Hi,
As part of a subband processing with the input running at 44.1 kHz I get 33 subband gains running at 1378.125 Hz.
When I want to multiply the gain vector with the DC source I have a problem with a sampling rate mismatch. It looks like whenever I set the DCSource sampling rate to 1378.125 Hz it rounds it to 1378.13 Hz. Is there a way to solve this and stick to 1378.125 Hz?
Thanks.
Pablo
1:18pm
Any thoughts?
3:43pm
Hey Pablo,
You're running into the 6 digit limit of precision for 32-bit Floats. We automatically round the source modules sampling rates to 6 digits, but if you naturally achieve something like 1378.125 (7 digits) in your signal flow, we don't round that wire's sampling rate. To get around this problem I suggest doing something like the following layout. Note that I'm setting the sampling rate to an integer before the Multiplier module, then setting the sample rate back to a 6 digit number after combining the wires. This avoids the potential differences in the rounding for the DCSource and SetWireProperties modules if a full 6 or 7 digit number is used.
The sampleRate in the wires is only used to calculate filter coefficients, and having an error of 0.05 Hz should be negligible for anything downstream of this.
I hope this helps, and sorry for the delay.
-Axel