I've done quite a bit of amp testing using REW, my external interface, and special attenuators. The distortion measurements, both stepped and swept, are very useful and revealing. But I have a question regarding measurement of the noise floor of the device under test (DUT).
I typically run long FFT and multiple averages when testing distortion, so as to pull the harmonics out of the noise for better resolution. But it seems that this will minimize the DUT noise measurement, meaning the noise floor on the graph is deceptive. REW derives a noise floor from these measurements, but how? If it's averaging, then it seems that the noise floor will be artificially low. If it's a single sample, is there weighting applied? I want to get a good measurement of the DUT noise, and I'm not sure what the best configuration would be in order to do that.
I typically run long FFT and multiple averages when testing distortion, so as to pull the harmonics out of the noise for better resolution. But it seems that this will minimize the DUT noise measurement, meaning the noise floor on the graph is deceptive. REW derives a noise floor from these measurements, but how? If it's averaging, then it seems that the noise floor will be artificially low. If it's a single sample, is there weighting applied? I want to get a good measurement of the DUT noise, and I'm not sure what the best configuration would be in order to do that.