Jump to content

Phase coherent loudspeakers ?


thiptoman

Recommended Posts

What are the opinions on phase/time coherence in loudspeakers here ? There are some companies praising the benefits of a time coherent speaker over a non time/coherent design . I've never heard such a speaker and was wondering if this is really such a big deal in loudspeakers . Any opinions on the subject would be appreciated . Thanks !

Link to comment
Share on other sites

>What are the opinions on phase/time coherence in loudspeakers

>here ? There are some companies praising the benefits of a

>time coherent speaker over a non time/coherent design . I've

>never heard such a speaker and was wondering if this is really

>such a big deal in loudspeakers . Any opinions on the subject

>would be appreciated . Thanks !

This is an idea that I think started in the 1970s. One of the pioneers in this field was DCM Time Windows. It became a trendy fad, a lot of manufacturers jumped on the bandwagon for awhile. The theory tries to deal with what is called propagation delay, the time difference between when voltage is applied to the speaker voice coils and when the cone responds. In general, the larger and more massive the cone, the longer the delay. So called time coherent loudspeakers attempt to address the difference from one driver to the next in the crossover range by delaying the output of the tweeter just long enough for the woofer to catch up so that the move in the same direction at the same time. Some use crossover network elements, some carefully try to match woofer and tweeter (so called light "fast" woofers) and some will place the tweeter further away from the listener by the way the cabinet is constructed. Dick Shahanian tilted his Rectilinear V backward 4 degrees to accomplish this goal (also putting the listener off axis and increasing reflections from the ceiling at the same time.) The expectation is that the coordination will result in the equivalent of a single driver having a single motion. In theory it can't work. The fact that the drivers are physically displaced in space means that they are two centers of propagation which will each radiate as a souce causing complex interference patterns anyway. As an analogy, imagine two rocks thrown into a still pond landing on the water at exactly the same time but in different spots. Each will cause ripples to eminate outward but they will not be equivalent to one large rock landing causing a single pattern of waves traveling outward in circles. Interestingly, in the 1950s, the problem was viewed from a geometric point of view rather than a time based point of view and the solution was to employ coaxial and triaxial speakers where the cones had the same center of radiation. But those were not time compensated. This is analogous to the two rocks falling in the same spot in the pond but at slightly different times. Same problem. If a coaxial or triaxial driver was matched timewise, it could produce a single coherent wave...but only on axis. As soon as you move off axis, the difference in distance between the listener and the two cones would result in the same interference patterns again. By the way, these interference patterns result in peaks and dips in frequency response whose precise nature will depend on the geometry of the speakers, the location of the listener with respect to them, and the propagation delay time difference. As far as I know, nobody has ever conducted double blind tests using time coherence as a controlled variable to determine whether or not it is even audible and under what circumstances. I don't think most manufacturers worry about it anymore. They have entirely new fads to profitably market.

Link to comment
Share on other sites

Guest Barrydor

I owned and enjoyed Acoustic Research AR9 speakers for 27 years. In that period of time, I never found another speaker that could equal their performance until I heard time/phase coherent speakers. I now own Vandersteen Model 5As and I have not looked back

The easiest way to implement a time/phase coherent speaker is to use a single driver, which by definition makes the transducer time/phase coherent. A good example of this design is the Quad electrostatic speaker. Unfortunately, a single driver design almost invariably has serious frequency response and SPL limitations

The objective of a multi-driver time/phase coherent design is to minimize the phase shift and propagation delay of the crossover network. Many speakers use high order crossover networks with steep roll off slopes so they can better control driver power dissipation and thereby use less expensive drivers. The phase shift and delay that are inherent in higher order crossover networks is substantial and varies with frequency

Time/phase coherent speaker designs often use first order crossover networks, which exhibit a constant phase shift. The shallow crossover slope of a first order crossover allows substantial out of band material to reach the drivers so the drivers must be designed to handle a wider frequency range and greater power without distortion. With well-designed drivers, time/phase coherence has little or nothing to do with the size and/or mass of the cone

Time/phase coherent speakers often physically align the drivers so that the distance from the acoustical center of the driver to the listener is the same. In other words, usually the midrange is set back from the woofer and the tweeter is set back from the midrange. In addition, the driver radiation patterns are often carefully controlled to ensure that interference patterns and reflections are minimized, allowing the drivers to behave as a point source

The goal is to have every frequency reproduced at close to the same time and phase as possible while the drivers reproducing them are equidistant from the listener. By doing this, all frequencies arrive at the listener in phase at the same time. This theory can and does work very well but is unfortunately far more expensive to implement properly. The analogy of ripples on a pond in the previous response is irrelevant because it places you beside the speaker, not in front of it at the listening position

The mangled square wave/transient response plots of many speakers (where one can actually see each of the drivers start and stop in turn) make one wonder how the sound can bear any resemblance to the original. In addition, the typical high order crossover often exhibits ringing in response to transient events. A good time/phase correct design presents an almost perfect transient response plot

The timbre of musical instruments consists of many related frequency multiples beyond the fundamental frequency. It only makes good sense that it is desirable for a good transducer to reproduce all of the frequency components in phase and at the same time. Unfortunately, there are those who consider the arrival time differences to be inaudible and as I said before, time/phase correct designs are inherently more expensive to manufacture

In my opinion, a good time and phase coherent speaker is far more resolving, accurate and musical than anything else I have heard. They are also far easier to place in a room and they exhibit far less room interaction than non- time/phase coherent designs. In addition, unlike most other speakers (including my AR9s), I have *never* experienced listening fatigue from a time/phase coherent speaker design, regardless of how long and how loud I listen

Several audiophiles and musicians who have listened to my Vandersteen Model 5As and other time/phase coherent speaker designs share the above opinion. I strongly encourage you to audition a well-designed set of time/phase coherent speakers and decide for yourself if the design offers benefit. As I said earlier, I enjoyed the AR9s for many years, but I have not looked back

Barry

Link to comment
Share on other sites

"The analogy of ripples on a pond in the previous response is irrelevant because it places you beside the speaker, not in front of it at the listening position"

I have to disagree. The analogy is very apropos, the only difference being that the medium of the pond is water, for the speaker drivers it is air, and the interference on pond surface is in two dimensions while the interference pattern for the speaker system is in three.

There are many differences between AR9 and Vandersteen 5a. Aming them the Vandersteen speaker has a rear firing tweeter whose arrivals at listener are substantially out of phase (by hundreds of degrees) not only with the front generated arriving midrange sound but with the front tweeter as well. Those who have read my other postings about how I've changed my AR9s know I've added three indirect firing tweeters and their output constitutes more than 90% of the high frequency sound energy generated into my room. I can listen to mine all day also without any listener fatigue. IMO, the improvement comes from eliminating the shrillness of a single front firing tweeter, not from any time alignment.

Link to comment
Share on other sites

Guest Barrydor

Perhaps the word "irrelevant" was a little strong. I think I was confused by the two dimensional aspect of the analogy

Looking at it another way, the major interference patterns would occur close to the stones, much like the major driver interference occurs in the near field. By the time the ripples reached the observer, they would look like a unified wave front, or a point source

The rear-firing tweeter on the Vandersteen Model 5A is designed to use for replacing rear wall reflections as required and can be turned off completely. Many listeners, including myself, do not use the rear-firing tweeter at all

Barry

Link to comment
Share on other sites

Once the distance between the centers of propagation of two drivers becomes a significant percentage of the wavelength of the sound reproduced, interference patterns in the audible range are inevitable. The wavelength of a 5khz sound wave is about 2.5 inches. The best way to "time align" drivers would be to bi or tri amplify them and use digital time delays. The best way to avoid interference patterns in the crossover region would be to use digital filtering with very sharp slopes. Digital filters can be made considerably superior to analog filters. I think NHT has just such a model. That should result in considerably improved performance of a satellite/subwoofer system in linearity of midbass response. Linkwitz-Riley has an interesting idea, it recognizes that the interference is inevitable but the desgn places the physical location of the greatest interference pattern in a area of the room where the listener is not likely to be. Read a description to get an explanation of how it accomplishes this. It doesn't yield as flat an overall electrical response as Butterworth alignment but it may have advantages once the acoustic fields are taken into consideration. Again, nobody to my knowledge has ascertained under what conditions so called time smear or time incoherence between drivers is audible by testing the same speaker system with one controlled variable only. All so called time corrected speakers have many other differences from other speakers including differences between each other and it is not clar that their supposed time coherence is responsible for any increased preference or perceived accuracy so far. However, if you have a reference, I'd be glad to take a look at it.

Link to comment
Share on other sites

From the link

"The air pressure is irrelevant!"

Hmm, that's not the way I remember it. The speed of sound is given as 1090 FPS at sea level. As I recall it, as you go to higher elevations, the velocity decreases because the density of air decreases. The denser the medium, the faster the speed of sound propagation. For instance, sound travels five times as fast in steel as it does in air. It's so efficient, they say you can hear a train by putting your ear to the rail when the train is still 100 miles away. Any comment about sound velocity versus air pressure? Wouldn't this be the same reason sound velocity changes with temperature, because colder air is more dense? I'm going to have to look that up.

Link to comment
Share on other sites

With the two links as reference, I'm now wondering if it's actually possible to build a phase coherent loudspeaker that will properly perform over a normal household temperature range.

Second, given the extremely small delay we're talking about, can someone actually hear it, let alone measure it at 10 feet from the speaker cabinet? And if so, how much of an effect, if any does temp and humidity have?

Link to comment
Share on other sites

>With the two links as reference, I'm now wondering if it's

>actually possible to build a phase coherent loudspeaker that

>will properly perform over a normal household temperature

>range.

Thiel Audio has done a pretty good job. Check out their stuff.

>

>Second, given the extremely small delay we're talking about,

>can someone actually hear it, let alone measure it at 10 feet

>from the speaker cabinet?

TDS (time delay spectrometry) using a TEF analyzer can do the job.

And if so, how much of an effect, if

>any does temp and humidity have?

I suspect it's negligeable. However, the theorists will included it in a discussion just to show it exists. Can you hear a difference at 20% RH vs 90% RH? Who knows? I know I can't.

It's all about the music!

Carl

Carl's Custom Loudspeakers

Link to comment
Share on other sites

>With the two links as reference, I'm now wondering if it's

>actually possible to build a phase coherent loudspeaker that

>will properly perform over a normal household temperature

>range.

>

>Second, given the extremely small delay we're talking about,

>can someone actually hear it, let alone measure it at 10 feet

>from the speaker cabinet? And if so, how much of an effect, if

>any does temp and humidity have?

From the second website;

"The key variable in this standard atmosphere is temperature as all other basic properties can be computed from it. Since the speed of sound is a function of density and pressure and both density and pressure are functions of temperature..."

It's a nice looking website and it is printed unambiguously but it doesn't make any sense to me, in fact it defies common experience. Here are some examples. BTW, as I recall, 0 degrees F is 454 Rankine for reference. Is there a significant difference in air pressure between the south pole where record temperatures have been recorded around minus 120 and death valley where temperatures are often around +114...in the shade? That's a difference from about 330 R to 570 R about a 60% increase in temperature. Is the temperature of the contents of a can of compressed air different from that of the air in the room it is stored in? What am I missing here?

The proof that for any location off axis of a multiple transducer speaker system no matter how carefully physically aligned and timed cannot produce a "phase coherent wavefront" over an overlapping frequency range is obvious from simple wave mechanics and consideration of its geometry. The question of whether it matters for subjectively accurate reproduction of music is one which as far as I know has not been determined but I am very dubious. If it were, speakers such as AR3 which in no way ever claimed phase coherence could not have been made to sound even remotely close to subjectively accurate yet the demonstrations AR conducted proved otherwise. In my experience based on my own experiments, other factors far more important should be the real variables targeted for more accurate speaker systems. The intuitive notion which audiophiles and even engineers have in this regard are usually way off the mark.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...