HeadRoom Supreme headphone amplifier
It was a flight attendant. "You're annoying the other passengers," she said, "Your stereo is way too loud."
I was embarrassed, of course. I was also pissed. I was also intrigued. Could the Brahms leaking from the Sennheiser HD560 Ovation headphones be audible above the sound of air rushing past the fuselage and engines turning kerosene into carbon dioxide and water? I tried listening to the headphones with them in my lap. I tried holding them in front of me about as far away as I was from my all-elbows neighbor. Just a faint rustling noise could be perceived above the 100dB low-frequency ambient background.
My listening time had already been partly eaten into by the new blanket rule forbidding the use of "approved electronic devices" until the airplane was well embarked upon its voyage. I realized what was happening. Rival headphone amplifier manufacturers were trying to abort my forthcoming review of the HeadRoom. And well they'll need to, I thought, as I took advantage of the opportunity by asking the stewardess to bring me another beer. (American's flight attendants could usefully learn from British Airways in the art of keeping passengers well supplied with alcoholic beverages—"Boozehounds Fly British!" I must've read somewhere.)
Wait a minute. What other headphone amplifier manufacturers? Especially manufacturers of cute little portable battery-powered headphone amplifiers? Yes, the Melos SHA-1 has been selling very well since Corey Greenberg reviewed it for Stereophile in 1992 (footnote 1), but the bulky tubed Melos is strictly a housebound creature. Grado sells a battery-powered amplifier to go with its HP series of headphones, but the amp costs $750—too rich for my blood! And mail-order musician catalogs have always offered pro-audio headphone amplifiers that look like they've escaped from the early days of solid-state. But mobile high-end headphone amplification is a rather deserted play area.
Which brings me to Tyll Hertsens' unique HeadRoom amplifiers.
The quest to make the headphone listening experience more equivalent to normal speaker listening is not new. Back in the '60s, the late Ben Bauer of CBS Laboratories devoted a lot of study to various equalization, interchannel crossfeed, and time-delay schemes intended to work the trick (footnote 2). Indeed, an article in the November 1962 Audio, authored in part by engineer John Eargle (better known today for his excellent-sounding Delos recordings), described a DIY device based on Bauer's crossfeed circuit. The amplitude responses of the Bauer circuit's two channels and the time behavior when fed a signal to just one channel, calculated using a circuit-analysis program, are shown in fig.1 (footnote 3). The driven channel has a degree of HF lift applied to it, while being crossfed to the undriven channel with an approximately complementary amount of HF cut so that the total power in the treble remains about the same. In addition, the crossfeed signal to the undriven channel is delayed by about 0.4ms below 1kHz.
Fig.1 Calculated stereo frequency response, compensated for insertion loss, of Bauer stereophonic-to-binaural network from JAES, April 1961, with delay to undriven channel also plotted. Input is swept tone input to left channel only (2dB and 0.2ms/vertical div.).
Why should it be necessary to so drastically mutate a stereo signal? When a binaural recording (one made with a dummy-head two-channel microphone) is auditioned on headphones, all the sounds appear to come from outside the listener's head. When a conventional amplitude-stereo recording is listened to via headphones, however, musicians and singers seem to be sitting on a clothesline stretched from ear to ear inside the head. "The instruments form a 'musical hat' on the observer's head," was how Bauer described this phenomenon; his EQ, time-delay, and crossfeed are intended to eliminate it by synthesizing the natural ear-to-ear crosstalk signals that a listener experiences when hearing a sound source from the front. (Sound to the side of one ear is heard by the other ear delayed by up to 0.7ms; it is also rolled-off in the treble due to the acoustic obstacle presented to the sound wave by the head.) This processing is intended to render the headphone listening experience more natural, more comfortable.
Fast-forward a quarter century. Any number of researchers are using Digital Signal Processing (DSP) to work the inverse trick: make listening to conventional stereophonic recordings via a single pair of speakers resemble the binaural experience, with sounds coming at the listener from all directions. (This is also something that was first suggested by Ben Bauer; QSound is the commercial realization of one of these algorithms.) Yet I thought Bauer's stereo-binaural quest had been long forgotten—until I saw an advertisement from a new Montana company, HeadRoom, in the March 1993 Stereophile.
The HeadRoom dual-channel headphone amplifier can operate conventionally, amplifying its inputs and driving a wide range of headphones of differing impedances and sensitivities. But when its unique Process switch is thrown, it feeds each channel through to the other with an amplitude and time delay that are carefully controlled with frequency. In this manner, it parallels what the obsolete Bauer circuit did: synthesize the around-the-head crossfeed information that a listener hears from frontally placed sound sources.
The HeadRoom's block diagram is shown in fig.2. The input impedance is defined by 100k shunt resistors; these also tie the inputs to ground in case the DC-coupled amplifier is turned on with a source component not connected to its inputs. With the Process disengaged, the signal passes through to the volume control, then the output op-amp, discrete complementary buffer stage, series 8 ohms resistors, and the ¼" headphone jack. (There is no 3.5mm stereo jack: this amp is meant to drive Arnold-sized cans.) A gain-setting resistor is looped around the output stage. With the Process switched in-circuit, each audio signal passes through Crossfeed EQ and Crossfeed delay networks and is summed into the opposite channel prior to the volume control.
Fig.2 HeadRoom headphone amplifier, block diagram. The numbers refer to the pins on the HeadRoom module. (Click on picture for larger image.)
Construction is to an excellent standard. The entire amplifier is housed in an extruded aluminum case. Two dual-AA battery holders from the English Bulgin company occupy the center-rear of the housing; the circuitry and front-panel hardware are carried on a U-shaped, double-sided printed circuit board that surrounds and is supported by the battery holders. On the right rear is a pcb-mounted AC power jack that takes in 5V DC from the supplied AC adaptor; on the left rear are two pcb-mounted RCA jacks, these recessed inside rear-panel holes so that the RCA connectors are securely supported by the HeadRoom's chassis.
Footnote 1: See CG's review in Vol.15 No.10, October '92, p.195, and my 1992 "Products of the Year" feature in Vol.15 No.12, December '92, p.7.
Footnote 2: See "Stereophonic Earphones and Binaural Loudspeakers," B.B. Bauer, JAES, Vol.9 No.2, April 1961, reprinted on p.373 of Stereophonic Techniques, edited by John Eargle and published in 1986 by the Audio Engineering Society. This excellent anthology also includes a 1977 paper (p.382), "Improving the Stereo Headphone Sound Image" by M.V. Thomas, Ph.D., which implements an actively realized version of the Bauer circuit using LM301A op-amps. (Is this the same Dr. Martin V. Thomas who did amplifier design work for the British Musical Fidelity company in the early '80s, I wonder?)
Footnote 3: A similar response diagram appeared in the January 1968 issue of what was then plain Hi-Fi News. I only include this information for pedantry's sake, however, as HFN/RR now only keeps a limited supply of very recent back issues. Back before I edited that magazine, its publisher decided to save warehouse space by trashing almost all its stock of back issues. It might just have been the fact that I embarked upon my current career because I was a regular and enthusiastic reader of HFN, then HFN/RR, from 1967 through 1976. Nevertheless, I regarded this act as a tragedy, given the superb quality of its technical articles. There was nothing any of the magazine's editorial staff could do about this historically wrongheaded decision.