I have carried out other tests on the level of the measured signal and actually the measurements have a good accuracy from the maximum level of the output signal to the power amplifier equal to 8.9 Vrms to the minimum signal of about 10mVrms below this minimum level is affected by noise.
But it is precisely on the noise level I have some doubts, in fact by turning off the generator the RTA shows a much lower noise level than that measured by the FLUKE 8920A and the Picoscope.
The noise measurement is 28mVrms with the Fluke and Picoscope while REW measures three different values, the level meter displays 2.35mV the noise in the 430uV distortion box and the level measured by RTA 951uV.
The noise level measured by RTA should be reported to the highest peak at 47-48Khz the noise level of the distortion box should be much higher because it includes the noise generated on the entire audio band, instead it is only 430uV. The only higher value is that of the level meter but we are well below the 28mV around -20db.
Maybe I'm missing something?
Thank you