September 2010

Audiolics Anonymous Chapter 131
Headphone Potential Realised
A review of the Smyth Research Realiser A8.
Article By Dr. Bill Gaw

  Oh lucky me!! Over the past three months I've been able to solve two gremlins which have bugged me and my musical listening pleasure for 30 plus years; equipment vibration with the Black Ravioli Footers and AC noise with the Pure Power PP 2000 units; both of which lead to distortion. Now the third gremlin in my life has also been cured: being able to listen to headphones with imaging that is not only out of the head, but comes extremely close to music reproduced by the two systems I've tested it with.

Over the past 30 years I've owned at least four high end headphones, including electrostatics from Stax and ElectroVoice, and dynamic units from Grado and Beyer Dynamic, and an in-the-ear pair from Etymotic Research, and have ended up selling each at a great loss as I could not stand how they imaged. All, no matter how accurate they are at reproducing frequency and dynamics, have the same two problems; they put the whole hall into the space between your ears, and the chest thrill of deep bass is missing.

While there have been several attempts by various companies to ameliorate these through signal processing, all have failed to one extent or another. The system that comes closest is called "Dummy Head Recording", where two microphones are placed in false ear canals of a Styrofoam head. I actually have several of these, mostly produced in Germany, and they do somewhat image outside the head. While this does approximate where microphones can be optimally placed for playback imaging through headphones, and do give a panoramic feel to the sounds, it does not alleviate three problems:

1. They can only be played back through headphones as over speakers they sound constricted.

2. They only correct for one of the causes of image problems, the head transfer function, which is not the same for the Styrofoam head as yours, and do not correct for your headphone response.

3. They do not allow for a stable image in space, as one moves one's head the image moves with it. Normally when we move our heads a sound remains localized in the same point in space. With headphones it moves in the same direction.

I don't know whether you've ever looked at other people's ears, but if you do you'll notice that the external ear shape, distance between, and tissue construction varies not only from person to person, but also each ear on each person is different. It is the external ear, shape of the head and body that changes how the ear captures frequencies, with the difference in the time of arrival between the two ears and volume of the signal determining the location and distance of the sound. This is called the "Head Related Transfer Function". Both the shadowing of the head which decreases the total volume of the sound and the relative volume of various pitches, and the time of arrival of the sound to each ear affect our perception of both positioning and distance of the source. Thus each human, and each ear, actually picks up and perceives sound differently, but through the use of our other senses, especially sight, in childhood, one learns how to correlate the sounds with the environment.

This presents multiple problems, both with using loudspeakers and headphones, in presenting a soundstage which approximates the original performance. With loudspeakers, we have anywhere from one to many speakers in the environment that give cues to the brain as to placement of the voice or instrument in space, with more speakers producing a closer approximation.

One actually has to relearn how to position an image in space when using multiple speakers reproducing multiple microphone feeds as the only true image is produced by the speaker itself and any image in between is actually a phantom with the brain imagining it in open space. This is dependent on how the two ears pick up the distance and frequency differences and how they are perceived by the brain. I actually have a friend who prefers mono recordings as his brain has to work hard to perceive stereo imaging. At least with speakers, one can use the external ear to perceive the placement.

Not so with headphones, though. Since the headphones are either applied over the ears, or actually placed inside the ear canals, one loses all the external ear cues that allow the brain to perceive space. Thus all the sound ends up in the head.

Wouldn't it be great if someone could find a way to allow these cues to be presented through headphones so that the original placement of the sounds in space could be reproduced, plus allow the image to stay in the same place in space, and be able to get back some of the other body bass cues?  Gone would be the expense and hassle of setting up a listening room with expensive loudspeakers, cabling, amplifiers, room treatments, room, etc. And if you have a superb system, wouldn't it be nice to be able to perfectly reproduce the sound of that system through headphones any place or environment your little heart desires. Then wouldn't it be nice if you could reproduce anyone else's music system and environment any place you like, and let them pay for the expense of the whole system. Think of all the money you could save!! Well, there's now a way.

About six months ago I read an article in Widescreen Review about a new method of digitally adjusting an audio signal in such a way as to almost perfectly reproduce those cues that your own ears produce. The inventor is Stephen Smyth, Ph. D., who was the developer of the algorithms that are used by DTS for their cinema and home audio playback systems, and subsequently our home systems. He went on to develop several other algorithms, and has come up with  a method of almost perfectly recreating what our ears perceive in open space through headphones that he has called SVS for Smyth Virtual Surround. Subsequently he designed a unit which can be used for up to 8 channels that he has called the...

Smyth Research Realiser A8
Until now, the company has specialized in marketing it to professional movie and recording studios, and has been very successful, with one company buying 29 of them. This allows the studios to have one superbly set up sound stage (which can cost 7 figures when done right) with several workstations with only sound consoles using the Realiser as a virtual soundstage rather than a complete studio with its great expense.

The unit is 10 x 10 x 3 inches, and on the rear it has eight RCA multi-channel inputs to be connected to your pre-pro's outputs, eight RCA outputs which are used only for setup, which connect to your pre-pro's direct inputs, two RCA outputs to the main headphone amplifier and two outputs to a tactile speaker system for either a seat shaker or a subwoofer. In addition there's an optical digital output for an external D/A decoder, and a USB connector, an input for a 9 volt wall wart, and an input jack for the headphone position tracker receiver.

The front has a small display, an SD card input, multiple control lights, microphone input and headphone output phono jacks and a USB power jack for the position tracker. It uses 24-bit/96-kHz A/D and D/A converters on a high quality circuit board with op-amp amplification.

The kit contains the main unit, a head tracker which attaches to the headphone's band and a headphone tracker reference piece which is placed centered in front of the listener, a set of in-the-ear microphones with multiple foam ear canal blockers and cable, remote control, and even a Stax SR 2050 Headphone with its amplifier box. One can also order the unit with the Stax SR 4040II headphones. They prefer that the unit be used with Stax products but certainly other top of the line headphones may be used.

The manual, which can be found on their site, is at present 90 pages long, but is written in a very understandable fashion for so complex a unit. The first 24 pages are on setup, the next 12 basic running of the unit, and the last half is on all the tricks that can be performed. Thus it took me a while to comprehend all of the things the unit can do for your musical pleasure.

Setup is somewhat complex but intuitive, Mr. Lorr Kramer of Smyth Research, (who graciously took two days out of his annual vacation to demo the unit) was able to get the unit operating in a few minutes and have the first measurement set of my ears done in about 15. 

The measurements consist of the following steps:

1. Speaker Angles. One measures the angles between the listening position and the speaker in relation to the center front speaker which is at 0 degrees azimuth. For instance the left front will probably be around –30 degrees and the right front at +30 degrees. One enters that data for each speaker.

2. Calibrate Speaker Volume. The microphones are placed in the ears such that they are flush with the opening and the included foam pieces block the canal. The unit gives off a signal that adjusts each speaker's volume such that it is in the optimal level for the measurements.

3. Speaker PRIR Measurement. The Personalized Room Impulse Response measures the information given out by the entire system, including the room, and how it interacts with your external ear. The unit sends frequency sweeps through all of the speakers, and asks to have the listener turn his head towards each front speaker in turn to gather information on head response. This is then saved.

4. HPEQ Measurement: Headphone Equalization. To get the interaction between the output of the headphone and the external ear, one places the headphones, and the unit does sweeps for each ear. This is then saved. While each person will have to do a PRIR measurement for each room one wishes to emulate, only one Headphone Equalization must be done per person and individual headphone.

5. Data Saving. If both measurements were accurate, one then saves both to one of 64 internal memories or an infinite amount of external memories that can be saved on the SD card. Thus an infinite number of individuals in an infinite number of rooms can be stored.

6. Presets. The four most important measurements can be stored in four presets on the remote for easy call-back.

So what have we accomplished with all this?

1. The Tracker senses which direction your head is moving up to the angles measured, and automatically rotates the signal such that the channels stay where they are in the room. Thus for instance, move the head to the right and the center channel moves toward your left ear, thus giving the sensation that the original image stays where it belongs.

2. The PRIR measurement records how your external ear corrects for sound placement in the room and also measures the sound qualities of your music system.

3. The HPEQ measurement corrects for how the Headphone interacts with your external ear and corrects any frequency anomalies.

4. Numbers 2 and 3 therefore theoretically allow the headphone to give a perfect reproduction of what your entire system in your room sounds like to your ear-brain system thus negating the in-the-head experience, the frequency anomalies of most headphones, and allows a stereo pair of headphones to actually reproduce up to 8 different channels of information in space.

5. One can then take a source, the Realiser box and the headphones and reproduce any system's sound anywhere there are AC plugs for the equipment. Theoretically you could go to the best sound system in the world, spend about a half an hour doing the measurements and reproduce that sound while sitting on the John.

6. In my room and many others, with video, the speakers are below the screen. Thus my brain has to work to place the sound at the proper level. With this system, the headphones will automatically raise the height of the sound to match the screen just by pointing your head directly at it.

For the first week, I did multiple in room responses, first to become familiar with the unit's instructions, and then to see if varying its parameters improved on the playback. One can adjust the time of the sweeps from 3 to 12 seconds, with the former only good for quick setups, and the latter for final measurements. One can also do one to many sweeps per speaker.

Using the 4 presets, one can instantaneously switch to different response tests to hear the variances and which appears best. While one would think the longer and the more one does, the better the outcome, there seemed to be a flattening of the curve at three sweeps per speaker, which on an 8 speaker system took about 15 minutes, about the same as room equalization with the  Audyssey System. Beyond that I could hear no difference in the sound between the trials or in their relationship to the original speaker sound.

In addition, one can theoretically purchase only one super high end stereo amplifier, two speakers and cables, place the speakers in each of the best places in the room for sound, and capture the results to the unit as a complete surround system. Or one can leave the speakers in their positions, and rotate the head through the 360 degree arc to reproduce a 7.1 system. Likewise, one can move a subwoofer to several positions in a room; take measurements, and record, which should decrease significantly measured standing waves and bass frequency anomalies.

Further, the unit can correct for reverberation and other room anomalies for each speaker, and correct for volume of each channel. Delay of each speaker can be adjusted for those rooms that don't allow for the speakers to be equidistant from the listener.

For those individuals who can still hear a difference between what the unit perceives as correct reproduction of your loudspeaker, the unit can also run pink noise of varying frequency bands so that one can hear where the emulation varies from the original, and adjust the volume of the pink noise so that the speaker and headphone have the same relative volume, thus improving on the correlation between the two.

The unit also has outputs for two tactile transducers which are normally applied to one's seat, thus emulating the deep bass vibrations one normally feels from speakers.

Thus, theoretically, one could connect a multi-channel source directly to the unit without intermediate preamp-processor. Once the room measurements are done, to listen to a recreation of your entire system would require only your sources, six high quality analog interconnects, high end AC cords for sources and headphones, the unit and headphones, and possibly some tactile stimulation.

So how well does the unit work. In a word, outstanding!!!

First, the between the ears in the head symphony orchestra effect is completely gone. Imaging is exactly where it should be in space. I have an SACD of a group of seven singers and in one piece, each person is placed in a separate speaker. With the headphones, they are placed at exactly the same point in space as that speaker would be, and even with head movement, the image remains stable in the horizontal plane.

Second, by tilting the head back, so that I am looking directly at my elevated video screen, the front channel sound is directed at the screen, not below it where the speakers are placed, thus actually improving on the audio-video correlation.

Third, with my Stax 404 headphones, even without adjusting the frequency bands of the loudspeaker emulation, the image produced is extremely close to what my speakers produce, except for the deep bass information. Adding the Tactile transducer I have in my listening chair improves but doesn't completely eradicate this. On the other hand, by turning on my subwoofers in the room, I can almost get the bass to match what the system normally produces. The sound produced by the headphones is actually cleaner than obtained by the speakers, but with a slight decrease in mid and upper bass information. In addition, there is a slight loss of ambiance information, which is probably due to the signal having to go through the A/D and D/A processing.

Fourth, as an experiment, I used one of my main front horn loudspeakers and by rotating my head and body, rather than moving the 800 pound speaker around the room was able to emulate 7 loudspeakers in a surround configuration in my room. While the sound didn't recreate the room acoustics, it did do a wonderful job in throwing a concert hall feel.

Fifth, while my system is fairly quiet for a 7.1 system with multiple tube amplifiers and several amplifiers and projector with fans, being able to shut off most of my equipment and have the headphones over my ears, normal room and equipment noise was practically eliminated, thus dropping the noise floor. This actually allowed playback at lower volume through the headphones.

Sixth, and probably most important, while I haven't been able to get to the point where I can honestly say that I can't tell the difference between listening to the headphones and the speakers in direct instantaneous A-B testing, the emulation is close enough that I've had my entire system shut down for almost a week now and have been very satisfied listening only to the headphones.

Seventh, since the multiple tube amplifiers have been off, with the heat wave that's been here over the past week, I've been able to listen in relative comfort.

Eighth, I have now been able to collect information on one other system based on the $70,000 TAD Reference One loudspeakers. Thus I can now listen to either his or mine without moving from my listening chair. The more systems I evaluate, the more money I'll save.

Ninth, I now have files on my audiophile friend's ears that he could then use to listen to his system in other venues if he had my headphones. While this doesn't do them much good, he theoretically could come over to my house and directly compare the sound of his system to mine.

Ten, and maybe the most important, I now have a Wife Acceptance Factor (WAF) of 100%, in that she hasn't complained of the noise emanating from my listening room, as there isn't any.

So what are the negatives?

One, rather than having a separate box, it would be wonderful if some producer of pre-pros or receivers would license the system software for inclusion into their boxes similar to what Audyssey does for room and speaker correction.

Two, the signal has to go through A/D and D/A conversion at 24-bit/96kHz. Hopefully they'll come up with the idea of an HDMI input which will be able to work with Blu-ray's 24-bit/96kHz 7.1 PCM output. Then there'd be only one D/A process, with the player doing the Dolby or DTS decoding and the unit doing the D/A conversion.

Three, the bass thrill is missing, or with tactile transducers doesn't quite mimic a good subwoofer as it only stimulates the butt and back. Maybe someone will produce a suit that fits over the body that will help this.

Four, only one person can listen at a time. While the Stax has outputs for two headphones and the unit also has two, yet only one emulation can be produced at a time. Thus, one would need a unit for each person.

I'm digging at straws now, so I'll stop. I'm so happy with the unit's ability to recreate my and other's systems at any time in any place with AC, that I've purchased it. Considering its cost at $3360 which includes a $760 Stax headphone and amplifier, the unit is a steal!!! Get one before they realize the mistake they've made in the pricing. Further information can be found at Smyth Research and the units can be purchased through directly through Smyth Research, from Glenn Poor's Audio/Video, and from several dealers in Japan and Europe.