Hello

I'm very new to PD, but I used MAX/MSP a few years ago for a small project. Now I doing another small project and I think I might be able to use PD for this.

Here is roughly what I need to do
1. Load and play a wave file (shouldn't be too hard
-> This is a normal audio wave-file but it has a special signal embedded in it. The signal is really at tone that plays at a specific frequency (19.200 Hz).
2. I need to detect when this "tone" is on and when it is not. A "bang" when it is on would be perfect.
3. I will use the bang to set a data-pin on the parallel port to high. I don't expect this to be a big problem either. I think it has been done before? If I somehow can write directly to a memory address I'm sure it can be done.

So the big problem is step 2. Maybe FFT could be used for this? I think FFT looks very complicated and I have no idea where to start. Can somebody help me please? Please remember that I'm very new to this program and I don't know a lot about audio-signals, fft...

I should also mention that:

1. I expect the 19.200 Hz tone to appear very frequently - sometimes many times per second
2. I may sometimes only appear in the left or right side of the stereo-signal.
   
   Any help would be greatly appreciated

hey kristian--

for 1: you're right, shouldn't be hard...just use tabread4~ or sfplay~ or whatever
for 2: try the fiddle~ and bonk~ externs...they are the most popular beat/pitch recognition externs around. they should do the trick. if the 19.2 Hz tone is the only sound in the wave file, you could probably just use an envelope follower (env~). if there are other sounds it gets trickier. i'd filter out all of the audio above a certain cutoff (low pass filter (lop~) or you could use a bandpass filter (bp~ or vcf~) for a more selective filter).
for 3: i know there is either native or external support for serial in/out in Pd...not sure about parallel. i am sure someone has written an extern for parallel i/o but you'll have to track it down.

sounds like an interesting project. if you don't mind my asking, what is generating the 19.2 Hz tones and what is hooked up to the parallel port?

--zac

I'm trying to make a free alternative to the decoder found here: decoder - It is too expensive for me and for a lot of other people, I think. I'm not trying to put them out of business, but to make an alternative for people who are willing to mess around with electronics and who can't afford a readymade decoder and glasses.

for 2: I have already had a look at the fiddle~ and bonk~ externs. They look promising, but I don't understand them well enough to actually use them.
The audio contains other sounds + the 19,2 KHz signal, so I guess the envelope follower is out.
I hadn't thought about filtering the signal, as you suggest. Maybe I can use that in conjunction with fiddle~ ?!

i just did some testing with a [bp~ 19.2 5] object (bandpass filter with center frequency of 19.2 Hz, Q or selectivity of 5). through the filter i mixed some ordinary music (pretty wide range of frequencies present) and somewhat random 19.2 Hz sine wave pulse. i when i wrote the output of that to a table, it was clear to see on the graph when the 19.2 Hz pulse was on or off. you could probably use an envelope filter to track this. fiddle~ would probably work as well if you're willing to take the time to learn how it works and get all of your parameters straightened out.

hmmm...i just read over your post and am now realizing the signal is '19 thousand 200 Hz'. these american/british/continental conventions can make things confusing. so, in anything i have written, i was thinking we were talking about a low frequency (19.2 Hz) but we are actually talking about a high frequency. i suppose it doesn't change much--just change my low-pass to high-pass and 19.2 Hz to 19200 Hz.

sorry about the confusion.
-z-ac

Yes I'm sorry about the 19.2/19,2 Hz/kHz confusion. You are right it's 19.2 KHz

I've have been looking at the bp. My start out point was the example file called "H02.bandpass.pd" which is located in doc/3.audio.examples

However it seems to work better with low frequencies. Maybe the setup you have been using is better than mine. If you still have the .pd file, I would very much like to see it. Maybe you could send it to me? My address is: gandalf@beatnik.dk

Thank's for your time!

the patch i was doing some tests with was designed for 19.2 (19,2) Hz. i didn't test 19.2 KHz. my guess is that it will be trickier to filter out high frequency signals because you are bound to get interference from the rest of the audio on the file (especially if it's music). i am away from my Pd computer this weekend, so i can't really test that.
--zac