Hi,

I tried using SndBuf instead of SinOsc (both on Rasp Pi and a MacBookPro), but SinOsc seems to offer better performances.

On the Mac (2.7 GHz Intel Core i5, 16 GB 1867 MHz DDR3, MacOS Sierra 10.12.6), running the code you'll find below, I've got this CPU usage (I used top from terminal to check the CPU usage):

SndBuf x 50 - CPU usage 24.0%

SinOsc x 50 - CPU usage 16.4%

the same seems to happen on Rasp Pi 3, despite I had to use 25 oscillators to avoid glitches.

here's my code

------------SinOsc Version------------

SinOsc s[50];

for(0 => int c; c<s.size(); c++)
{
    s[c].gain( (1/( s.size()) $ float) * 0.25);
    s[c] => dac;
}

while(true)
{
    for(0=>int c; c<s.size(); c++)
    {
        Math.random2f(300, 500) => s[c].freq;
    }

    200::ms => now;
}

------------SndBuf Version------------

SndBuf s[50];

for(0 => int c; c<s.size(); c++)
{
    s[c].read(me.dir() + "/sine_1024.wav");
    s[c].pos(0);
    s[c].loop(1);
    s[c].gain( (1/( s.size()) $ float) * 0.25);
    s[c] => dac;
}

while(true)
{
    for(0=>int c; c<s.size(); c++)
    {
        Math.random2f(300, 500) => s[c].freq;
    }

    200::ms => now;
}

-------------------

cheers,

Mario