The LFE channel on the XMC-1 has a fixed low pass filter. I believe it is at 120 Hz (I'm not sure if that's the exact frequency, but I believe it is.) That low pass filter on that channel is a requirement of the Dolby specification. (Both the presence of the filter, and its frequency, are specified by Dolby, and so are "just part of a properly functioning decoder".)
In theory we could allow you to turn it off or move it when you're using formats other than those provided by Dolby. Likewise, we could make it adjustable, and just hope Dolby didn't notice. However, we chose to just do it "the Dolby way" on the XMC-1.
The yellow plot is subs connected to normal sub output (HDMI channel 4 in REW ie LFE) going through the XMC (set with 24db slope in XMC)
The purple that is exactly the same trace is subs connected to the right channel, set as large and LR4 120hz filter added to my subs via DSP in the amps (inuke). I added 110hz and 130hz so you can see what it isn't.
Last Edit: Jan 25, 2018 14:09:43 GMT -5 by markymiles
Arcam AVR550, 2 x Nakamichi AVP1 7ch Power amps, DIY LCR, 4 x Volt 8LX Surrounds, 2 x BMS 18N862 DIY sealed subs.
There is a legitimate objection to using REW plots (or any other room plots) to prove a point involving the frequency response of an electronic device. The objection is simply that both speakers and room acoustics have the potential to have a truly profound and sometimes unpredictable influence on the measurements.
The bottom line is that, while you may have carefully established a baseline, taken your measurements carefully, and calculated accurate results, not everybody does. It's very easy for huge variations in room acoustics to completely obscure legitimate data - or for a huge difference in measured response to turn out to be due to a shift of a few inches in microphone placement. (And there is a somewhat long history of false claims and bad data based on poorly handled in-room measurements.) In contrast, it's simple to measure the frequency response of an electronic device with great accuracy by plugging an oscilloscope or RTA into its output terminals. As a result of all this, most folks with an engineering background simply consider in-room measurements to be a really bad way to accurately characterize the frequency response of an electronic component. (To the point where, if an in-room measurement produces an interesting result, they assume that "it's probably the result of a measurement error".)