Some speakers slant, some don't. Why?

[Jud]

53 minutes ago, firedog said:

Yes, if you want perfect time alignment, DSP is the way to do it. Some add on boxes like DEQX do it, and some actives like the Kii Three or Beolab 90 also do it internally to a digital signal.

 

There are also some speakers like Vandersteen that try to get time alignment right. They do a good job, but the speaker setup itself can't do it as precisely as DSP.

 

As an owner of Vandersteen speakers, I'm interested: Why can't speaker setup be as precise as DSP?

[Jud]

10 minutes ago, Keith_W said:

Jud, it is because time alignment by driver alignment and analog crossover are able to do it within a narrow band only. The actual crossover itself plays havoc with the phase, which introduces delay. There is only so much an analog passive crossover can do before it becomes incredibly complicated and starts introducing losses. After all, if you think about it, an analog passive crossover is lossy enough already. It is powered by ... of all things ... your music signal. 

 

DSP is able to accomplish all this within the digital domain, where the losses are comparitively small. Also, with modern computing power, DSP units (like my setup) are able to achieve time alignment to within 1/1000 of a second. 

 

When I first heard the theory of DSP driven crossovers, I was totally won over by it. The reality is somewhat different (isn't it always!!!). The reality is - doing a DSP driven crossover means that YOU (or ME), as a complete speaker design novice, is tasked with designing what my friend (a speaker manufacturer) calls, "the heart and soul of the speaker". And to do this without training, or if you are like me, only a high school understanding of maths. I have learnt plenty, and I have achieved plenty (my speakers already sound better than they did from the factory), but I have a long way to go. 

 

In short, if the DSP is handled by someone who knows what they are doing, the results can be phenomenal. Ask a twit like me to do it, and the results are ... variable. 

 

Hi Keith (and thanks to coot for starting this very interesting thread).

 

My Vandys are what Vandersteen calls in marketing-speak "phase correct," meaning in normal non-marketing language the crossovers are linear phase, so no monkeying with the phase and no group delay.

 

Also please note that many DSP crossover filters are minimum phase, so these in fact will do what you call "play[ing] havoc with phase" and have group delay themselves.

 

I too have only a high school math (no "s" here in the States  ) education, so I understand exactly what you are saying about having to try to learn a little bit about a subject where a solid mathematical foundation is necessary for real understanding.  Fun "learning on the job," though, isn't it?

[Jud]

Just now, firedog said:

I think you know the answer to this.

With well done DSP, you can basically take all the factors into account (including the slight alignment problems caused by the analog/mechanical attempt at time alignment within the speaker itself); so, as Keith said, you can get virtually perfectly aligned results. 

With analog you will get, by definition, less precise results. Doesn't mean it isn't good. The OP asked if time alignment could be done in DSP.

Of course, in both scenarios you also have to place your speakers precisely to preserve the correct time alignment. 

 

Yes, I did think that was what you were going to say.

 

I would say DSP might be more *flexible* than driver placement, but I don't know necessarily about more *precise*.  Using the instructions in the Vandersteen manual, I've got speaker placement down to about 1/8" in all dimensions.  Sound takes about .000009 second to move that far.  I haven't tried DSP crossovers - would you notice a difference of that amount in their timing? 

 

Of course I move my head more than 1/8" when listening to music, but then I would with DSP crossovers as well.  As you said, effectiveness of both DSP and driver placement will depend on the physical relationship between drivers and listener's ears; as soon as you move your head, there's a little less precision.  

 

Where I can see DSP working better is for situations you cannot solve by changing room characteristics and speaker placement in your particular listening space.  But you will want to watch out for things like the phase effects Keith and I mentioned.

[Jud]

4 minutes ago, audiventory said:

 

Hi Jud.

 

If you mean time delay precision, DSP have no limit by time shift size. It is more precise than sampling period T=1/[sample rate] and not only multiply T.

It is so, because digital signal, restored to analog, is not "stairs" of voltage. But it is curve interpolated by analog filter. Therefore time delay may be T, 2*T, T/2, T/9, T/1.4, T*2.1 etc.

 

Hi Yuri.  Yes, completely agreed.  But when you implement the digital crossover in a physical space with speakers and a listener, then the limitation is how precisely you can locate the speaker drivers and the listener's ears with respect to each other, just as it is with analog (physical) crossovers.

[Jud]

To put this question of timing and precision in a physical space another way:

 

Let's think about the 1/8" = .000009 second I mentioned before.  For convenience, we'll refer to this as 9 microseconds.  The highest frequency of the audible range usually quoted for humans is 20kHz.  One wavelength occurs in 50 microseconds.  Reasonably settled research results have shown humans can distinguish interaural time differences on the order of 10 microseconds, so we're getting close to that 1/8" distance.  (Kunchur got a result of 5.6 microseconds, but as with many things related to audio, his research has come in for a fair bit of criticism, so let's go with the greater bulk of results in our discussion.)

 

Figuring that we don't need to quibble about a microsecond among friends  , let's say we'll notice if distances between speaker drivers and listener's ears are off by more than 1/8".  Thus if we position our speakers with a precision of 1/8", we need to hold our heads absolutely stock still.  On the other hand, if you want to say we can use instruments to measure arrival times at the listening position with arbitrary sub-microsecond precision, then that gives us an entire 1/8" of freedom to move our heads during our music session.  Happy listening!  

[Jud]

11 minutes ago, audiventory said:

 

I re-read posts above, but I can't understand what difference between digital and analog crossovers relatively speakers. Both can provide delay.

Analog crossover can shift phase signal (time delay) in small range and depend on frequency, I suppose. It is hardly managed.

Digital crossover have wider range of delay value and almost infinite low step of altering of delay.

Digital crossover can model almost any analog one.

 

Hi Yuri -

 

This interview with Richard Vandersteen discusses his use of first order crossovers to avoid phase effects and group delay (it has a bunch of the usual marketing stuff, too, which of course you can ignore): http://www.soundstage.com/interviews/int07.htm

[Jud]

9 minutes ago, audiventory said:

From mathematical point of view it allow calculate phase (time) delay for each frequency range for desired interferention in ear point without accounting "internal diffractions or reflections" (if I correctly understand what is mean).

 

Hi Yuri.  This is marketing.  At that time Vandersteen did not house its tweeters or midrange drivers inside boxes, and did not mount them on "baffles" (horizontal surfaces of considerably larger dimension than the drivers).  See the image below:

 

 

[Jud]

4 minutes ago, coot said:

Wow. You guys are way over my head with your tech-talk - (maybe I should listen from above my head?).

So I guess if I'm looking to buy I should look for speakers that are either slanty or have digital crossovers and some mention of digital time-delay correction?

 

Or digital crossovers.

 

Not the worst idea in the world.

 

What would your notional budget be?

 

[Jud]

21 minutes ago, audiventory said:

 

 

Of course, all depend on implementation. As usual

 

By my experience of tuning sub by ears, it is non-trivial task.

 

Jud, are there digital crossovers with auto tuning by microphone?

 

Yes, no way I'd try this by ear.  You should ask Keith, I think, because I have no experience using digital crossovers.  I'm sure you can *tune* digital crossovers using mics, but I don't know if there are crossovers that auto tune by microphone (I understand this to mean the crossover would tune itself using a connected mic).

 

I also wonder (Keith?) whether the tuning that one can do oneself extends not only to frequency but to measurements of group delay and phase.

[Jud]

25 minutes ago, coot said:

$30 max; like you say, "notional"!

Specifically, Legacy Aeris

 

They sure look nice.

[Jud]

59 minutes ago, audiventory said:

 

Hi Jud.

Looks like he use soft walls for avoiding reflections. Wave interference is too complex thing. I don't ready say something about it. Need model such system for learning theoretical abilities, first. Interestingly, how they design systems? They are use modeling software?

 

Hi Yuri.  Yes, I don't think there are internal reflections, but that is likely true of most speakers.  I don't know how they do their designs, but I would imagine they must use software.  

 

Richard Vandersteen likes to project a certain image (former truck driver over 2 meters tall; credits his wife with inventing the "baffle-less" design when she played music through speaker drivers lying on a table not installed in a box; and you read in the interview the stuff about dragging shovels around to design the crossovers), but if I remember correctly, Vandersteen was one of the very first speaker companies to use Fast Fourier Transform analysis.  And I can tell you everything I can feel under the cloth that wraps my speakers is absolutely rigid.  It takes some good work to create a design that can be built relatively cheaply and have such rigidity.

[Jud]

14 minutes ago, audiventory said:

 

I think digital digital crossover may be tuned in system with room correction.

 

9 minutes ago, audiventory said:

 

Theoretically, active speakers are better way, because amp tuned by drivers, crossover too.

 

It is traditional scheme for studio monitors. That should provide (theoretically again) flat response.

 

Although studio monitors are not necessarily famous for great sound.

 

As you said, implementation is critical.  A lot of what I have read about room correction software packages involves frequency response, and it is my impression the filters used are very often minimum phase.  (I do see time correction mentioned in the Legacy and KEF materials, so perhaps digital crossovers pay more attention to this than most room correction software that doesn't come from speaker manufacturers.)

[Jud]

25 minutes ago, Keith_W said:

BTW I hope that you don't think I am some kind of DSP guru, because I am not. It just seems as if CA's real DSP gurus don't seem to be around very much these days, which is a bit sad considering the whole reason I joined CA in the first place was to learn from them. 

 

@The Computer Audiophile, maybe not a bad idea for an article, or a series of them?  Use of room equalization software to guide physical layout of listening room and system placement, and use as the basis for DSP; then how to do the DSP steps....

[Jud]

37 minutes ago, esldude said:

 

The squarewave from the Vandy isn't too bad.  Not as good as some Quads. Still quite good for a multi-way speaker.

 

 

 

Impulse response of the Vandy 2ce.

 

 

 

And for comparison the Quad Esl 63.

 

 

Of course in regards to super precise alignment, I do believe a 20% change in humidity will change points of focus due to change in the speed of sound more than 1/8th of an inch.   And a 2 degree change in temperature will make more difference than the humidity. 

 

 

Strange to me the Vandy would stop ringing quicker than the Quad.

 

I think an 1/8 inch movement by the listener may happen more often than 20% humidity changes or even 2 degree temperature changes.

 

Have any graphs for the Sound Labs or upline Vandys with built in amps?

[Jud]

9 hours ago, esldude said:

Those were all from stereophile measurements, which I should have credited.  The old software would do that for links from Stereophile automatically.  They don't do impulse or squarewave more recently.  The Treo does show a nice step response which is near textbook.

 

Of course how important is that?  Here is the step response of some Revel Ultima Salon 2 speakers which sound quite fantastic by most accounts.  This step response is similar to that of all the Revel speakers.  Most of which get near universal acclaim for sounding good.

 

I believe the ESL63 had a resonance between the panel and metal cover around 13 khz which is why it rings on.  The later near identical design of the 989 did reduce that problem. 

 

Thanks, esldude, informative as always.

 

Wish I knew how to read the "waterfall plots" that are so popular these days.

[Jud]

23 hours ago, esldude said:

Yes, the issues you point out are a result mainly of the choice to use 1st order crossovers in the Vandersteen.  

* * *

JA's description of the Vandy 2CE:

In the time domain, despite the Vandersteen's multiway design, its impulse response (fig.7) is as time-coherent as that of the single-driver, crossoverless Fujitsu Ten Eclipse TD712z, reviewed elsewhere in this issue.
Read more at http://www.stereophile.com/content/vandersteen-2ce-signature-ii-loudspeaker-measurements#0eG5hLm0CRCwkWxG.99

 

The step response (fig.8) also features a time-coherent, right-triangle shape, though there is a rather faster decay than I expected.
Read more at http://www.stereophile.com/content/vandersteen-2ce-signature-ii-loudspeaker-measurements#0eG5hLm0CRCwkWxG.99

 

So one philosophy is to have very even responses, controlled off axis dispersion, and steep crossovers to control the drivers well with a step response that isn't time coherent.  The other is to allow wide driver overlap, and have time coherent response at least at some points in front of the speaker. This usually results in an uneven off axis response particularly above and below the listening axis. 

 

Both have their fans and adherents.  I've heard both sound good.

 

I would say I think the PSB/Revel/JBL approach seems to produce speakers which sound good easier with less fuss in more rooms.  Which one ultimately sounds best comes down to details of execution.  What information there is indicates our hearing in the upper frequencies at least is not fussy at all about phase or time coherence.  That leads me to think letting that go makes it much easier to produce a satisfying speaker.

 

Richard Vandersteen's interview indicates a deliberate decision to achieve time/phase coherence at the cost of flatter frequency response across the band.

 

Ricardo has mentioned before this bothers him.  It didn't bother me with the 2Ci and 2Ce, and assuming the 3A Sig is similar, there either.  One way to look at it is that I learned to hear this sound signature as "correct," or at least what I was looking for in audio reproduction.  Another viewpoint is that since people do hear differently, perhaps I was among a sub-population predisposed to pay greater attention to time/phase coherence in evaluating the realism of reproduced sound.

[Jud]

3 minutes ago, semente said:

 

I also find this a plausible explanation and one's favourite musical genres also play a part here.

 

Sphile's 3A measurements show a more even tonal-balance although the upper mids and treble are shelved up by a few dB, maybe the 3A Sig is even better in this respect.

There seems to be quite a bit of resonance coming from the midrange driver (kevlar?):

 

 

Vandersteen 3A, cumulative spectral-decay plot at 50" (0.15ms risetime)

 

Read more at http://www.stereophile.com/content/vandersteen-audio-3a-loudspeaker-sidebar-3-measurements-page-4

 

I'm pretty sure Vandersteen has never gone in for Kevlar.  Carbon fiber or fiber/balsa sandwich, maybe.

[Jud]

7 minutes ago, semente said:

 

The reviews is pretty worthless in that respect.

The specs page mentions 4.5" midrange unit with patented, low-diffraction magnet, curvilinear polycone, and ferrofluid cooling

 

From Vandersteen's FAQs (http://vandersteen.com/support/faqs):

 

Quote

Cone materials that excel in one application may not be appropriate for another. For example, aluminum, Kevlar, and carbon fiber are very good for subwoofer cones and tweeter domes where they remain in the piston mode through the pass band and the inherent nasty peak from their first break-up mode is either way out of the pass band (subwoofer) or above audibility (tweeter). Woofers and midrange cones however, need a neutral material with high internal dampening like treated paper since their first break-up mode will be in or near the pass band. Since it is very difficult to control the initial quality of treated paper cones and to maintain the quality over time, we use proprietary or selected mineral filled polypropylene cones in these critical applications.

 

[Jud]

1 hour ago, audiventory said:

 

Sound perception have no common points for people. Because we can't access to foreign mind.

 

I knew you were going to say that.