I added a second hi pass filter to test PEQ 2 and the result is about the same. I have attached screen shots Image 5 and image 6 with the pink noise file playing. I'm using internal volume at 100 with the loudness unchecked for this test. The lowpass roll off looks about the same to me at 24 and 48 dB/octave, I did not yet try to re-run the REQ tests. I also see the long low slope in the low pass filter when playing music - its harder to get a good screen capture that way. The Echo audio sound card has 6 analog outputs (with no EQ) that I have connected to the AVR-3805 external inputs. No other up or down mixing. The External inputs bypass the Denon DSP in the AVR and only go through the volume control and amps.
Thanks,
Mike
So I did some testing myself to see if I could replicate your results. Here's what I found:
When I dialed in a lowpass (or a highpass for that matter), and looked at analyzer, it showed me a very similar rolloff pattern to the one you're seeing (sharp immediate rolloff, which gradually becomes more shallow). Different filter settings showed the shallowing at different levels of attenuation, but when I used a loud enough pink noise signal all high or low pass filters I tested (I tried many different frequencies and slopes) showed that flattening in analyzer at a certain point.
However, when I measured the filter effects in Holm Impulse, the filters I tested behaved correctly (i.e. it did not show a similar shallowing of the rolloff, but instead showed the expected filter slope down to the noise floor of my measuring rig at around -95dBFS).
For that reason, I think the analyzer issue is a bug in the way the analyzer shows signals, not a bug in the lowpass or highpass filters. Each of the cross lines in the analyzer represents about 5 dB, so in your first example the filter doesn't show "flattening" until it's 30 or so dB down from the passband.
I think the analyzer is just poor at showing large differences in volume. If I play an 80Hz pure sine tone, the analyzer shows a much wider area of signal playing with a long tail off in either direction, whereas an external 1/32 octave RTA just shows much narrower 80Hz-ish tone with some very small sidebands.
So that's my explanation for the analyzer issue; I'm still baffled about your REW measurements. All my measurements show the filters working exactly as expected.
Sorry I'm not more help.