HifiZine
The enthusiast's audio webzine

A Paradigm Shift in Hi-fi

I’ve been an audiophile since I was fairly young, and I have the string of purchases of traditional hi-fi gear to prove it. Like most other people, I was “ecstatic” with my system, until the itch hit and I felt it was time to upgrade to something better. It’s a vicious cycle, and one that cost me more money than I care to admit. But the beauty of it is that it keeps you forking out money to “almost” reach perfection, only to require a new set of expenditures the next time you upgrade.

Often these upgrades are because you want “better imaging,” or “better tonal balance,” or “more dynamics,” or “greater placement flexibility,” or whatever. In other words, you want something new that retains all the strengths of your current equipment, but fills some gap or shortcoming that bothers you a bit at first, and later starts to gnaw at you.

The root problem here is that you trust the skill and expertise of the designers of your equipment more than you trust your own judgment. Break the cycle! I’m here to show you how. First, understand that there is no magic in hi-fi. What others have done, you can do too. Second, know that no one is a greater expert on your preferences than you are. Third, realize that the idea of an “absolute sound” is only a way to get you into the general ballpark of your final system, and that fine-tuning that last bit of sound is always subjective – and this variation is what makes us interesting and unique. Taking this responsibility is the paradigm shift I’m referring to in my title.

Gaining Control – Going Active

Most people buy speakers that have capacitors, resistors, and inductors that divide the highs, mids, and/or lows into separate ranges. Usually they are located inside the speaker cabinet, and they are called a “crossover.” This is more accurately called a “passive crossover” because it occurs after the amplifier and uses passive parts to achieve this division. Basically, a speaker manufacturer selects a set of drivers, builds a box to put them in, and designs a passive crossover to give you the best sound possible.

But there are some serious problems with passive crossovers: 1. they suck up power; 2. they cause horrible phase shifts (except 1st order crossovers, which are phase-coherent, but have the drawback of causing more harmonic distortion because of their shallow slopes); and 3. they are voiced according to the preference of the designer. The solution to (1) and (3) is to go active. The solution to (2) is also going active, but only in a very special case: the DEQX.

After owning a large number of passive speakers, it occurred to me that I simply would never find the “perfect” speaker for me out there, for the simple reason that the builders of those speakers all voice them for their priorities, their preferences, and their environments. So I built my own speakers. I call them “The Ellas,” after Ella Fitzgerald. Originally they were a vented MTM-WW speaker, but I sit fairly close to my speakers (about 10 feet away), and that was not enough room for the MTM setup to integrate properly. So I switched it to a MT-WWW setup instead. I used a pair of upgraded Dynakit VTA ST70’s built by Bob Latino for the mids/highs, an Audio-GD fully-differential solid state C1 amp on the bass, and a highly modded DCX2496 as the active crossover with frequency response equalization and room correction built in. I also used a Mapletree 2a SE Octal-based tube preamp, and a NOS Havana DAC for a source, with a USB Thingee doing USB-SPDIF conversion from my PC-based playback system.

You can see my build of the Ellas in this thread on AudioCircle – warning though, there’s lots of pictures.

With an active system, it’s critical that you have a good calibrated microphone and measurement software. I have a Behringer microphone that was calibrated for me, and I use the HOLM Impulse measurement software with a USB-based M-Audio preamp hooked to my laptop.

The Journey

I was very happy when I went from passive to active in my speakers – the overall dynamics, clarity, ease, and coherency all took a major step forward, especially once I figured out the measurement software and got the system dialed-in well. The DCX2496 is very nice in that it lets you set any crossover point, and use almost any crossover slope. It even lets you use asymmetric slopes (or no slope at all). As a means for really dialing in a sound you like, and having almost god-like power over your speaker’s output, it’s unsurpassed. And the mods took it from a good sounding piece to a world-class sounding unit.

One thing I discovered as I listened, is that I’m fairly sensitive to phase and timing errors. When I used lower-order slopes with less phase and timing errors, I liked it better. Using higher-order slopes sounded “cleaner” in the short term, but I found them fatiguing in the long term. I lived with my speakers using 2nd-order Linkwitz-Riley filters very happily, for quite a long time.

Enter the DEQX

Then I heard about the DEQX, which does most of the things the DCX2496 does, but takes it a step further. It measures (and corrects) phase and timing errors within a single driver. It does this with very steep crossover slopes, giving you the option to use (for example) 96 dB/octave slopes that are phase and timing perfect. It time-aligns all of your drivers. It measures and corrects group delay.

I started reading about the unit, and I was very intrigued, but at close to $6k for the unit that I wanted, it was out of reach. Then I saw the exact unit I wanted come up for sale on audiogon.com, for half the new cost. I figured I should at least try it, and if it didn’t work out, I could sell it for exactly what I paid for it. On the other hand, if I did like it, I could then sell my much-loved DCX2496, MHDT Havana, and Mapletree 2a SE.

Arrival and Setup

The first thing to note is that the HDP-3 is far more advanced and better built and engineered than the previous 2.6 model. Gone is the switching power supply, replaced by a very good analog linear supply. The balanced outputs use super high-quality Jensen output transformers. The grounding has been improved. The DACs are much better. The op-amps have been upgraded. The chassis is more robust and befitting of a truly high-end piece of equipment.

After making all the connections, installing the measurement software, and hooking up the DEQX via USB to my laptop, I spent quite a lot of time reading the 200-page manual. Yes, 200 pages. And you need to read all of it. This thing is amazingly flexible and powerful. But there are a lot of ways you can screw it up. So read the manual. Grok the manual. Once you have, you are ready to proceed. Hook up your microphone, put your speakers in the center of your room, elevate them 3 feet off the floor (this is all to avoid reflections and floor bounce), and start measuring.

After a few sweeps, you are done with the individual driver measurements. Next, you configure the crossover points and slopes, instruct the unit to correct for phase, group delay, and time alignment. The slopes are calculated, all the corrections are calculated, and then it has you do more sweeps to verify that the output actually matches the input and all the corrections.

A couple of helpful hints. Figure out the best crossover points by understanding where beaming will occur in your drivers. This affects total in-room power response. If you are going to listen on-axis all the time, then you have less to worry about. But if you (or anyone else) listens even a tiny bit off-axis, you need to realize at what point drivers start to beam. For a 6.5 inch driver, it’s 2 kHz. You don’t want to cross over higher than this. For a 5.5” driver, it’s 2.5khz. Again, don’t cross over higher than this. You can Google “length of sound wave calculators” to figure out the exact frequency that is best for your drivers. Also, with most speakers, there is a “floor bounce” problem at about 200 Hz (sometimes higher, sometimes lower, depending on where your drivers are located). If you have a 3-way speaker, you can circumvent this by setting the mid/bass crossover point around this area.

Finally, put the speakers in their normal spots in your system and run more sweeps for room correction. Once this is done, you are ready to rock and roll!

Impressions

The DEQX is lovely! It’s a bit of a pain to take the initial measurements (as you have to move the speakers to the center of the room and elevate them off the floor). But, once the measurements are done, you can set the speakers up any way you want. I have loaded three different sets of crossover slopes and points into it at a time, and can toggle between the three in real-time. Then I can keep the one I like best and try out a couple of new slopes/points. It’s letting me very quickly cycle through the settings to end up with the best sound.

Moving from the DCX2496 to the DEQX, the biggest improvement (in my opinion), is getting perfectly flat frequency response. That alone makes the whole listening experience more relaxed and pleasurable. Your ear/brain is no longer having to “work” to fill in gaps in the presentation. The DCX2496 was very good, and got me much closer than any other crossover/speaker setup I’ve heard, but it is limited to maybe a dozen or so points of correction before it runs out of CPU power. With the DEQX, you have a thousand points available for correction. That level of granularity is a big improvement in what you hear.

The second big improvement is phase correction and time alignment. That just moves the whole sound out of the “digital” realm and gives a very relaxed, analog-sounding presentation. Of course, this is the HDP-3, which has things that the old 2.6 does not, like a much better power supply, better op-amps, better grounding, and balanced outputs using Jensen transformers. So it’s going to sound less “digital” than the 2.6 for those reasons as well.

Finally, I’m able to enjoy my old Billie Holiday, Louis Armstrong, and mono classical recordings much more than I ever have before, on any of my systems. The DEQX is the real deal, especially if you use it fully active and can take advantage of the super-steep slopes and phase correction.

In fact, right now I’m listening to a recording from the early 1940’s of Rachmaninov playing his own piano concertos. I’ve had this recording for a long time, and I’ve always respected the performances, but I’ve never loved them. I have now fallen in love with them, thanks to the extra color, emotion, passion, and beauty that the DEQX lets through. And when I listen to more modern recordings, like the difficult Eva Cassidy recordings, or the superb Patricia Barber recordings, it just gets better and better.

Conclusion

I got it for a steal, but I still have to (reluctantly) sell the modded DCX2496, the Mapletree, the Havana, the Scott Nixon, the USB Monica, and the Bolder modded Duet to recoup my investment. I’d prefer to keep these pieces around for a second system, but finances dictate that they must go.

The upside is that my system is much simpler now, with way less cables and junk in my AV cabinet. And I’m re-discovering a bunch of music that I’d shunted off to “poor recording” land a while ago. How cool is that, an upgrade that lets you enjoy more of your music collection, not less! It’s the opposite of those pieces of hi-fi gear that “show you how bad this recording really is.” Instead, you can hear that the recording is not perfect, but the DEQX presents it in the best possible light. Or, put another way, it shows you what there is to love in any given recording, and does not focus on the things to dislike.