You can use NAudio and create a derived WaveStream that outputs sine or square waves which you could output to the soundcard or write to a WAV file. If you used 32-bit floating point samples you could write the values directly out of the sin function without having to scale as it already goes between -1 and 1.
As for accuracy, do you mean exactly the right frequency, or exactly the right wave shape? There is no such thing as a true square wave, and even the sine wave will likely have a few very quiet artifacts at other frequencies. If it's accuracy of frequency that matters, you are reliant on the stability and accuracy of the clock in your sound card. Having said that, I would imagine that the accuracy would be good enough for most uses.
Here's some example code that makes a 1 kHz sample at a 8 kHz sample rate and with 16 bit samples (that is, not floating point):
int sampleRate = 8000;
short[] buffer = new short[8000];
double amplitude = 0.25 * short.MaxValue;
double frequency = 1000;
for (int n = 0; n < buffer.Length; n++)
{
buffer[n] = (short)(amplitude * Math.Sin((2 * Math.PI * n * frequency) / sampleRate));
}
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…