> difficult, and it would be if you were to try to code it in BASIC,
> but fortunately there are DLLs available to do the job. One of
> best known is the Fastest Fourier Transform in the West (FFTW)which
> you can get as a free pre-compiled DLL for Windows:have
> Although I've not personally called FFTW from Liberty Basic, I
> from another dialect of BASIC and it was quite straightforward.Hi Richard!!!
I'm not a programmer but I love all about the computers and,
specially the sound issues, because I'm a music teacher.
I understand your points about the audio representation. In fact, I
worked on it but not with Liberty Basic. So I have a great doubt
about it: all the samples you're talking about are 8 bits per
sample, one audio channel *.wav files, but everybody knows the
standard audio files are 16 bits per sample, two or more channels
(stereo or some surround system). Now, supposing all the people in
the forum knows about how to work with the 16 bits/more than one
audio channel files, my doubt is: IS LIBERTYBASIC ENOUGH FASTER TO
PROCESS THE AUDIO DATA STREAM TO REPRESENT THE GRAPHIC AUDIO WAVE IN
Thanks for your time!!!
GOD BLESS YOU ALL!!!!!!!
Juan J. Paredes G.
From Curicó, Chile, South America, with love
> > In practice it's very hard to determine anything other than theI accept that FFT will provide precision and the .dll link looks
> > fundamental frequency that way. Instead, one conventionally
> > performs a Fourier Transform to convert the audio waveform (time
> > domain) to its spectrum (frequency domain).
useful. However you can still have a lot of fun trying to analyse it
yourself. http://www.lbdownloads.com/files/viewtopic.php?t=84 This
simply tracks the changes of direction and gives you a spectrum
spread of frequency and number of occurrences of that frequency. Full
of bugs and and limited to 8bit samples but it kinda works.
> the samples you're talking about are 8 bits perIt is relatively straight forward to move from 8bit you just need to
> sample, one audio channel *.wav files, but everybody knows the
> standard audio files are 16 bits per sample, two or more channels
> (stereo or some surround system).
deal with bigger numbers in the format that .wav files store stereo
and surround sound. usually each channel is listed in sequence. 16bit
just uses two bytes instead of one so you need to uprate the analysis
Real time? well that would require a real time audio stream. What we
are talking about so far is analysis of recorded .wav files.
Anyways the point I make is it isn't impossible its just numbers and
Liberty BASIC is vastly adept at handling numbers.