If you look at the SpecLab cap of the AVR BW-limited pink noise cal tone, it's dominant at 50 Hz, right in the center of 20-80 Hz, so that will dominate the RS meter reading when you set the SW level.
Correction files say to add 1.5dB at 50 Hz to the reading to correct for the RS meters C-weight curve, so if you're reading 75dB, it's really 76.5dB or higher. Some say as much as +3dB over the RS meter reading.
MKT, depends on how you run your sweep through REW. If you set your AVR to stereo and use L&R, redirecting bass to the SW channel, you're getting double bass back into REW through the mic.
Here's the experiment I did a few years back running a sweep in stereo, then switch to all channel (5 sats) stereo, and it added the predicted +9dB to the sub output:
This is a very interesting subject because conventional wisdom says that phase coherent same content in all 5 channels plus LFE summed at the SW output equals 121dB peaks. IOW, basically what my experiment shows happens when you send the same signal to all 5 channels vs 2 channels and redirect that info to the SW out.
That would mean that the sub is up to 6dB hot already with some soundtracks.
So, take the average Joe who uses the RS meter and rumble tone and set the SW level to 78dB, thinking he's a few dB hot. He's actually potentially 10dB hot. He then bumps the SW trim level to demo a scene or 2 and now he's 15dB hot.
That's cool with a commercial sub that's limiter and filter choked don low, but try that with my system and bad things will happen to the sub hardware, the house and the persons in the way.
Carp, Just use 1/3 octave smoothing and that should give accurate level matching without the peaks/dips.
Beast, yep, if you can't get the job done with 40+ liters of displacement, you're doing pretty much everything wrong. It would be awesome to see the results of any diff in the 2 shelf filter methods. Ping me when you post results so I don't miss them?
When I first bought clone amps 3 years ago, I didn't bother with the gain limit and voltage peak limit settings. DON'T make that mistake if you have these amps! Every other pro sound amp has global gain/limiter settings, making it much harder to set up a proper gain match. The clones are unique in that sense, but that can work against you if you don't use them properly, either clipping/limiting when you shouldn't be or damaging the amp or drivers.
In one example, playing a heavy hitter scene with strong ULF would cause the VPL lights to flash momentarily but everything sounded OK and the drivers were handling the scene at ref level with no problems. That's what I was testing for, to see how much displacement was needed with 'X' amplifier at ref level in 3500 cubes while playing heavy hitter scenes.
This indicated that the system was amp-limited.
After we spent the day measuring voltage out of the AVR SW output while playing those scenes at '0'dBMVL and '0'dB SW trim level, we realized that some scenes measured a lot higher voltage than the AVR spec. We then sent that same signal through the L/T and measured up to 10V out!
With the clones dip switch settings at global, the input signal was way too much and sending the amp over its limits on those scenes. We knew the L/T wasn't clipping because we specifically designed and tested the power supply to handle worse-case scenarios.
We then ratcheted the SW trim down to -8dB and designed and installed a circuit to ratchet the L/T signal down to a predetermined level and remeasured. With the new voltage out numbers, we set the gain limit and voltage peak limit settings so that a 2 O'clock front panel gain setting would run the amp to full power with no clipping and give the best signal to noise ratio.
After that, when playing the same scenes, there was no limit or clip lights flashing but the drivers soft-bottomed. This revealed that I was using too much amp for the system!
Minding signal chain roll off, gain matching, matching the power to the system and proper calibration may be the difference between a happy camper or a trip to the landfill, FWIW.