|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
Hi everyone,I've read the websites of Thiel and Meadowlark which promote their "time and phase coherent" designs. This begs the question-"Why doesn't everyone build speakers that way." It seems that in audio there are always tradeoffs. I'd just like an explanation of the tradeoffs inherent in the designs of Thiel, Meadowlark, etc.
Follow Ups:
http://www.audioasylum.com/scripts/t.pl?f=hug&m=60107
If you really want an intelligent discourse on this topic I highly recommend you take the time to read this AudiogoN thread. It is both interesting and informative.Happy reading,
DB
try reading the thread about wristwatches (here in the Speaker Asylum) in which DawgByte is a very active particpant!!
JimBop -Please take the time to read through the Time & Phase thread on AudiogoN and then make a comment regarding the credibility of my post. I think you'll find it both interesting and informative. Any time consumers have the ability to speak directly with an audio engineer and ask them probing questions regarding their designs and theories, we will be in a better position to make informative purchasing decisions.
Sorry DawgByte - just giving you a little ribbing for the wristwatch thread. I wasn't trying to question your integrity, and no offense intended. My apologies if you were offended by my remark.
No worries, I'm not that thin skinned. ;-)I know I've now posted this three times, but I swear the AudiogoN thread is fascinating. I've read it twice and some guy emailed me telling me he's read it four times. I learned a lot about speaker theory from that thread. It also put me in direct touch with Roy Johnson of GMA. I've met some smart people in my time and Roy is right up there with some of the brightest. I've heard amazing things from an owner in New York about GMA's C-3 speaker. Amazingly transparent, wide and deep soundstage, halographic, articulate, fast and weighty bass, plenty of air around each instrument and sounds great with ALL kinds of music.
The basic problem lies with the fact that 1st order crossovers must be used that have a 6 dB/octave slope. Thus, each driver is called upon to reproduce a much wider frequency range. That creates two problems: first, because there are two drivers reproducing the same frequencies, there can be destructive interference between them, which complicates getting a flat amplitude response and second, all cone/dome drivers do not radiate a uniform amplitude response into the front hemisphere at all frequencies.I've read the same ad copy that claims that the other guys are "too cheap or too lazy" to design for "phase coherence." Nonsense. If that were true, then at least all of the mega-buck, "statement," cost-no-object speakers by such folk as B&W, Dynaudio, JM Lab, etc. would be "phase coherent."
In my own limited experience, other things are more important. I have never cared for the sound of Thiel loudspeakers, for example; and I don't think that has to do with their "phase coherence."
If you're buying a pair of speakers, it's probably better to listen to them than read about them.
"This begs the question-'Why doesn't everyone build speakers that way.'"Because they are difficult to execute, and often require cabinet/driver placement designs that go beyond merely mounting drivers on a simple box. Hence with similar materials, such speakers are more expensive...
When time/phase coherent designs are executed properly, the speaker will have a clean step and impulse response, and will have a flatter impedance curve.
I have studied a long time from all availible e- and paper magazines, what are reviewers' comments on what speakers and what is more hearable. You cannot tell it from just your own experience, it takes statistically meaningful bunch of different conditions. I've come to a conclusion that some other things give more hearable effect than phase coherence - but it has no disadvantages as such, just that is ok in just one listening point. Killing all kind of resonances in speakers and enclosures is most important as far as drivers are decent.
and I can't find a single disadvantage. Mine are 3-ways with ScanSpeak drivers, 9700 tweeter, 8640 mid and 8565-01 woofers. I did not use first order crossover, but rather John Kreskovsky's filler driver theory. The roll off of the high end of the woofer is approximately 2nd order LR, the low end roll-off of the tweeter is approximately 2nd LR, while the mid (band-pass) is approximately 1st order on both ends. Crossover points are at 500Hz and 3500Hz. It is one hell of a lot of work to get the crossover right as the drivers sum over wide ranges. The response will sum flat and yield transient perfection only if the relative spacings of the drivers are very precisely located. 1/8" makes a difference. The system is transient perfect at only one point in space, namely on the design axis (midway between mid and tweet in my design) and only at one distance from the speaker face (I designed mine for 10', my listening position). The inter-driver phase differences necessary to create "transient perfect" vary slowly with distance once you are beyond (say) 6 feet from the speaker face so the speakers do in fact sound really great at any reasonable listening point.Remember that what phase distortion does is smear things in time. By going to a transient perfect design, things such as percussion, piano, attack on notes, etc. are more dynamic....you get a much greater instantaneous SPL and it is very noticable. Detail is well defined, and very complex detail such as echoes and reverbations in an auditorium seem to go on forever, while time smearing turns these into background noise in ordinary speakers.
Finally, there is nothing different about TP designs from other designs from a development standpoint. The principal difference is that you must work with drivers that can be made to respond at the required levels over a broad frequency range. My mid-drivers respond up to 11kHz, and believe me, shaping that high end response is challenging. How good a job did I do? Well, the impulse response of my speakers is less than 1/44,100 seconds, the resolution limit of my soundcard. Pretty damned good! And you hear it when you sit down and listen to very complex orchestral music. For jazz, vocals, chamber music and anything else involving a relatively small number of instruments what you need are good speakers, not transient perfection.
This terms are very much mixed up.Phase coherency
Phase alignment
Phase symetryTime coherency
Time alignment
Time Symetry
This is just a another "hi-end audio" myth. IOW bullshit. A multiway loudspeaker *may* be coherent in time and phase near the crossover point at only one place in your room and not at all audio frequencies. Even single drivers have problems.If we had a small sphere that had 360 degree radiation with both perfect linear phase and amplitude (ie. no crossover or group delay) than we could truly evaluate what your asking. Then we'd have to look at cost, drive-ability, sensitivity, power compression and other types of audible distortion.
Magnetar
Hi Magnetar;> > > ...If we had a small sphere that had 360 degree radiation with both perfect linear phase and amplitude (ie. no crossover or group delay) than we could truly evaluate what your asking... < < <
Correct me if I'm wrong, but isn't that pretty much what a Tesla coil induced plasma sphere speaker does? I've read rave reviews of the things, but they're way out of my price reach.....
Magnetar
Magnetar
As far as I know they don't go very low, so you'd have to mate it with a conventional (non-plasma) woofer.
Magnetar
Hi Magnetar;> > > ...As far as I know they don't go very low, so you'd have to mate it with a conventional (non-plasma) woofer... < < <
Absolutely true for the 'store bought' variety. I was thinking more along the lines of a DIY job with a big enough plasma ball to make some bass. Problem there is that in addition to hideous equipment cost, the electricity usage would be genuinely obscene just to generate the plasma ball. Still, I think it would meet the criteria you laid down, at least as a research instrument if not a viable home system.
A second possibility would be a 'flame speaker'. which has far lower equipment cost, though that uses an expanding column of combusting gas as the diaphragm. As you can imagine, a large enough "Wizard of Oz" style flame speaker would be capable of heating the house (and burning it down if one was not careful!) so the practical limitations outweigh any audiophile considerations.
Still, they're both interesting concepts that could be used to actually try out some of the theories about idealised radiators. Pity I don't know of any research facilities that maintain 'em. I'd love to hear a set that are bigger than tweeters...
But from reading about this, I gather that one issue is the geometry of sound distribution with first order crossovers. In most cases, it seems when I read reviews of Phase/time coherent designs, the reviewer mentions narrow "sweet spots" and critical placement issues.At the same time, most seem to feel that within that narrow sweet spot, imaging is more precise.
For my part, I've heard very good sounding systems based on phase/time coherent designs as well as non-phase/time coherent designs. Based on my experience, it's just not a good litmus test for saying one speaker sounds better than another.
My 2 cents,
Even the Quads are quite directional and they don't have a crossover.You're quite right. It's how the speaker sounds rather than the design type that is important.
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
If you take the time to read through this thread, you might change your mind - purely from a physics standpoint it is hard to argue with the logic of this design concept.
I think I saw a review of his speakers in AIG, in an issue I can't locate, and the conventional measurements for frequency response and dispersion were spectacularly good. Indeed, if I could listen to them, they might well be in the running for purchase.For me in my role as a consumer, I really don't much care about the issue, although speculatively I have some interest in it. As a consumer, I want a speaker I like that is good and reliable so I can enjoy music recordings and other recordings using it. I have to admit the the speakers I liked enough to buy have been pretty phase coherent in the range where the ear is most sensitive.
Oh, and thanks for the interesting link.
______________________________________
____________________________________________________________
"Nature loves to hide."
---Heraclitus of Ephesus (trans. Wheelwright)
Hi jholcomb;An experiment is worth a thousand posts - so why not try a phase coherent design and discover what YOU think? Buy a pair of inexpensive fullrange drivers and try 'em out. With only one driver (and in the smaller examples like the FE127, FE103, Tang Band line, etc. there's not even a whizzer cone), it's about as "time aligned" as you can imagine. Thus for anywhere from $20 to $70 you can find out the advantages of time coherence for yourself.
Be cautious though - if you aren't careful, the next thing you know, you'll be building backhorns, low power SET's, etc.! Your mainstream audio friends will think you're "touched" and before you know it you won't even care what the audio press is stoked over this month.....since you'll be busy enjoying your music to care.
Morse---Yeah, and then he can replace his time problems with distortion, compression and freqiency response problems and be reminded of why engineers went to multi-way speakers in the first place. :-)
Hi Tom;Each to his own, each to his own....but after going the old fashioned route and trying to go back to "modern" designs, I discovered that multiway speakers sound pretty darn lifeless in comparison (to my ears anyway).
My experience has been just the opposite.I have multi-ways,and "full range"....Multi-ways win....no contest.
Hi Don;> > > ...My experience has been just the opposite.I have multi-ways,and "full range"....Multi-ways win....no contest... < < <
It just goes to show that each of us judges subjectively using criteria that is personnally rather than universally relevant.
If you're after the widest dynamics, highest SPL's, and deepest bass then I can see why you'd go multiways - particularly multiway horns.
Personally, I'm a midrange freak, do not require the deepest bass, listen at apartment friendly levels, and need something that a middle aged back can lift without risk of damage - and I've found that the fullrange driver solution "works" for this set of criteria quite well. What about compression and peak limiting? I doubt that my peak SPL's ever hit 90dB! With 96dB SPL speakers it's just not an issue, and the speakers have a lifelike quality that's quite musically enjoyable - in my subjective experience.
Enjoy your music in whatever way suits you best!
Jhol,The disadvatages of time & phase aligned speakers are that the crossovers tend to be a bit harder to design and the speakers are less efficient by the use of 1st order crossovers. The inefficiency of the speaker typically means that the speaker will not play as loud as other speaker designs unless the design utilizes multiple drivers and employs a tremendous amount of power.
tube---Crossover slope in and of itself has no effect on system efficiency; I've built a number of super-efficient 1st order horn systems, so have PWK, Bruce Edgar and many others.
I see no relationship between time/phase coherency and efficiency.Why do you say this?
Magnetar
...physically that is.Are they time coherent then?
I guess phase coherent will depend on the crossover, yes?
We are talking about less than 2 or 3 ms of seperation right? I don't know just asking.
The audibility is dependant on distance and crossover frequency so it varies.John Hilliard, a big engineering honcho at MGM and later Altec, did studies on this after the famous "double-tap" incident at MGM in the early 1930s. When monitoring an Elenore Powell tap-dancing scene through a 2-way Western Electric horn speaker double taps were heard. Hilliard, then an engineer in the MGM sound department, traced the effect to the path differences of the bass and treble horn drivers to the listener which was on the order of 15' or so. When Hilliard moved the treble horn back the efect went away. Hilliard came to the conclusion that with a crossover in the 500-1000hz range a path difference of 3 milleseconds, about 3 feet, was inaudible.
This incident helped lead to Hilliard's (and his team's) design of the famous Shearer Horn at MGM, a much improved horn loudspeaker with far less path difference than the WEs. It also led later to Hilliard's Altec VOT's in which the path diference was totally eliminated.
Someone comes along and points out it really works in three. Thanks for enlightening my ass. I guess that makes more sense since the "time" component is dependent upon that old friend time vs. wavelength. Did not Meridian allow this with there DSP 8000 speakers. I also remember Atkinson measuring them and finding them "not" time-aligned. Meridian commented that it sounded better without it. However, that just might be a blurb. Again thanks
"Hilliard came to the conclusion that with a crossover in the 500-1000hz range a path difference of 3 milleseconds, about 3 feet, was inaudible."Not saying itīs like this, but maybe;
He found out that the sound of two taps was eliminated at or below 3ms (GROSS time distortion), but this does not mean further improvements couldnīt be made by using a time coherent/TP design.
My speakers (digital X-over) allow for direct comparison of time coherent and non-coherent behaviour. Even for my tin ears there's a difference. It's subtle but it's present. It only affects the image, with small bands of 4-5 musicians the musicians seem to change location slightly.Time coherent designs are the only accurate, so for the sake of accuracy the designer should care about this issue. The tradeoff IMO is higher overall cost.
Most competent designers do this out of hand.. some don't even mention it.. otheres 'grasping at straws' make much of it.
It's a phrase with little actual meaning left in it.. suggest avoiding anyone/thing who makes issue out of it.
The great majority of speakes out there has major timing distortion.The are some brands that claim phase coherent/time coherent speakers but obviously has serious time distortion anyway, such as;
Dynaudio, Sonus Faber, Lumen whitelight and more.The only brands Iīm aware of that use transient perfect/phase coherent solutions is; Meadowlark, Earthworks, Thiel, Vandersten.
There is also some confusion about what these terms mean which makes the whole thing more complicated.
John Kreskovsky has some nice things to read about regarding transent perfect speakers/x-overs.
/Peter
"Most competent designers do this out of hand.. some don't even mention it.. "Do you actually know what you are talking about?
While I am no engineer, I am an owner of Meadowlark Kestrel Hotrods. The design of this speaker isn't perfect; which are? It is very fun to listen to. The sloped baffle design and high quality drivers produce a realistic if a bit narrow(on some recordings) soundtage. Imaging is incredible, very pallable with natural detail. Is it the end all design, no, is it a great design that works well with a variety of material-yes! The trade offs seem minor to me but again that is just my opinion. I find them to present music in an incredibly believable way ..as many say "natural".
In passive design, both drivers have to output significantly beyond their xo point thus the driver and the cone material has to behave well extensively so to speak. cost goes up because you need better materials and drivers.Also, depending on the size of the cone, the sound can have a tendency to beam messingup the power distribution, soundstage and the summed up response of your freq. curve (tonal balance). It is one thing to design a stand-alone phase coherent xo, but once it is integrated with the rest of the design...its not coherent anymore, compounded by the fact that if the material is not up to it you need to compensate with complicated xo design. cost goes up again.
time alignment is another one where cost goes up some because its more labor intensive to built compare to a box. Time alignment is only maintained within the design listening distance too...
active speaker are the ones that can truly do away with the above constraints because you can built in very steep slopes w/o delays, and so not overextend the drivers/cone. but few manufacturer have the combinattion of skills to built the amps and the xo and the driver and the housing, thus the steep price for good active speakers...better designs are more expensive.
Maybe it's because the time/phase thing is only effective over a relatively narrow frequency range and listening window. Also, crossovers for this type of design tend to be pretty complicated with all sorts of corrective circuitry to get the drivers to behave. Ever seen a first order Thiel crossover? Some pre-amps have less components!Lastly, maybe the whole idea is somewhat over rated. If it was the nagic elixir its proponents make it out to be, don't you think more designers would use it?
"Lastly, maybe the whole idea is somewhat over rated. If it was the nagic elixir its proponents make it out to be, don't you think more designers would use it? "No, becasue so far with analouge means, itīs hard and much more expensive. However with the new digital era coming, the nubers of time correct speakers will raise, no doubt about it, becasue eliminating errors is what hifi is about.
This post is made possible by the generous support of people like you and our sponsors: