Headphone frequency response measurements, conducted with a microphone in front of the headphone driver, much as speaker measurements were conducted, observed that headphones with a measured flat frequency response did not sound as flat as they measured. Comparing the sound of a flat measured speaker with a flat measured headphone revealed extreme tonal differences that started off a whole lot of investigation into, firstly, why they sounded different and secondly, how a headphone’s frequency response could be altered to make it sound like flat measured speakers.
An experiment was set up as follows:
1. A loudspeaker playing a frequency sweep was recorded by a high quality, miniature microphone in one of two types of room – either an anechoic chamber or an approximation of an ordinary room (see later) – and the frequency response was charted.
2. The same microphone was inserted into a subject’s ear canal and the speaker replayed frequency sweep was charted again.
3. It turned out there was quite a difference between the charted frequency responses of the two recordings.
4. It was postulated that the frequency response of the recording made by the microphone in the ear (see 2) could be altered by pre-equalization to ultimately match the shape of the frequency response of the recording of the same microphone when not in the ear (see 1). In theory, pre-equalised headphones could then sound the same as listening to speakers.
5. This gave rise to a target measured frequency response for a headphone to sound like a flat measured speaker i.e. if the headphone had a measured frequency response that looked like the target response, it should sound flat when reproducing a recording that had been mixed with speakers in front of the mixer, and should also sound as if one was listening to speakers in front of him/her, rather than via headphones with drivers parallel to the ear pinnae.
6. The concept of pre-equalisation of headphones was thus born.
Pre-equalisation could either be mechanical (i.e. the driver frequency response was manufactured to behave that way e.g. the AKG 240DF – not so easy) or electrical (which should be cheaper, easier and field-adjustable), and meant that, although the headphones now had a frequency response that had been altered to something that looked decidedly non-flat when measured, it reproduced the sounds coming from a sound source with the same frequency response at the ear canal as if recording and replay over headphones had not been introduced into the chain i.e. the headphone replay should now sound the same as sitting in the room and listening to the speakers.
Two main theories of the correct pre-equalisation curve were forwarded. The first, called free-field equalization, suggested that the above experiment be conducted in an anechoic chamber (like an open field, free of reflective, absorptive and refractory surfaces). So, to reiterate, a free field equalized headphone is designed to sound like the reproduction of speakers in an anechoic chamber. Although an anechoic chamber is more reproducible as a standard, it was argued that nobody listens in an anechoic chamber (and indeed, most listeners find even speaking in an anechoic chamber uncomfortable) and a reasonable approximation of a standard listening area be used to conduct the above experiments. This was called diffuse field equalization.
There are many things that alter sound between the release from the sound source and arrival at the ear canal. Reflections, diffraction and absorption from objects in the listening environment, reflection, diffraction and absorption by the head, hair and ears all contribute to alteration of sound before it reaches the ear canal. Diffuse field equalization, as mentioned before, is an attempt to make the replay of a recording via headphones sound like you are listening to the same recording through speakers in a non-anechoic room. Experiments were also done such that headphone users were asked to equalise various sharply limited frequency bands’ playback on headphones until they had matched the loudness of the same playback through speakers and with headphones removed. A good correlation was obtained between this method and the probe microphone recording method mentioned above, apparently validating the theory.
The direction of sound (from the front in a reverberant field) with speakers is far removed from actually injecting the sound directly into the ear canal.
Stax’ president at the time, Naotake Hayashi, pondering this problem, decided to use some brilliant lateral thinking, possibly because Stax had not, at that time, successfully created a diffuse field equalized headphone. Also, any electrical diffuse equalization would have to be a custom-designed unit for a particular headphone. He decided to create a new headphone that coupled its own reproducible miniature room (complete with uneven diffractive, reflective and absorptive surfaces) to the listener’s ears called the Stax SR-Sigma Panoramic Earspeaker. It had headphone drivers that fired sound from anterior to posterior and bounced off irregular “wool” into the canals, instead of being directly injected into the ear canals. This is the same as creating a miniature diffuse field room for each ear as well as having "speakers" that fired sound from the front, rather than straight into the canals, much as one would experience in a normal listening room when listening to speakers. It was partially successful, but listeners either hate it or absolutely love it. Personally I love it, but they were inefficient headphones and sounded quite rolled off at both ends of the frequency spectrum. They were also huge and very odd looking.
Better drivers than the original Sigma units (which were the same ones as used in the later Lambda Semi-Panoramic earspeaker) improve the frequency extremes, allowing the merit of Naotake Hayashi’s theory to finally shine through (e.g. the very rare Sigma/404 hybrid or, to a lesser extent, the Sigma Professional).
Stax later decided, instead, to bite the bullet and build custom equalisers to electrically equalize their latest headphone range to provide individual target diffuse field responses for each of its various headphones (the ED-5 for the SR5 normal bias headphone, the ED-1 and SRM-Monitor for the Lambda Professional high bias phone and the ED-Signature for the high bias Lambda Signature). The design of the equalizers was a collaborative effort between Stax and Dr Gunther Theile at the German IRT broadcasting network. The headphones could then be less bulky than the Sigma (as the room effects didn’t have to be physically created by the headphones) and therefore more fashionable (The Sigma definitely had a style only a mother could love). Again, reactions to Stax diffuse field equalized headphones literally polarized listeners into “hate it” or “love it” camps. I would guess that, economically, this proved to be a dead end, and Stax has never publicly mentioned any research into diffuse field equalization again, and the Stax Sigma and Sigma Professional earspeakers were sadly discontinued.
As mentioned above, the three ED diffuse field equalisers were designed for three different Stax phones (the ED-5 for the SR-5 normal bias, the ED-1/SRM Monitor for the Lambda Pro professional bias and the ED-Signature for the Lambda Signature professional bias). The ED-5, ED-1 and ED-Signature were placed between the source and the headphone driver and are connected by way of RCA cables. The ED-1 matched the construction and size of the SRM1 Mk2 and was finished, like those units, in either black or silver. The ED Signature matched the black of SRM-T1. The SRM-Monitor incorporated an ED-1 and an SRM1Mk2 Professional into one larger package and was finished in either black or silver, and had both RCA and XLR inputs.
The ED-Signature would most likely also closely match the 404 and Lambda Nova Signature. The ED-1 equalisation (in my case, provided by a very rare SRM Monitor) sounds rather nice with the Lambda Nova Signature and surprisingly good on the Omega 2 Mk 1, despite being the wrong equalisation for the latter. The upper midrange/lower treble, in particular, sounds quite a bit flatter and the low end remains in good balance with the mid and high. As Bill Sommerweck said in his review of the ED-1 in the April 1989 issue of Stereophile, track 9 on Stax' own “Space Sound” CD changes from objectionable (without the equalizer switched in) to quite listenable with the equalisation switched in. In my opinion, there is no magical out of the head experience, except when listening to the aforementioned CD, or the Ultrasone binaural tracks (i.e. binaural recordings). These are seriously spooky. Try them with someone who is not used to listening to headphones and see what happens when you cue up track 1 or 2 of the former, or the fireworks track of the latter. Sabine whispering in your ear - Oh yes! Shower spraying on your shower cap - OMG! Fireworks on the Ultrasone CD – duck!
Now, here is where things start getting weird. I had a listen to the SR-007 Mk 1 phones with the equaliser on and it also sounded great - it may just be happenstance, but I've never heard Miles Davis’ "Kind Of Blue" sound so wonderful and with plenty of lower bass (which even the SRM-717 doesn't seem to match). This is strange, because the frequency response for the ED-1 is absolutely flat in the bass and should neither increase or decrease bass presence, and the equalisation should not be suited to the SR-007 anyway. I would have thought it unlikely to be a de-emphasised treble spike on the SR-007 Mk 1, as noone has commented on that phone having a treble spike before. I don't know, but whatever, this pre-equalisation is not just scientific theory and sounds superb. To me, the sound has moved from lots of good hi-fi parts and to an organic and holistic experience. In my opinion, the Stax SRM Monitor is quite simply the single best piece of equipment I have ever purchased, and galvanized me to write to Dr Gunther Theile to congratulate him on his pioneering work (in conjunction with Stax) on this unit.
Edited by John Buchanan - 2/5/11 at 4:00pm
