Go Down

Topic: DTMF decoder library (Read 112145 times) previous topic - next topic


i did buy 6 of these chips $6 ea. haven't tried them yet but im sure they will work fine.
Electronics Engineering, "EE=MC^(OMG x WTF?!)"
Professional Auto Collision/Custom tech


I'm trying to use el_supremo's DTMF library. No matter what I do, I can't get a reading/signal from it.
I am wiring up the handset output of my telephone with this schematic:

Taken from: https://coolarduino.files.wordpress.com/2011/02/dc_bias11.png
I know this circuit is working because other libraries that do different things correctly analyze the audio.

Is there something I need to change? I would prefer not to by a DTMF Decoder.


Which other libraries, what do they do?

Don't send me technical questions via Private Message.


This is a basic speech recognition software: https://github.com/arjo129/uSpeech. It worked great.
This gave a response, but because it can only detect one tone at once, it didn't work: https://github.com/jacobrosenthal/Goertzel



Works very well for me to display the code on the computer.

For cons, I can not seem to use a DTMF to turn an LED on output.

How to make the LED lights up when you press the button 5.

thank you


Post your code (using code tags).

Don't send me technical questions via Private Message.


Oct 14, 2013, 07:00 pm Last Edit: Oct 14, 2013, 07:22 pm by hb3ywu Reason: 1
Thank you for the reply,

I took your code and have just added this:

Code: [Select]
if(thischar == 5) {
digitalWrite(led, HIGH);

I know (thischar == 5) is not fair. But I do not know what to put.

with this code
Code: [Select]
if(thischar) {
digitalWrite(led, HIGH);

The LED lights up, but with all the DTMF.

Code complet:
Code: [Select]
//DTMF Remote
//Led pin 13
//Audio IN A0

#include <DTMF.h>

int sensorPin = A0;
int led = 13;
float n=128.0;

// sampling rate in Hz
float sampling_rate=8926.0;

// Instantiate the dtmf library with the number of samples to be taken
// and the sampling rate.
DTMF dtmf = DTMF(n,sampling_rate);

void setup(){
  pinMode(led, OUTPUT);     


int nochar_count = 0;
float d_mags[8];

void loop()
  char thischar;

  /* while(1) */

  thischar = dtmf.button(d_mags,1800.);

// Led ON receive DTMF "5"
//??? no (thischar) fonction
  if(thischar == 5) {
digitalWrite(led, HIGH);

  if(thischar) {
    nochar_count = 0;
    // Print the magnitudes for debugging
//#define DEBUG_PRINT
    for(int i = 0;i < 8;i++) {
      Serial.print("  ");
  } else {
    // print a newline
    if(++nochar_count == 50)Serial.println("");
    // don't let it wrap around
    if(nochar_count > 30000)nochar_count = 51;

I started, but you've probably noticed ...

thank you very much


The code returns the detected button as a character so you need to use:
Code: [Select]
if(thischar == '5') {
  digitalWrite(led, HIGH);

This allows it to also return the symbols such as '*', '#' and, if you have the full 16 button pad,  'A', 'B' etc.

Don't send me technical questions via Private Message.


The simplest things are the most obvious.

Thank you, now it works fine.

FYI, perfect decoding at a speed of 50ms, but only with the wiring you suggested.

thank you very much for the help and the job.

HB3YWU / Mike


DTMF.cpp gets compile errors for me. I use the current version of the 'arduino' package  from the Fedora 'updates' repository on Fedora 20. Perhaps something has changed in the avr-gcc compiler.  I get error message:
    unable to find a register to spill in class 'POINTER_REGS'
The line causing the error is:
   dtmf_mag = sqrt(Q1*Q1 + Q2*Q2 - coeff*Q1*Q2);
I fixed the problem by splitting up this expression into 3 lines:
    float tmp;
     tmp = Q1*Q1 + Q2*Q2;
     tmp = tmp - coeff*Q1*Q2;
     dtmf_mag = sqrt(tmp);



Pete, it is possible that your library can decode CTCSS tone? this guy make something that : http://whatcomradio.wordpress.com/2013/05/28/dtmf-decoding-with-arduino/
but this radio can't read DTMF tone or i made a mistake ?


The library is specifically designed to decode the 8 DTMF tones. Decoding CTCSS tones is much more difficult because in some cases they are less than 3Hz apart whereas the DTMF tones are at least 70Hz apart.
If you were aiming to decode only a few of the tones and could choose a few that were much further apart, it can be done, but trying to decode all of them is probably impossible on an Arduino without external hardware.

Don't send me technical questions via Private Message.


When I try to compile the example dtmf_test I get the following and some more error-messages:

C:\Users\Peter\Documents\Arduino\libraries\DTMF\DTMF.cpp: In member function 'void DTMF::detect(float*, int)':

C:\Users\Peter\Documents\Arduino\libraries\DTMF\DTMF.cpp:179:1: error: unable to find a register to spill in class 'POINTER_REGS'

Arduino: 1.6.1 (Windows 7), Platine: "Arduino Uno"

Has anyone an idea how to fix this?



Mar 18, 2015, 06:04 pm Last Edit: Mar 18, 2015, 06:11 pm by el_supremo
The code compiles with Arduino version 1.0.6 but it appears that there is a bug of some sort in the compiler in versions 1.6.0 and 1.6.1 - probably related to the -Os flag.
I have a workaround. Replace the entire DTMF:detect function in the DTMF.cpp library file with this:
Code: [Select]

// return the magnitudes of the 8 DTMF frequencies
void DTMF::detect(float dtmf_mag[],int adc_centre)
  int index;
  float d_tmp;

  /* Process the samples. */
  for (index = 0; index < N; index++)
  // Calculate the magnitude of each tone.
  for(int i=0;i < 8;i++) {
// El_Supremo 150318 the compilers in Arduino verisons 1.6.0 and 1.6.1
// generated "unable to find a register to spill" error in the original statement
// here. Breaking it into pieces worked around the problem.
//     dtmf_mag[i] = sqrt(Q1[i]*Q1[i] + Q2[i]*Q2[i] - coeff[i]*Q1[i]*Q2[i]);

  // This is the equivalent of sqrt(real*real + imag*imag)
  d_tmp = Q1[i]*Q1[i];
  d_tmp += Q2[i]*Q2[i];
  d_tmp -= coeff[i]*Q1[i]*Q2[i];

    dtmf_mag[i] = sqrt(d_tmp);

I didn't notice/remember the report of the same problem in message #24 by griff2a until after I'd found a fix. My first attempt at splitting up the statement was similar to his but it didn't help so I had to split it even further.

Let me know if this works for you. I will update the link to the library.

Don't send me technical questions via Private Message.


Mar 18, 2015, 06:23 pm Last Edit: Dec 19, 2016, 04:49 pm by el_supremo
The most recent version of the library is attached to the first message in this thread.

Don't send me technical questions via Private Message.

Go Up