Anthem Room Correction (ARC) for the Anthem D1 and D2 SSP



It's all about hitting the target.

When we refer to the target, we are talking about what we want the sound to be, never mind for a minute how we actually do it.  Counter-intuitively, ARC does not attempt a perfectly flat speaker response for several very good reasons.  The first is that it is impossible to get a perfectly flat, or even uniform, response for every listener in a room, so there is no point in trying.  The best systems, like ARC, aim for a “best overall” fit.  Thats not new, as other systems, such as Audyssey, do the same.  What is unique is the fact that Paradigm/Anthem, through their many years of user listening experience and leveraging the Athena project we mentioned earlier, really know what constitutes “good sound”.

Some stuff matters, some stuff doesn't in terms of psyhcoacoustics.  What ARC brings to the table is unique: its particular set of audio and room characteristics which their research has shown to be the most significant and "worth" fixing, as well as that what is not and is best left undone.  One of the more tangible aspects of this is the concept of room gain, and is one of the target settings you have control over.

Room gain is basically the amplitude boost any audio source (be it a speaker or even your own voice) gets in the upper mid-bass range from the boundaries of the room.  Filter, or correct for, this natural boost, and you end up with a very unpleasant, dull sound because our ear-brain just “expects” room gain.  ARC assumes you are not using poor speakers, and thus when it sees a common “swell” in the frequency response of the measurements, considers that to be room-gain and, here is the key, does not target it for correction the way it does a room mode.  In other words, if you look at the target curve in the graphs, it has a very shallow hump (in my case about 4 dB) in the upper bass, and yes, we want that.

But isn't that distortion?  Yes and no.  First, it is by any stretch very mild in amplitude as compared to other aberrations in the system's response, yet correcting it is in a very real way worse than keeping it.  I was challenged by Anthem to “just try it” by accepting all the recommended target settings for Movie mode, and zeroing out the room-gain for Music mode.  Indeed, correcting room gain makes it sound very dead, literally.

Another parameter you can play with in the target settings is maximum frequency to bother filtering or correcting at all.  Anthem elects 5 kHz as a default here.  It's not that their filters can't keep up, but it gets so complicated and messy up there that there is virtually nothing you can do electronically (unless you have really bad speakers, which ARC assumes you dont').

Here is a prime example of how/why acoustical treatments will always be equally important along side ARC:  nothing controls and focuses treble like some absorptive panels strategically placed at first reflection points.

Also part of the target settings is the crossover frequency for each speaker.  This is one that is really going to ruffle some feathers as, with little exception, ARC will tend to use higher frequencies than people would expect, sometimes much higher.   It's not uncommon for ARC to select upwards of 160 Hz for what some people consider full-range speakers.  It's also not uncommon for it to pair that with a subwoofer low-pass of 80-100 Hz.  In other words, ARC will in many cases intentionally create a “hole” in the system's electrical response, because through a combination of the filtering and the room itself, the difference ultimately is taken into consideration (plus you get a nice headroom bonus as compared to crossing everything over at, say, 80 Hz).  People with big tower speakers are not going to like learning that the best fit for them in the room with ARC's help means high-passing them at something close to 100 Hz, but they are going to have to get over it.  What comes first: your pride, or getting the best sound possible?

You can, of course, override these crossover frequency choices at this point, and although Anthem is confident with what the system chooses, depending on your particular speakers and moreover the position of your subwoofer, you may want to force a lower crossover.  In my case for example ARC elected upwards of 160 Hz crossover for my front channels, which yielded excellent results but ended up leaving me with a subwoofer which simply drew too much attention to itself.  Forcing everything to THX-80 Hz still "worked", though I may have sacrified a little of what ARC is capable of by doing so.  Regardless of why you want to change the crossover settings, if in fact if you do, you must do it here within the ARC PC software, as the calculation of the final filters are done here and takes the crossover frequencies into account.  If you change the crossover frequencies in the D2 after the fact, ARC will be thrown completely out of whack.

Once you are satisfied with the target settings, the next step is for ARC to calculate the actual filters it will apply.  This is heady stuff.

ARC uses minimum phase IIR (Infinite Impulse Response) filters instead of the more common FIR filters, since these last, especially at low frequencies, incur long delays which are complex to account for.

For those who love numbers:  from the microphone's 16 bit integer data, signal processing calculations on the PC are done with 64 bit floating point values, yielding 24 bit fixed point coefficients downloaded to the D2.  The IIR filters in D2 have 56 bit fixed point accumulation and use 1st order error feedback when converting to 24 bit fixed point.

The net result of all the number crunching is a green “calculated” curve which gets added to the graphs.  These are what the response would be if a hypothetical “post” measurement were taken.  This curve is very revealing in that it does NOT look like a ruler-flat response.

“Anechoic chambers sound like crap”.   I don't know who first said it, but we all repeat it now because they are words to live by.  Seeking a ruler flat response with no “sound” from the room is not what we want.  Graphs of Impulse Response measurements are really misleading in this respect in that they make things look a LOT worse than they really are.  Again, this goes back to knowing what matters and what doesn't.

Take a look at the ones I'm showing you here.  You would think my room and/or speakers are garbage, but I assure you neither is the case.  Thats not to say there aren't some major issues to correct, sometimes more, sometimes less depending on the room, but it's not as though rooms are as bad as an Impulse Response measurement graph looks, or more to the point that we would want it to look ruler flat.

The point is, what you do see in the visuals is an overall correction of the most egregious distortions.  But no visual aid or graph is going to convey to you what ARC does to you system . . . trust me, you have to hear it to believe it.  I'll explain in a minute.