PortAudio: Play multiple generated sounds by calling function

I've started to play with PortAudio library about week ago. I've checked most of tutorial/test examples but haven't seen solution for what I need. I'm making simple sequencer - I've already done drawing sounds as blocks on the piano roll, but now I need make it sound somehow. I wonder if there is possibility to get it working like this:

run a method that will play one sound,
give frequency and duration of sound as parameters to that method,
play multiple sounds(i.e. three sounds at the same time, bu calling three methods).

In example files it looks more complicated. When I want to play multiple sine waves, so I have to merge all that waves, and then give that data to stream.

Maybe someone got better solutions to resolve this problem?

Solution

I'll just give you a simple example (for a sine wave) and then you can create other types of waves you're interested in. The input parameters needed are

sampling rate (e.g 8000, 16000 etc)

amplitude (the actual values depend on the output format but it's best to have values in the range 0-1 and convert them to whatever format you like/need)

frequency (expressed as a fraction of the sampling rate)

duration (in seconds)

The buffer length for tone data (samples) is determined by the tone duration in seconds multiplied by the sampling rate.

The actual code to create a sine wave may look something like the following

//global variables
const float PI = 3.141593;
const unsigned samplingRate = 8000;
const float amp = 0.8;

float *GenerateTone(float frequency, unsigned duration, unsigned &bufferLen){

     const float freq = frequency/samplingRate; //(e.g 440 / 8000 = 0,055)

     bufferLen = samplingRate * duration;

     float *buffer = new float[bufferLen]

     for(int i = 0; i < bufferLen; i++ ){

        buffer[i] = amp * sin(2 * PI*freq  * ((float)i)/samplingRate);

        }
    return buffer;
}

You can call this function like

unsigned len;
float *pTone = GenerateTone(440, 1, len);//len is an out parameter
...
delete [] pTone; //deallocatone memory when you no longer need it

In C++, you can also use std::vector to store the samples. This way you don't have to worry about memory allocation/deallocation.

std::vector<float> v; //make vector global  
const float PI = 3.141593;
const unsigned samplingRate = 8000;
const float amp = 0.8;

void GenerateTone(float frequency, unsigned duration){

  const amp = 0.8f;

  const float freq = frequency/samplingRate; 

  const unsigned len = samplingRate * duration; 
  for(int i = 0; i < len; i++ )
        v.push_back(amp * sin(2*PI*freq  * ((float)i)/sampleRate));

}

You can also pass amplitude as a paramater but in the examples above the amplitude is hardcoded. Also, see (https://en.wikipedia.org/wiki/Triangle_wave, https://en.wikipedia.org/wiki/Square_wave )