Cones, Domes, and MQA
It has been many years since I have been to an audio show, so I was pretty excited to be invited by Archimago to attend the Vancouver Audio Show. I must say I had a great time! I was very fortunate to be able to sit in the sweet spot for most of the demos, including the MQA demo with the Tidal Sunray G2 loudspeakers and Burmester 909 Mk 5 amp above.
Rather than detailing equipment specifications and prices, I take the approach of how each exhibit sounded to my ears. As a reference, I compare to a sound reproduction system that has been calibrated for accuracy. My definition of accuracy means the frequency and timing response of the music arriving at my ears matches as closely as possible to the content on the recording. I wrote a book on the subject.[PRBREAK][/PRBREAK]
Speakers and rooms typically have the greatest influence on the tonal response for any given sound reproduction system. Therefore, my review is from this perspective. What do I listen for? If I were to categorize from bass, mids to high frequencies, I listen for smooth bass response (+- 3dB tolerance) and is it balanced with the rest of the frequency band? Does the bass transient response sound tight with no overhang (i.e. narrow impulse response)? Or is it blurry sounding (i.e. non-linear distortion)? Integrating subwoofer(s) without sounding boomy (i.e. peaky room modes) takes considerable skill and effort. I feel this is the primary reason why I did not see any subs on exhibit.
Listening to midrange frequencies, in addition to being balanced with the rest of the frequency range, (i.e. tone) does it sound smooth or nasally sounding (i.e. peaky frequency response)? Does the midrange sound distorted or harsh at program reference level (i.e. non-linear distortion)?
The show was dominated by dome tweeters. To me, the sibilance sound that our voices make when singing an s word is the hardest to reproduce accurately, assuming one prefers accuracy. It can range from real s sounding to overly sibilant SSSsss sound, to slightly rolled off ssshhh sound, to a ssszzz zingy top end at reference program level (i.e. 83 dB SPL C weighting, slow position on a Sound Pressure Level meter measured at the listening position).
Reproducing “s” sounds were further complicated by the number of turntables at the show. I have not seen that many turntables since audio shows in the 70s and 80s. Cartridge and the mastering/pressing of the vinyl vary with respect to how sibilant the sound reproduction is, which may mask what the speaker’s tweeters are capable of. Notwithstanding if the cartage vertical tracking angle was set up right in the first place. Kudos to the all of the exhibitors for carting and setting up the gear – having done it myself, it is a lot of work and hard to get right, especially being room dependent. Which brings me to…
It’s a crying shame man… that the room acoustics of hotel rooms are difficult as they are small, mostly square-ish rooms that typically sound boxy and/or boomy, no matter where the speakers are placed. Naturally, the larger rooms, or rooms with a favorable room ratio, tended to sound better than the boxes.
I would love to have heard these systems with the rooms neutralized to hear more of the system’s sonic characteristics. I am not talking about anechoic chambers either. Physical room ratios play a large role on how boomy and/or boxy a room will sound. Further complicated by every room at the show had different room ratios, much like our own listening environments.
Want to gain a better understanding of why we hear what we hear in small room acoustics? I recommend James (JJ) Johnston’s presentation on Acoustic and Psychoacoustic Issues in Room Correction, including the Power Point presentation to slide 31.
What to do for future audio shows? One suggestion is for the Vendors to agree on a target frequency response (i.e. tone) and have every system tuned to that response in all demo rooms. That way, I (we) would have a chance to actually hear more of the characteristics of the sound reproduction system that is not being influenced or masked by the room acoustics, or lack thereof. The thing is, the research has shown over 40 some years of objective measurements and subjective listening tests, that the target frequency responses are not only well known, but are also well correlated:
Here is an example of four target frequency response curves that result in the most neutral or balanced sound to one's ears at the listening position:
The point is that all four target frequency responses are close together gathered from 40 years of research and listening tests. One can read the details for free, including links to the research, by clicking on Look Inside and clicking on Recommended Target Responses in the ToC.
Room acoustics are frequency dependent based on room ratio, construction, and absorption. Given that top-end DSP loudspeaker and room tuning systems use 64 bit calculations, for all practical purposes, are completely transparent, so one would be listening more to the gear than anything else. I know right, a giant leap for many, but once one hears an accurately tuned system, both in the frequency and time domains, it is hard to go back to boomy, boxy, and/or nasally with poor imaging. Bottom line, it would simply be nice to hear more of these systems capabilities with less interference from the hotel room acoustic environment.
What about the timing responses of these speaker systems? From Stereophile, few speakers are time coherent. If you have not heard and lived with a time coherent speaker system before, how can one compare? To me, this is an area where speaker designers should optionally offer time coherent crossovers for their speaker lines for folks that want a more accurate representation of the audio content. There were a few time coherent speakers at the show. Time coherence takes design skill and is expensive to implement with an increase of passive XO parts or active XO with electronics and/or DSP software.
The timing response of loudspeakers also has a tie in with MQA, but I will explain later.
The ArtVibes Audio are interesting looking speakers that can be ordered with original artwork from select artists. From a sonic perspective, they sounded a bit boxy and somewhat rolled off. Most likely the room… I would love to drop a measurement mic and take some measurements at the listening position to see what I was actually listening to. I think folks would be surprised to see the variance in frequency response. The frequency response is unlikely to be smooth in the bass and typically has a boxy or nasal characteristic which is representative of the rooms influence on the frequency response of the system at the listening position. Nonetheless, they did sound pretty smooth. These are not time coherent speakers.
Totem’s Element loudspeaker sounded very smooth with good imaging. Not time coherent. Sounded to my ears on the warm side with a bit rolled off top end. I am wondering if I am hearing the natural roll off of the woofer as there is no bass XO on the woofer. Unlikely, but did sound on the warm side, I am not complaining. Very interesting design and implementation.
Audio Note loudspeakers in a less than ideal room = boom city. Really too bad. If you look closely, there is a Black Sabbath album under the amplifier. Should have played that to take advantage of the boom. The room was so boomy, I could not really hear the balance on the tweeter.
Wilson Audio Sabrina sounded quite smooth and balanced to my ears. Some room boxiness to the sound, but they imaged incredibly well. I attribute the excellent imaging to Wilson’s claim that the speakers are time coherent. Looking at Stereophile’s measurement of the step response, Figure 6, indeed shows time alignment. I am also impressed to see in Sabrina’s service manual a large section on room acoustics and speaker placement. I have used the Wilson Audio Setup Procedure (WASP) before in lieu of measurement gear with good results. One can never underestimate the value in properly setting up one’s speakers to get the best sound quality for any given room.
Devialet Silver Phantom sound very impressive. I would love to review these in my home environment. Big sound from a small speaker, with excellent imaging. Manufacturer claims they are time coherent, but I could not find anywhere on the net any actual frequency or time response measurements for this DSP speaker system. It would be interesting to see measurements of this speaker, especially their step response.
Kef Blade 2’s hooked up to Naim’s Statement amplifiers. The Statement amplifiers remind me of early Cray computers. The Blade 2’s sounded incredibly smooth to my ears, with very good imaging. Kef makes the claim that these speakers are “single apparent source”, but Stereophile’s step response measurements Figure 9, show that the speakers are technically not time coherent. The tweeter leads first, then the midrange, then the woofer. Compare the step response to the Wilson Sabrina’s above to see what I mean. Aside from that, the speakers measure incredibly smooth and that is how they sounded to my ears. They especially sounded good on the s’s sibilant sound. I remember they image quite well, but something about the depth of field did not sound quite right to me. After looking up the step response (timing) measurement linked above, I can now understand why the depth of field sounded slightly different to my reference. There is a step response measurement below of the reference time alignment that I am referring to in which we will get to a bit later.
The Davis Acoustics Renoir sounded really good, even in the smallish room they were in. Smooth and extended in both frequency extremes. These are not time coherent speakers.
A blast from the past, the infamous Speakerlab Super Sevens. When I was in high school, a long time ago, I built a version of these for a friend and they sounded fantastic to rock out on. Even though they are sitting on the floor in this demo, they have a tight acoustic suspension low end response, coupled with what arguably could be the smoothest top end of the show. S’s sounded real, yet not strident sounding like some of the dome tweeters when pushed hard, at least in my experience. Those planar magnetic midrange and tweeter really sounded super smooth and fast in reproducing transients. Best value buy at the show, even with the speakers on the floor. Must have forgot stands, look at that pile of Boulder electronics.
Magico S5 Mk II sounded very flat to my ears, meaning a ruler flat frequency response past audibility. Very revealing highs with extended low frequency response with some room coloration. Not time coherent from any material I could find on the speaker. Would love to see a step response of these ones.
Dynaudio Focus speakers. Another technically advanced active DSP loudspeakers. Very clean sounding at volume, a trait I note in all bi- tri-amp powered systems, including my own. This is another loudspeaker I would love to measure up and have a look at the step response.
I could not help sneak in a picture of VKMusic’s KT150 push-pull amplifier with micro-controller bias adjust. A combination of old school tubes with modern day micro-controller. Very cool.
Before we get into it, I want to discuss again the timing response of loudspeakers as to how it pertains to MQA. As mentioned in Stereophile, very few loudspeakers are time coherent or time aligned. It is worth reading the entire article to better understand why a step response is used to measure time coherency as opposed to viewing an impulse response.
What does an ideal time coherent loudspeaker step response look like? Pretty easy to model using a digital filter that is flat from 20 Hz to 20 kHz, with the bass rolling off at 15 Hz at some dB per octave as loudspeakers don’t reproduce 0 Hz or DC very well. Here is the step response of the model specifications:
The vertical “step” represents all frequencies playing at the same time and the downward slope is the corner frequency rolloff of the filter.
From the Stereophile article above, here is what a good measured step response of a time coherent speaker looks like:
Note the vertical step is the tell-tale sign of time coherence, and pretty much matches the shape of the ideal step response above, except that this is a measured response of speaker and room. Below is a step response, from the same article, but representative of a non-time coherent speaker:
Here we see the tweeter arriving first with a positive going step, just after 4 milliseconds, then we see the midrange arriving next, with a negative step, and then finally the woofer arriving with a positive step. This, and variants, are typical step responses of non-time coherent, multiway loudspeakers. Have a look on Stereophile if you can find your speakers measured step response.
Below is a measured step response of my 3-way, high efficiency speakers. These use a simple 3-way passive crossover, with high quality parts:
This is a zoomed in view on the time scale. Tweeter arrives first, negative step, then midrange, also negative step, then woofer, positive step. Even with frequency correction for a flat response, the timing response is still representative of the chart above. Meaning these speakers are not time coherent and therefore distort the timing of the music arriving at my ears.
However, by disconnecting the passive XO and using digital XO and time aligning the drivers, I am able to turn my speakers into time coherent speakers that do not distort the timing response that arrives at my ears:
I have also overlaid the ideal step response from the beginning of this section to show how close to ideal I was able to transform a typical 3-way speaker system into a time coherent system that can accurately reproduce the timing response that is on the recording. It does make a difference to my ears, especially the transient impact and imaging. Especially the depth of field imaging that make the speakers “disappear.”
What the heck does all of this have to do with MQA? If part of MQA’a claim is to deblur or
Photo courtesy of Element Acoustics
With respect to the MQA demo with the Tidal Sunray G2 speakers, Burmester 909 Mk 5 amp, Burmester 111 MusicCenter/preamp and MQA decoding from the Meridian 808 V6 Signature CD/DAC, sounded excellent to my ears.
Disappointingly, there was no comparison between a regular recording and an MQA encoded version of the same recording. Therefore there is no way to detect, determine, or decide MQA’s contribution to the sound quality other than one is listening to a $300K system with excellent recorded, mixed, and mastered source material.
If one’s speakers are not time coherent, and distort the timing response similar to the measured step responses above, then what sonic value is MQA’s deblurring filter? This precludes that one is striving for accurate playback of what’s on the recording. I certainly am interested in playback accuracy, both frequency and timing response to my ears at the listening position. If that is your preference, then either investing in a pair of time coherent speakers or applying frequency and timing correction to an existing pair of speakers by using DSP are the only two ways today to get playback timing accuracy to one’s ears.
I had a great time at the show. There were only one or two systems that succumbed to hotel room acoustics (i.e. the room was simply too small). Most systems sounded very good, even though I could still hear either boom or boxiness or both depending on the hotel room and speaker setup. Many exhibits sounded excellent despite the acoustics.
What I find fascinating is that the majority of these systems were in the tens of thousands of dollars and with quite a few in the hundreds of thousands. If I were to setup a system worth that amount of money, I would want to neutralize the effects of the room and ensure the interchannel frequency response is as identical between the speakers as one can get across a broad seating area so that the listeners were getting the best representation of the speakers and equipment as possible. While it is ideal to audition audio gear in one’s home, it is often very difficult to make those arrangements. Sometimes the gear one is interested in can only be heard at an audio show.
While loudspeaker and room DSP software is available, not too many use it at these audio shows. In the pro audio market for touring bands that take their speaker rigs from venue to venue, all use some level of speaker and room correction both for frequency and timing response tuning. For sure, other than the push button, auto-tuning correction software, one does need to understand what is going on in order to correct both frequency and timing response effectively across a listening area. I wrote a couple of CA articles on basic DSP and advanced DSP if you want to get a flavor of how it is achieved.
In addition to the skills required to neutralize the effects of a room, there may be fear that if all of the systems were smoothed the same way using the same target fr curves, then the unskilled listener may not be able to tell the difference between systems. However, the referenced target curves are a great place to start, many will fine adjust to one’s own preference, which may bring out the voicing of the speakers even more. That’s the beauty of designing your own custom frequency response and timing correction filter to match one’s own speakers to one’s unique acoustic listening environment.
My preference is for accurate sound reproduction so that the music arriving at my ears is as identical as possible to the music that is on the recording, regardless of format. In my case, I have used loudspeaker and room DSP software, as the number of time-coherent speakers are still far and few between. However, I am excited by the audio show to see the number of speakers that do use DSP, I wish there was more of them correcting the speakers timing response for accuracy.
Enjoy the music!
Thanks to Archimago for letting me use your pics from the show.
About the author
Mitch “Mitchco” Barnett
I love music and audio. I grew up with music around me and worked ten years as a professional recording/mixing engineer. Recently, I wrote an eBook on Accurate Sound Reproduction Using DSP.