This review page is supported in part by the sponsors whose ad banners are displayed below
Our baseline was that Fred hadn’t spent years working on his speaker to end up with what we heard. But since he'd been quite satisfied with the sonics in his place, the first thing to our mind was the preset correction curve in the Deqx. Fred had left us the Deqx calibration software install CD. Thus our next step was to install it on a laptop, then connect the laptop to the Deqx. The Behringer measurement microphone was set up in the listening position and connected to the Deqxl. We did not have any experience with this software but owned a TacT room correction system years ago whose basics were similar. First the type of loudspeaker gets entered whereby the Deqx can be used as an active crossover. It took some trial and error to get the right setting for a single driver but we managed. Then the Deqx determines a base line by emitting a series of clicks though the left and right speakers which are picked up by the measurement microphone to load into the calibration software. In our case the visual representation of the in-room measurement was not unlike the Himalayan mountain ridge with a severe dip from ~750 to ~1250Hz. Lowering peaks is no problem for room correction software but filling out dips is. We spent considerable time trying to smooth the curve with several corrective options. No matter how hard we tried, we did not manage to get it to our liking.


A few weeks later AER’s Filip Keller took the train from Stuttgart to pay us a visit. He was not only interested to hear his latest driver mounted in our Pnoe horns but also to give the Axjet a curious ear. Over the weeks before we had intensive contact with Filip about the options to getting the Axjet to work properly. At his studio a pair of the British speakers was working quite well. The biggest difference was the means of correction. Filip was working with Audiovolver to develop a small version of their high-quality correction device. In his backpack he'd brought a sample of it but not the accessories needed to do the measurements and correction. But he completely agreed that the Axjet cold do much better than it did for us. As a result he introduced us to Peter Schippers of Audiodata in Aachen, Germany. Peter has worked in audio electronics and electroacoustics for more than 30 years. Over the years he has tuned hundreds of rooms and systems all to the fullest satisfaction of his clients. Together with the Frauenhofer Institute—yes, the same brainiacs who brought us MP3—he developed the Audiovolver II. The Institute is responsible for its mathematics, Peter for the sound quality since music and whiz kids aren’t the best bed fellows.


With Filip as go-between we contacted Peter who offered to pay us a visit and set up an Audiovolver II for the Axjet in our room. Peter turned out to be one of the kindest most knowledgeable men in acoustics. His unique track record of helping people get the best from their systems not only led to a wealth of anecdotes but enables him to perform magic within hours. He came with a few suitcases of magical stuff. Not only an Audiovolver II emerged but a measurement microphone an laptop. Peter’s modus operandi is to visit a customer and perform an initial in-room measurement not unlike we described above before loading a correction curve into the Audiovolver II. Unique here was that Peter retains a copy of the initial measurement and correction. The customer now can contact Peter by phone or mail when something needs changing. Peter can adjust the correction curve and the customer can download it onto a USB stick. That stick is then inserted into the Audiovolver which automatically loads the new curve.


In our case Peter ran the measurements not only with his own software but also his well-trained ears and a clever app on his iPhone. These three data streams were combined in the first correction curve. Just like our own Deqx measurement, Peter’s showed the severe dip around 1kHz. To be more precise, the dip spanned a full octave from 600Hz up, not the most insignificant range. As mentioned before, lowering a peak with correction software is feasible unlike filling out a suckout.  Hence the only solution was to lower all of the output except for the depression range. This meant at least 10dB of wholesale attenuation to correct most of it. Some attenuation was saved lest the signal get too low. Peter then suggested having the correction curve in general follow the overall frequency response of the Axjet, that is with a gentle slope from low to high frequencies. In the graph you can see the measured and correction curves. Next to the obvious dip of the in-room measurement, the drop in the low frequencies is clear. Below 70Hz there wasn’t much going on.


It took several attempts to arrive at the most satisfactory correction. Peter not only adjusted the imperfect response but enhanced the impulse behaviour. The result was that the virtual image was raised from near the floor towards ear height. Switching the correction on and off had a strange but wonderful ping-pong effect in the vertical plane. By the time it was completely dark outside, we had been left with a system that sounded many times better than before. Now we could asses the true capabilities of Fred’s design. But would this lend itself to a formal review of a ready-for-commerce product? We didn’t think so. The Axjet of our meeting was still in the beta-testing stage. Hence we decided to write it up as an industry feature. Why not just cancel the review? Because we love this stuff.


The combination of a well-dialed Audiovolver and Axjet played music quite to our satisfaction. What was initially thin and shouty was gone. So was beaminess. Listening well off-axis was no problem and phase and timing were good enough to give a sense of stereo even at 90°. Overall the Axjetas we had them set up were able to paint a nice virtual image between and around them. Bass extension was limited but the lower mid quality made up a great deal. Many box speakers do less. A quality about the AER drivers we like is their ability to vanish completely from the scene. This was present with the exception of frequencies around 1000Hz. Here the drivers very obviously ‘showed up’. Fred had to definitely adjust something here. That awareness of the speaker being there was a metallic edge. We managed to lower its presence by sticking an ASI sugar cube just in front the afterburner on the aluminium midrange horn.


The time we spent with the Audiovolver was limited as Peter needed his machine back. Nevertheless we listened for long hours using various sources like CD, streaming via the La Rosita Beta and vinyl. A final setup included a WLM Duo subwoofer. Peter deliberately did not cut the signal at 50Hz per Fred’s advice but let the natural roll-off do its work. That meant the signal output from the Audiovolver contained frequencies all the way down. Using the pre-out of either the Devialet or Audio Note Meishu and sending that signal to the WLM active crossover and subsequently a cheap Behringer A500 stereo power amp added the lowest two octaves to the sound. Now we were talking!


When Fred accompanied by his son Fred came to pick up the Axjet we had quite spirited conversations on audiophile expectations versus PA, music room versus by-the-pool use. But how to report our findings to a man who had invested thousands of hours and £££ into his dream? The only way was to be honest and conclude that the Axjet was almost there. Fred took the message very well. We saw in his eyes that he was already thinking of a solution to our problems. Just as we were getting a little used to the empty spaces which the friendly red horns had left behind, we received word that Fred had found the culprit for the frequency response dip and metallic sound around our problem area. It was the gap between the driver’s whizzer edge and the start of the afterburner’s tube. By closing the gap between the two, the air leakage apparently causing the problem disappeared. Eureka! He was also working hard to build a matching pair of subwoofers so that a complete set of Axjet would cover the full audio spectrum. Here is Fred’s latest on his progress:


"You rightly pointed out that we were elated to solve the 1.2K dip and now have a completely different picture in the mids. Although I still may build a tapped horn-bass speaker in the future (probably a single central one using Filip's AER Studiobass), I have performed open-heart surgery on the Axjets here and extended the first part of the exponential rear horn by 1200mm to achieve a half wavelength at the 50Hz cut-off following your observations in regard to the length of the Pnoe horns. When I looked back at my XLS calculations I could at first not understand why the bigger 40Hz horns we have built sound great at low frequencies but the Axjet were rolling off early with a peak at 80Hz. I suddenly realised the problem. I had used the same expansion co-efficient as the bigger horns where the length developed a half wavelength at 40Hz. When the Axjet parameters are put into this formula, it leads to a horn that is just too short to achieve the required bass. I should have experimented more but was persuaded to go straight to CAD and machine the foam master.


"I now have a completely different sound in the bass horn with a modified co-efficient for the calculation and managed to re-engineer the Axjet without altering the qualities of the afterburner or the overall shape. I am happy to tell you that we now feel very confident that the Axjet will work happily without any extra EQ. For a perfectly flat response Audiovolver can provide a solution with much more ease now that the dips have been dealt with. There is a comfortable warmth and resolution to the bass that feels quite natural, not as full and flat as a full woofer system but an entirely different picture to your audition. I am sorry that I allowed myself to rush things but as soon as we get the chance—perhaps when we take a modified pair to Filip—you will get a chance to hear them properly! So in summary we plan to market the modified version for general use with cables and a starter amplifier (DDA 100) so the user can simply plug in a laptop. They will come in their jackets and flight cases of course and be available in any RAL colour. We can supply the Audiovolver Micro for flatter response."

Corrected response with the Audiovolver of our original loaners.

Here ends our journey with these remarkable speakers for now. It led us to three very special people in the audio realm. Fred with his extensive history in the kind of music that gave our youth its flavour and still does; Filip who took the effort to come visit us and his wonderful drivers; and last but not least Peter with his extraordinary acoustic skills and experience. All three deserve a follow-up in these pages...
Axjet website
Audiovolver website