I am using the circuit blocks RTA thats supposed to show level in db and PEAKENV in conjunction with READBACK to display the HEX of the peak value. I have calibrated the input level such that the RMS meter shows about +1dB and the hex around 8,0,0 this being empirically determined by listening tests and reference to the input level as the input full scale. The source is the generator in the well known REW program. Furthermore the DSP input is directly connected to a DSP DAC and after the recommended LPF is routed back to the mic input of the PC and level displayed again by REW. My problem is the dynamic range between the full scale input and the noise floor.
Basically REW confirms this to be around 72dB, not bad for using a crappy laptop soundcard HOWEVER the RTA block insists the noise floor is only -36dB!! this is in fact supported by the ratio of full scale and noise floor levels indicated by the peak hex values (8,0,0 and 0,1D,A0). Surely a full scale value of 8,0,0 from the ADC indicates the ADC result has been divided by two somewhere and maybe this accounts for the two fold difference between internally and externally measured dynamic range ?
I am extremely concerned about this issue as my project will rely heavily upon the accuracy of internally measured signal levels, please help I have searched the forum to no avail, surely someone else has experienced this ?
Thanks in advance