The Center, Measured

I've written before about building a center channel to answer many questions I had about existing 3-way uber centers.  I came across Troels Gravesen's Illuminator Center and had questions... so I started looking at some commercial designs with similar parameters:  3-way, center channel with about 7" woofers. 

I came across models from Legacy, ATC, and even Wilson.  Besides cost one thing that struck me was the huge physical size of the speaker cabinets and the (in my mind) pathological attempts to reach near full-size speaker bass extension.  

The 3-way designs with 7" woofers had excellent potential for frequency response, off axis performance and dynamic range.  Where I really didn't agree with any of these esteemed brands or Troels was in the attempts to go as deep as they do because of size and also dynamics.  Here I agree with THX/Lucasfilm's early specs for home theater satellites:  Cut them off at 80 Hz and let the subwoofer do the rest. 

In exchange for commiting to a high-ish cut off point I get a much smaller speaker, with tons of dynamic range and maybe a fourth of the distortion of the big box designs.  With a Hypex DSP powered three-way plate amp we get the equivalent of 1,000 Watts but until now I've not actually published the measurements.

Any idiot can use a DSP plate amp to get a flat frequency response by using the EQ features like a sledge hammer to force obedience.  What I wanted was excellent on and off-axis performance, which I think I achieved. 

You are just going to have to take my word for the fact that the EQ adjustments are overall pretty light and subtle.  The magic is in the integration of the drivers, crossovers and digital delay.  If I had achieved excellent on-axis performance via EQ alone you'd see the off-axis performance turn into an absolute hot mess.

The Center of the Center

What I wanted to show here is the value of a midrange driver in a any speaker, but especially the center.  I also want to talk about why you need to angle the speaker so the midrange points towards the users.

Unfortunately due to time and space constraints I don't have the chance to do quasi-anechoic measurements for the center channel.  All I can do is 1 meter gated / blended measurements from the center actually on top of my Butcher Block rack, aided by IsoAcoustics center channel stand. 

Please ignore the little squiggly signals below 300 Hz as most of those move around with the mic placement, and the response in the bass includes the home theater's crossover response. 

The plot below shows the on-axis (red), 45 degree horizontal (green) and 45 degree above (black): 

 

 

Horizontally I achieved the goal of having a very wide and even dispersion.  Honestly it's even better than I expected.  The maximum drop-off is around 5 dB at 10 kHz, pretty outstanding for any speaker, and the crossover points of 350 Hz and 3 kHz are invisible on axis or horizontally, showing excellent driver time and phase alignment.  The only roll-off present horizontally is the natural behavior of the tweeter. 

If this was a two way design there would be 5" to 7" mid-woofers pushing towards 2 kHz.  If we assume we use the same 7" woofers I did there would be an off-axis deficiency because the mid-woofers become directional around 1 kHz.  By lowering the crossover point of the Revelators to 350 Hz we prevent them from being directional at all.  See the manufacturer's measurements, below:

From 350 to 3 kHz the 4.5" midrange takes over.  Assisting to achieve the results in the first graph is the use of 4th order Linkwitz-Riley filters.  These would be nearly impossible to achieve in a passive speaker with a flat baffle because of the time and phase alignment needed plus the sheer number of parts. 

I wish the solution was perfect but there is some interaction between the drivers.  Look at the black line around 3 kHz.  This is 45 degrees above the tweeter.  The dip you see is caused by the phase cancellation that happens when the tweeter is closer to the microphone.  In other words, the midrange is now delayed compared to the tweeter and causing destructive interference above the tweeter axis.  About an 11 dB dip at 2.5 kHz, with an overall effect between 1.8 kHz and 3.6 kHz. On the other hand, measuring 45 degrees above tweeter axis is a very extreme situation.  Stereophile for instance only measures +15 degrees.   This kind of behavior is typical for two and three way speakers with vertical alignments, so the lesson here is to sit with your ear no higher than the tweeter axis, and often below.