Can we hear the effects of time alignment? (1 Viewer)

[Michael R Price]

There seem to be two types of speaker manufacturers: those who think time alignment is important, and those who don't bother with it. I don't know which side I should be on.

Can anyone describe what one hears when a speaker has serious phase distortions? Or what the improvement would be if I, for example, switched to a speaker with the same drivers as mine but with a perfect impulse and phase response? I am trying to learn here, given that I don't really have an opportunity to hear a time-aligned system. Thanks.

 

[Kevin C Brown]

An easy test. If you have a home theater!
Purposefully set the distances to your speakers slightly off, and then even more off. Pick a selection of music you are familar with. You should notice "blurring" of things like cymbal crashes, snare drum hits, etc. Basically, the upper frequency should sound less crisp. Not exatcly the same as within the same enclosure, but it gives you an idea.

 

[Michael R Price]

Kevin, that is an interesting idea. Unfortunately I only have a two channel stereo system :frowning: I think that even the errors involved with speakers with very poor phase wouldn't be the same as the delays you could incur with your surround processor, since at 10kHz for example a 10,000 degree phase error would be only a few milliseconds (maybe 3 feet worth of difference).
I just asked this question because my speakers use high order crossovers and thus have pretty bad phase responses. I believe there are products out there that can correct your system's time domain properties.
I have noticed that the treble sounds clearer when only one speaker is playing. However, I think this is more of a room problem and less to do with time alignment (since the one speaker alone still has phase distortion). Thanks for the idea Kevin.

 

[ling_w]

Another way to do the time delay test is, if your speaker is bi-wireable, hook the low and right channels to an upper and lower rnage amp and hook those amp channels to the high and low freq drivers of the speaker. Then vary the time delay of the speakers.

I think that does as much as tilting your speaker back and forth, which is what companies like Spica and Wilson does with their WATT.

Anyway you could try to tilt your speaker back a bit so that the voice coil of the tweeter is in vertical alignment as the midrange or woofer's.

Usually good time alignment speakers has very good imaging. As the previously mentioned Spica and WATTs.

 

[Mark Seaton]

Hi guys,
"Time Alignment" is an interesting term, and realize it is a trademarked term, I forget by whom(Thiel? Dunlavy?). In any case, the problem with conventional loudspeakers is that multi driver speakers will almost always have problems of integration of the drivers meaning that you can only "align" things for one location, or at best for an arc in front of the speaker. Furthermore, a direct radiator has a theoretical 90deg phase lag which translates to increasing delay with lowering frequencies. Electrostatic and horn loudspeakers can in fact have a near zero phase angle, and can properly reproduce a waveform. The most significant problem with horns lies in making a full range speaker and integrating the different parts. Our Unity Summation Aperture design was just granted a patent last week and you get see some background on it in this thread at Audio Asylum:
Unity Patent
After working with these speakers and seeing refinements to the product, I would say that we can percieve changes in "time" and proper phase response. This design does allow for a time correct solution which appears in measurements as ONE coherent source, with zero indication of the crossover points or multiple sources. It is only at the upper and lower end of the speaker's passband where any shift is observed. In short, when frequency response has remained mostly the same, I have never heard an improvement to the time/phase response make things sound worse, only better.

 

[Michael R Price]

Even though this topic seems to be fading away, I have one more question:

It's obvious that making a time aligned speaker requires more resources than any other old speaker. (Special crossovers and cabinets for physical alignment, etc.) Especially at lower price points (under, say, $2000 or so)... Do we think it is worthwhile to using a time-aligned design as opposed to a conventional design using better quality components?

 

[Tom Brennan]

Michael---There is a difference of opinion on this; some makers think it worthwhile and others not. Interesting story follows. This whole thing came about in the 1930s at MGM. Eleanor Powell tapdancing was being played back on a Western Electric 2-way horn monitor with a 12'(yes, feet) path difference between the woofer horn driver and the tweeter horn driver. The result was a "double tap". John Hilliard, a sound engineer at MGM studied the problem and discovered the double tap was caused by the path length difference of the drivers. Hilliard found when he moved the tweeter driver back even with the basshorn driver the double tap went away. Hilliard eventually determined that with a -3' path difference such anomolies were inaudible IF the crossover was 500-1000hz. Hilliard was subsequently the project manager on the development of the famous Shearer Horn theater system with greatly reduced path differences compared to the WEs and was later the engineering Kahuna at Altec where he developed the VOTs, the VOTs having perfect path alingment. The VOT came out in 1945. People have been thinking about this a LONG time.

 

[Saurav]

Do we think it is worthwhile to using a time-aligned design as opposed to a conventional design using better quality components?

Another thing to realize is that the human ear isn't linear, it is more sensitive to phase/timing errors in certain frequency ranges than in other ranges. Specifically, I believe 1K - 4K is pretty much the sweet spot for our hearing sensitivity. Sadly, that's where the tweeter-to-midwoofer crossover point occurs for most 2-way or even 3-way speakers. So, what some people do to get around it is to use "full-range" drivers. That is, you design a speaker that uses a single driver for say 200Hz to 9KHz. Add a light and fast 10" woofer under it, and that'll handle it from 200Hz down to say 30, depending on your room - you won't get earth-shaking bass, but you won't miss the fundamental notes of a bass guitar either. Similarly, add a "super-tweeter" to handle things from 9K out to 20K or beyond. You'll still have the phase/alignment issues at that crossover point, but they'll be in a frequency region that the ear is much less sensitive too.
Of course, in reality it's more complex than that. There aren't many drivers that can handle that specific frequency range, for starters. The ones that exist have other problems, which mostly stem from trying to reproduce that frequency range accurately, so you're back to square one.
Still, there are some designs that work on this (or similar) principles... some people love them, others hate them. What else is new

 

[Kevin C Brown]

Vandersteen makes quite reasonably priced time aligned and phase correct speakers.

 

[ling_w]

It's obvious that making a time aligned speaker requires more resources than any other old speaker. (Special crossovers and cabinets for physical alignment, etc.) Especially at lower price points (under, say, $2000 or so)... Do we think it is worthwhile to using a time-aligned design as opposed to a conventional design using better quality components?

The Spica TC-50 was time aligned, with a sloping face. It was holographic, not much off from the Wilson WATT. And its price was only in the $500 range.

 

[Mark Seaton]

One detail which should be realized is that the designs above which have been badged as "time aligned" are only aligned for one location in space, or at best an arc in space. The energy off axis will not have this same relationship, and the reflections in the room will reflect this. Likewise, the area where this "alignment" holds true is rather small. Creating a Time Correct speaker is not a simple matter, as there is also the matter of phase, where most direct radiating loudspeakers have a 90deg phase lag which progressively delays lower frequencies. With this sort of phase response, drivers can only be aligned at crossover, but the response can never be completely Time Correct. This is easily measured with a TEF 20 or other TDS measurement which actually does measure phase directly, rather than interpolating it mathmatically from the frequency response as many measurement systems do. This method only holds correct where the drive units are behaving ideally. Any breakup or beaming will make this calculated phase response incorrect.

It is difficult to evaluate the time aspect of a loudspeaker as it is very difficult to have a system where ONLY the time aspect of the respone is being changed. Again though, I've never heard an improvement in time/phase response to do anything but better the percieved sound.

Regards,