g2312

lets say, we have a simple delay unit like this:

            const double in1 = (*(__in1++));

            const double in2 = (*(__in2++));

            double expr_3885 = ((in2 * 0.2) * 44100);

            double tap_3883 = m_delay_1.read_linear(expr_3885);

            double out1 = tap_3883;

            m_delay_1.write(in1);

            m_delay_1.step();

            // assign results to output buffer;

            (*(__out1++)) = out1;

The problem is the hard coded samplerate of "44100". If the resulting plugin runs in a host with a different samplerate all time values are wrong. Same applies for filter coeff calculations, e.g.:

double expr_3899 = safediv(((in2 * 2) * 3.1415926535898), 44100);

            double sin_3894 = sin(expr_3899);

I would propose to change the export mechanism in a way that the parameter "samplerate" multiplied with the time as factor is used in expressions like the above.

This exports to something like this:
		while ((__n--)) {
			const double in1 = (*(__in1++));
			const double in2 = (*(__in2++));
			double expr_3885 = ((in2 * 0.2) * 44100);
			double tap_3883 = m_delay_1.read_linear(expr_3885);
			double out1 = tap_3883;
			m_delay_1.write(in1);
			m_delay_1.step();
			// assign results to output buffer;
			(*(__out1++)) = out1;

		};
The problem is the hard coded samplerate of "44100". If the resulting plugin runs in a host with a different samplerate all time values are wrong. Same applies for filter coeff calculations, e.g.:
double expr_3899 = safediv(((in2 * 2) * 3.1415926535898), 44100);
			double sin_3894 = sin(expr_3899);

gen-export-samplerate-vs-delay-calculation

indeed, it would be cool if gen replaced [samplerate] by something like :

# ifdef SAMPLERATE 
#  samplerate = SAMPLERATE
# else
#  samplerate = 44100

in the beginning of the output code, and with 

well i don't know if the syntax is correct, but the idea would be to get the current samplerate value if it is defined somewhere in the architecture, which i guess is pretty standard but doesn't exist always maybe ; else it just goes with the current samplerate set as a global variable, which would be easier to change according to your currnet architecture anyway.

(Hmmm apparently it does export a "samplerate" variable somewhere in the code, but it's not used to replace user-written [samplerate]... strange)

in the beginning of the output code, and with samplerate variable everywhere else.

(Hmmm apparently it does export a "samplerate" variable somewhere in the code, but it's not used to replace user-written [samplerate]... strange)


Thanks for the report. I've logged a request for future releases about this.

Thanks for the report.  I've logged a request for future releases about this.


gen~ export - samplerate vs. delay calculation