Music in the Round #94: Benchmark & Marantz

Sometimes, I feel I'm two people. One, a card-carrying audiophile, is always looking for ways to optimize his enjoyment of multichannel music recordings, a purist pursuit that begins with file playback and leads to DACs, amps, and speakers, while eschewing anything that can complicate or compromise the sound. Thus, while his main system may seem elaborate to outsiders, to him it seems streamlined: NAS>player>DAC>preamp>power amps>speakers. In fact, it's possible to combine the NAS and player in a single device, if that device's CPU and RAM are capable of doing all the tasks—but these product categories continue to evolve so quickly that he prefers to keep them discrete. Another consideration is that most multichannel player implementations aren't very receptive to input from other hardware sources, like disc players.

My other self, attracted by logic and convenience, wants nearly everything completely and neatly finessed by a modern multichannel preamplifier-processor or AVR. He wants to just connect the NAS and the pre-pro to a LAN or plug-in a USB drive, et voilà: decoding, processing, control, and output, all in one, with multiple remote-control possibilities—with all the bells and whistles of a home theater.

But wait—you can't just do that. I have yet to find a modern, network-enabled pre-pro or AVR that will stream multichannel files directly. It amazes me that the designers of all these cutting-edge, microprocessor-controlled, multichannel behemoths have so far failed to include any support of multichannel streaming via USB or Ethernet connection. You have to buy an external server that can feed the pre-pro/AVR via HDMI—or, worse, buy a server and a multichannel DAC to connect via the pre-pro/AVR's 5.1- or 7.1-channel analog inputs—if it has them.

What works best for one person won't necessarily work best for another, so I'm not going to make any presumptions. What I intend to do in this column is to continue to pursue the urges of both Kals. We saw the fork in the road . . . and we took it!

Myriad Benchmarks
In my September 2018 column, I wrote of my discovery of miniDSP's U-DIO8, which converts multichannel data fed to its USB port to eight S/PDIF or AES-EBU outputs at resolutions of up to 24-bit/192kHz. Armed with a U-DIO8, you can use three or four of your choice of stereo DAC, or choose a favorite reference stereo DAC for the front L/R channels, and less expensive ones for the other channels. But I'm not optimistic about mixing and matching—DACs vary too much for me to be confident of success. Think of ensuring that all three DACs have identical gain and processing latency, and whether those vary with the data format. For the moment, I'm comfortable with identical triplets.

In the course of writing about the U-DIO8, I became attached to using a trio of Benchmark Media Systems DAC3 HGC DACs as part of my review of Paradigm's Persona 5F loudspeaker in last month's issue, and figured that they might benefit from a symbiotic relationship with Benchmark's AHB2 stereo power amplifier, which I reviewed in November 2015. Benchmark agreed.

The AHB2 has three gain options—or, more properly, input sensitivity settings: Low, Mid, and High. To produce full output, these respectively require inputs of 9.8V, 4V, and 2V RMS. Benchmark recommends Mid or High for unbalanced inputs, but makes no comment about balanced inputs. With my ear to a tweeter and the AHB2's sensitivity setting at Mid with no signal playing, the noise was lower than with my other amps. That made sense—even bridged, the AHB2 has 9dB less gain or even more than most of them. For the least noise, Benchmark's John Siau suggested that I use the AHB2's Low setting; although it requires a source capable of 9.8V RMS, John Atkinson has determined that the DAC3's maximum output level at 1kHz "was a very high 18.3V from the balanced output jacks."


So out went the preamp, and I went commando with the DAC3s directly connected to the three AHB2s. For the first time I can recall, I heard no noise from my system, even with an ear flush against a tweeter. On the other hand, no preamp means no volume control. I tried using Benchmark's remote control to run all three DACs. That worked okay, but I had to aim the remote carefully to ensure that all three got the same message. In the end, all three DAC3s were set to Home Theater Bypass, and I used JRiver Media Center's volume control.

From my first round with the AHB2s I had expected X-ray clarity, but sometimes too many veils dropped: The flaws in less-than-superb recordings could bee made so obvious that they distract from my enjoyment of the music. Using the AHB2s with the Paradigm Persona 5Fs demanded that I reposition the speakers to keep from aiming their tweeters directly at my ears. Nor is Bowers & Wilkins' 802 D3 Diamond speaker the epitome of flat on-axis frequency response. JA's measurements accompanying my review of the B&W in the June 2016 issue show a hump at around 10kHz that's somewhat broader than the Persona 5F's, though not the 5F's rising tweeter output.

This seems significant—in a surround-sound speaker array, the center-channel speaker is always aimed directly at the listening position. On the other hand, I always aimed the L/R speakers as a stereo pair, so that their axes crossed somewhat behind my listening position.

Now, with no preamp, a high-output DAC, and Benchmark's low-gain power amp, I definitely preferred having the L/R stereo pair aimed directly at me, though I can't really account for why. The change in my system's gain structure—more gain from the preamp/DAC and less from the power amps, meaning that noise in the former is less subject to amplification by the latter—precipitated a gross reduction in background noise, from fairly low to inaudible; was there a stochastic interference between the low noise and the signal that emphasized my perception of the speakers' treble irregularity? I'm grasping at straws, but who cares? I ended up with all three front B&W 802 D3 Diamonds aimed directly at me. The benefit is that they now focus the same mid-to-treble frequency response at the listening position.

I know that we audiophiles can obsess about small changes and make them seem more important than they actually are, but that's what this hobby is about: Eliminating as many audible flaws as possible, however tiny. In this case, the effect was striking at first listen, and endured through everything I played, even tracks that had troubled me before. Exsultate Jubilate, the classic 1984 collection of sacred music by Mozart with soprano Emma Kirkby and Christopher Hogwood conducting the Academy of Ancient Music (CD, L'Oiseau-Lyre/ArkivCD 411832), wasn't transformed, but Kirkby's voice recovered its bell-like clarity—and mezzo-soprano Marianne Beate Kielland singing Finzi's "Come Away, Death" (SACD/CD, 2L 2L-064-SACD), and Sara K.'s cover of "Can't Stand the Rain" from her Hell or High Water (SACD/CD, Stockfisch SFR 357.4039.2), recovered their warmth and resonance.

The sounds of large spaces were impeccable—even strange ones. Having longed for a multichannel recording of Berio's Sinfonia for Eight Voices and Orchestra to challenge that of Péter Eötvös with London Voices and the Gothenburg Symphony Orchestra (CD, Deutsche Grammophon 001708302), I was impressed with the vitality of a new 5.1-channel recording by the vocal ensemble Roomful of Teeth, with the Seattle Symphony under Ludovic Morlot (24/96 WAV, Seattle Symphony CD SSM1018). This is a much more engaged performance than the only 5.1-channel alternative, with Hannu Lintu conducting the Finnish RSO (SACD/CD, Ondine ODE 1227-5). Its sound is also far more immersive—Seattle's Grammy-winning sound engineer, Dmitriy Lipay, leans in that direction. I was surrounded by the orchestra and by the eight voices, variously amplified and unamplified. From the seeming hyperreality of the conductor's perspective, the effects were stunning but also somewhat intimidating. All seemed results of the new gain structure and clarity made possible by the Benchmark DACs and amps.

Marantz AV8805 multichannel preamplifier-processor
At the turn of the century, I was certain that the rise of home theater would be a springboard for multichannel recording and playback of music. Surely, owners of home theaters would hear what their surround-sound systems could do for films and realize that they could do the same for music-only recordings, and would adopt multichannel as the new standard for recording. After all, the necessary playback gear was already installed.

How wrong I was. Resisters today still point to the necessity of specialized equipment as an excuse. The solution is still that any modern audio/video receiver or preamplifier-processor can accept discrete multichannel via HDMI, and most will accept DSD and high-resolution PCM. Some offer multichannel analog inputs to accommodate external DACs, and multichannel analog outputs to accommodate multiple external power amps. Finally, room correction entered the consumer market via AVRs, and is now offered in an increasing variety of flavors.

jeffhenning's picture

Unfortunately, most AV pre-amps don't have serious output capabilities. Given that most consumer amps have upwards of 30dB in gain, there's not much incentive for them to offer it.

I used to own a Marantz 3800 pre-amp that I bought in the early 80's and it was rated at 9.8V output (or so...had it for 20+ years). The new Marantz 8805 is rated at one third of that or less. They don't offer much in the way of performance specs for their new stuff.

I've found the Emotiva XMC-1 to have some serious output, but, no, it's not near 18V. For my subs (Rythmik) and surrounds (Presonus S6), though, I do turn their levels down as far as I can and drive the pre fairly hard (it sounds great). Do the same with the Parasound that's powering my LS50's. I will be interested in seeing how my system sounds with an AHB2 in it. That's my next amp buy.

Kal Rubinson's picture

I just realised that the legend for the Audyssey graphs is not entirely accurate. The center panel is, indeed, the measured unequalized FR. The left panel represents the measured and unequalized FR above 300Hz along with the predicted FR in the equalized range below. The right panel is the predicted FR for a full spectrum correction.