IF the amplitude of the input signal is stronger between the range of 200hz to 400hz than the amplitude
of the same signal in the range between 1200hz to 1800hz THEN light a green LED.
IF the amplitude of the input signal is stronger between the range of 1200hz to 1800hz than the
amplitude of the same signal in the range between 200hz to 400hz THEN light a red LED.
I've used the basic stamp2 with the Nuts and Volts articles successfully with it's coding for a variety of
things over the years. Needless to say I'm quite loyal to Parallax because they've helped be get my brain
somewhat around micro-controllers. I'll gladly expend the effort to learn and adapt this if you think I'm on
the right track
That is sad news and thank you for letting me know, Seems like an amazing person.
Are there other people on the forum who understand how to analyze and respond to audio data once it is converted to digital?
That is sad news and thank you for letting me know, Seems like an amazing person.
Are there other people on the forum who understand how to analyze and respond to audio data once it is converted to digital?
Phil's FIR tool looks very useful. Does anyone have a copy of his perl script that does all the work, or know how to obtain it, since we don't know how much longer Phil's website will be up for?
IF the amplitude of the input signal is stronger between the range of 200hz to 400hz than the amplitude
of the same signal in the range between 1200hz to 1800hz THEN light a green LED.
IF the amplitude of the input signal is stronger between the range of 1200hz to 1800hz than the
amplitude of the same signal in the range between 200hz to 400hz THEN light a red LED.
I've used the basic stamp2 with the Nuts and Volts articles successfully with it's coding for a variety of
things over the years. Needless to say I'm quite loyal to Parallax because they've helped be get my brain
somewhat around micro-controllers. I'll gladly expend the effort to learn and adapt this if you think I'm on
the right track
Rich
Hi, once again,
have been thinking about your post for a while and what would be a good fit to your task and also to what you wrote about you knowledge. So if you want to dig into digital sound processing the Goertzel algorithm is certainly something to consider. In principle, I have done something like your task using P2 and Taqoz Forth. https://forums.parallax.com/discussion/173880/picture-of-sound-parallax-propeller-p2-as-a-tool-to-analyse-a-guitar-effect-pedal-or-an-amp#latest
Is is done in two steps: First get a buffer of audio samples and then analyse it.
Using a microcontroller shines, if you have to vary things like in my project the frequencies.
On the other hand, if the frequencies are fixed and especially if there are only these two frequencies present in the signal it will be very much easier, cheaper and faster to do at least the filtering and the conversion to envelopes with analog electronics. You would need a preamplifier with a band filter 200...1800Hz and low impedance output (one opamp) then in parallel a RC 700Hz high pass filter and a low pass filter and for each of them a simple rectifier to get the amplitudes. At the end a comparator (a second opamp) to compare the amplitudes.
Christof
I looked at your picture of sound project and I must say it's impressive and over my head. I am greatful for your responses to my post so I want to explain more as well as upload my schematic at present. a microphone is recieving a full frequency signal and trying to ignore the frequencies that are present outside of the desired range and respond to a comparison between the aplitude of the desired frequencies. It's an industrial floor test device that makes a scraping noise on the surface of the floor and listens for voids. The LM567 chip you see in my schematic is almost suitable for the task, but because I enjoyed the basic stamp2 so much years ago for other prototypes this seems fascinating as an option. I read the Goertzel description and the math fried my brain. Yet.. since the audio output never even has to be considered only the driving of and indictor light. Maybe it is still possible
I looked at your picture of sound project and I must say it's impressive and over my head. I am greatful for your responses to my post so I want to explain more as well as upload my schematic at present. a microphone is recieving a full frequency signal and trying to ignore the frequencies that are present outside of the desired range and respond to a comparison between the aplitude of the desired frequencies. It's an industrial floor test device that makes a scraping noise on the surface of the floor and listens for voids. The LM567 chip you see in my schematic is almost suitable for the task, but because I enjoyed the basic stamp2 so much years ago for other prototypes this seems fascinating as an option. I read the Goertzel description and the math fried my brain. Yet.. since the audio output never even has to be considered only the driving of and indictor light. Maybe it is still possible
Hi Rich,
I am not familiar with LM567. As far as I can see, the circuit detects one frequency in a rather small band. No comparison with a second frequency. I tried to make a sketch of my idea, which compares the strength of 2 frequency bands.
Do you use LtSpice? It is an extremely powerful tool and free.https://analog.com/en/design-center/design-tools-and-calculators/ltspice-simulator.html
I am sorry about this wikipedia article. I have difficulties understanding the math too. The pseudo code part gives the application and is rather simple.
You would need >=4000 samples/sec.
If you want to do sound processing with software you would need a processor, which can multiply relatively fast. I assume that Parallax P1 can do it, P2 will work for sure. But I don't think a stamp can do it.
As Parallax does not have anything comparable, I dare to recommend: https://pjrc.com/teensy/td_libs_Audio.html
Christof
Comments
I created an excel app for generating filter coefficients based on inputs.
http://blogcenter22.blogspot.in/2014/02/fir-filter-coefficients-using-excel-vba.html
-Phil
Hello Phil
Could it be adapted to perform this function?
IF the amplitude of the input signal is stronger between the range of 200hz to 400hz than the amplitude
of the same signal in the range between 1200hz to 1800hz THEN light a green LED.
IF the amplitude of the input signal is stronger between the range of 1200hz to 1800hz than the
amplitude of the same signal in the range between 200hz to 400hz THEN light a red LED.
I've used the basic stamp2 with the Nuts and Volts articles successfully with it's coding for a variety of
things over the years. Needless to say I'm quite loyal to Parallax because they've helped be get my brain
somewhat around micro-controllers. I'll gladly expend the effort to learn and adapt this if you think I'm on
the right track
Rich
Hello Rich,
Sadly I must draw your attention to this post: https://forums.parallax.com/discussion/175149/phil-pilgrim-rest-in-peace-january-13-2023-age-73
@VonSzarvas said:
Hello Rich,
Sadly I must draw your attention to this post:
Hello VonSzarvas,
That is sad news and thank you for letting me know, Seems like an amazing person.
Are there other people on the forum who understand how to analyze and respond to audio data once it is converted to digital?
Quite likely. I would recommend starting a new thread with a subject to catch attention on the matter
Hi,
what you want to do can be achieved with the Goertzel method. https://en.m.wikipedia.org/wiki/Goertzel_algorithm
Christof
Phil's FIR tool looks very useful. Does anyone have a copy of his perl script that does all the work, or know how to obtain it, since we don't know how much longer Phil's website will be up for?
Hi, once again,
have been thinking about your post for a while and what would be a good fit to your task and also to what you wrote about you knowledge. So if you want to dig into digital sound processing the Goertzel algorithm is certainly something to consider. In principle, I have done something like your task using P2 and Taqoz Forth. https://forums.parallax.com/discussion/173880/picture-of-sound-parallax-propeller-p2-as-a-tool-to-analyse-a-guitar-effect-pedal-or-an-amp#latest
Is is done in two steps: First get a buffer of audio samples and then analyse it.
Using a microcontroller shines, if you have to vary things like in my project the frequencies.
On the other hand, if the frequencies are fixed and especially if there are only these two frequencies present in the signal it will be very much easier, cheaper and faster to do at least the filtering and the conversion to envelopes with analog electronics. You would need a preamplifier with a band filter 200...1800Hz and low impedance output (one opamp) then in parallel a RC 700Hz high pass filter and a low pass filter and for each of them a simple rectifier to get the amplitudes. At the end a comparator (a second opamp) to compare the amplitudes.
Christof
Hello Christof
I looked at your picture of sound project and I must say it's impressive and over my head. I am greatful for your responses to my post so I want to explain more as well as upload my schematic at present. a microphone is recieving a full frequency signal and trying to ignore the frequencies that are present outside of the desired range and respond to a comparison between the aplitude of the desired frequencies. It's an industrial floor test device that makes a scraping noise on the surface of the floor and listens for voids. The LM567 chip you see in my schematic is almost suitable for the task, but because I enjoyed the basic stamp2 so much years ago for other prototypes this seems fascinating as an option. I read the Goertzel description and the math fried my brain. Yet.. since the audio output never even has to be considered only the driving of and indictor light. Maybe it is still possible
Hi Rich,
I am not familiar with LM567. As far as I can see, the circuit detects one frequency in a rather small band. No comparison with a second frequency. I tried to make a sketch of my idea, which compares the strength of 2 frequency bands.
Do you use LtSpice? It is an extremely powerful tool and free.https://analog.com/en/design-center/design-tools-and-calculators/ltspice-simulator.html
I am sorry about this wikipedia article. I have difficulties understanding the math too. The pseudo code part gives the application and is rather simple.
You would need >=4000 samples/sec.
If you want to do sound processing with software you would need a processor, which can multiply relatively fast. I assume that Parallax P1 can do it, P2 will work for sure. But I don't think a stamp can do it.
As Parallax does not have anything comparable, I dare to recommend: https://pjrc.com/teensy/td_libs_Audio.html
Christof