Benchmark Debuts High Dynamic-Range Amplifier

John Siau of Benchmark (right) and Laurie Fincham of THX (left) gave a provocative presentation on the final morning of the show, entitled "Why Most 24-bit Audio Systems Still Deliver 16-bit Performance." The thesis was that even with D/A processors capable of operating with a dynamic range >20 bits, there is still the resolution bottleneck imposed by the amplifier. As I have pointed out in Stereophile's reviews, amplifiers with a sufficiently low noisefloor and a sufficiently large maximum voltage swing to equal hirez audio's dynamic range are a) rare and b) necessarily expensive. Benchmark, using the unique, high-efficiency, feed-forward amplifier modules designed by Laurie Fincham, Owen Jones (the twin brother of TAD's Andrew Jones), and Andrew Mason, that I wrote about in my 2012 CES report, aims to address both those issues.

The Benchmark AHB2 amplifier, named in honor of the company's founder, Allen H. Burdick, who passed away just before the show, has a dynamic range claimed to approach 130dB, and outputs 100Wpc into 8 ohms, 170Wpc into 4 ohms and can be bridge to produce 340W into 8 ohms. THD+noise is said to be <–108dB relative to full output at 1kHz. John Siau explained that the THX modules are operated in a mode that gave the lowest noise and distortion rather than the maximum efficiency I wrote about in 2012. A key to the low THD is the use of 0.01%-tolerance resistors at critical points in the circuit. A switching power supply is used, Siau feeling that this was optimal because all the power-supply spuriae will be out-of-band and therefore more readily filtered. It also means that the ubiquitous 60Hz magnetic interference from conventional supplies, which you can often find in my amplifier reviews and is picked up by ferrous parts in the circuit, is absent.

The AHB2 is made in the USA and will be available in December. It is shown below with the excellent Benchmark DAC2, which is scheduled to be reviewed in the February 2014 issue of the magazine. Price was still be decided; it is expected to be about $2500.

jimtavegia's picture

I would think that would be hard to keep in stock at that price. Game-changer. 

Jason Victor Serinus's picture

It all depends on how it sounds. 

jmsent's picture


First off, where does there exist a room, a pair of speakers, or a set of ears, that has the ability to handle 130dB of dynamic range? Where is there a commercial recording in existence with anywhere near this much dynamic range? A noise floor 108dB below full power is all well and good, but I'm pretty sure other amplifiers have achieved this. My Pass amps produce output noise in the 200 microvolt range; totally inaudible with my ears right up to the speakers. In any event, it isn't just the amp that contributes to the noise floor. There's the preamp and even the DAC. We ultimately run into the thermal noise of the resistors themselves in the circuitry that establish the noise floor, but even so, we have already subjectively achieved total silence, for all intents and purposes. Whatever the amp's s/n noise spec,  if you can't hear any noise from the amp  at the listening position or even a couple feet in front of the speakers, and can't clip your amp even at insane listening levels, I'd say "Mission Accomplished". Plenty of existing systems have this capability. Anything more is just icing on the cake and useful for little more than ad copy. This strikes me as yet another solution in search of a problem. Or perhaps it's just an opportunity for THX to license yet another new technology.

fyzzics's picture

I think  the point here is that this amp actually matches the SNR and THD specs of many DACs out there.  And the Benchmark preamp in the picture does have comparable noise/THD specs, so the amp is not superfluous from that standpoint. And it is sort of affordable, to boot.

Benchmark can't help it if *your* preamp (or mine) isn't that good :-)

I totally agree with your assessment of the actual audibility in any real room of noise/distortion 110-130 dB down.

JL77's picture

I wish Benchmark all the best with the new amp, though would suggest that the true, unweighted, low THD, above-the-noise, systemic dynamic range using today's best-case audio program and electro-mechanics (from mic-source to home-delivery) is around 115dB. The good news is that systemic DR has grown predictably at an average rate of 0.8dB per year since wax cylinders, so by 2040, or so, DR should no longer be a limting factor, though I wonder if our ears really want to be in an environment with peaks above 120dB SPL, which some studies cite as the threshold of pain.

1likeh1f1's picture

At last year's RMAC, I had a great conversation with Cookie Marenco concerning the most significant challenge for hi-rez audio. 

I believe the real issue has nothing to do with hardware.  Rather, hi-rez is always going to be hampered because of the inability to access proper source materials (master tapes, etc.) of the music that we have in our libraries today (what we actually want to listen to).  The masters of so many recordings we've all grown up with are gone or damaged.  So, there is no way to create a 24 bit (or higher) resolution recording because all that remains is a lower res version to engineer from. 

Cookie told me several really interesting stories about how lacksadaisical, haphazard the record companies have been with masters over time for most of what artists have recorded (lack of cataloging, ownership changes resulting in lost tracking, inappropriate storage, etc.)  She mentioned that sometimes the artists themselves would damage the masters through neglect or even artistic conceit.  That, with fires and other destructive things that occur through the passage of time, narrows down tremendously the historic recordings available for hi-rez issuances.  For new recordings, it is certainly possible if the all of the recording/engineering dominoes fall right - but, for the lion's share of the music we love most in our libraries, hang it up. 

At the end of our conversation, Cookie said to me, "Never let go of your CD's and albums.  They're the best you're ever going to get in most cases."

Good advice.

seavan's picture

At a risk of sounding errogant, what does 24bit resolution have to do with 130dB dynamic range?

John Atkinson's picture

seavan wrote:
At a risk of sounding errogant, what does 24bit resolution have to do with 130dB dynamic range?

Each 1-bit increase in bit depth is equivalent to lowering the noisefloor by 6.02dB. The 16-bit CD therefore has a noisefloor at -96dB and change, hence cannot resolve information lower than -96dB. (This is without noiseshaping and dither.) By analogy, an amplifier with a maximum S/N ratio of 96dB would, if its noisefloor were random, have 16-bit resolution.

The Benchmark amplifier has a claimed S/N ratio "approaching 130dB," which, at 6.02dB per bit, is equivalent to more than 21-bit resolution.

This is superb performance, given that the practical limit of DAC resolution is currently also close to 21 bits.

John Atkinson

Editor, Stereophile

seavan's picture

John, thanks a lot for your response! I should have stated my question more clear. Both 24 bit and 16 bit have the same max output volume. The noise floor is better with 24 bits in a sense that the re-created analogue signal is "smoother". I do not know if this noise is audible in a way different from a 'regular' noise. I do not see, however, how it is related to the power amp output volume level (sound pressure level) that I think is discussed in this article.

Johnnyjajohnny's picture

No, the waveform on 24 bit is not smoother than on 16 bit, because it's not square waves coming out of the output on a DAC/CD player to begin with. Many people this is the case, and they also believe that bit depth means resolution, but it doesn't - it's only dynamic range.
Please read the following magnificent article that explains exactly what the misconception is and what bit-depth really means: