Adding a time aligned cable to time aligned speakers.

[John Holmes]

Does it make sense to do something like this?

Could the wire phase and or time adjustments have an adverse effect on a crossover that has been maximized (time aligned) for the speaker already?




"I came here, to chew bubble gum and kickass. And I'm all out of bubble gum!!!" My DVD's

[JonMarsh]

Hello John,

An interesting topic to start a Saturday morning on.

This is one of those things where, when people talk about cables and "time aligned cables" (which approaches codswaddling in marketing speak), compared with what we talk about in "time aligned speakers", the worst cables are much better than the best speakers.

Cables is one of those unfortunate areas, where what we know about what influences the sound (including human psychoacoustics, placebo effects, etc) is so small compared with what we *think* we know about other components.

Consider the measurments on my DIY speaker cable, as reported in the DIY forum in December:

DH Labs Silver Sonic Cable, 14 ft:

Loopback Inductance: 3.25 uH
Distributed Capacitance: 0.453 nF
Resistance one leg: 0.16 ohm
Resistance both legs: 0.32 ohm


DIY 8TC based Cable, 14 ft:


Loopback Inductance: 0.35 uH
Distributed Capacitance: 1.36 nF
Resistance one leg: 0.033 ohm
Resistance both legs: 0.066 ohm


Both cables are minimum phase (which most speakers aren't), both will pass square waves with fairly low to non-existent waveform distortion (don't you wish you had that in a speaker?), but there are differences in measured and subjective performance which clearly correlate with measureable *frequency response* differences. Those differences don't stem from differences in the "time alignment" of the cables.

Generally, when companie's claim to be able to measure differnces in the time alignment of cables, they're talking about using a TDS (Time Delay Spectrometer) and measuring cable reflections and glitches which occur in the RF range. Some companies feel strongly that this correlates well with what we hear; others do not.

Regarding minimum phase (time aligned) behavior in speakers, I built time aligned speakers in the 70's with first order crossovers, and very good reproduction of wide band impulse waveforms, such as a 1 msec pulse (which requires bandwidth and phase alignment beyond 1/10 and 10X the nominal pulse freuqency (1 kHz). However, it's possible to achieve that coherent response only in a relatively small window on axis. This is true for modern speakers using this approach, such as Theil, Vandersteen, and Dunlavy labs. Now, these first order crossover have significant drawbacks, including non-flat room power response (due to off axis crossover lobes), and relatively high distortion, because of the wide band response demands on the drivers, and the likeliehood of operating a driver well beyond it's pistonic region. What I've found over the years, is that I got better correlation for high quality sound reproduction by using drivers within their pistonic region (no cone breakup modes detectable, not even in impedance curve, the more senstive method), having very dead, insert cabinets (non-resonant), and selecting drivers with minimal inherent distortion, due to balanced motor design. Then, the rest is voicing for combination of axial response and very smooth off axis response (large sweet spot, flat room power response). This has, in my opinion, yeilded the best results, especially for the money, and considering the dynamic range of modern recordings. I'm harldly a trailblazer in this opinion, though- just take a look at Revel's philosophy- founded by Kevin Voecks, formerly of Snell Acoustics. He also used to believe in the mantra of pulse reproduction in speakers. Now, time alignment of the transfer function in the crossover region is very necessary, and often the easiest way to get that is by physical time alignment of the acoustic centers of the drivers in one plane- but it's not the only way. DCM Time Windows, anyone?

History lesson is over.

Best regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[Lex]

Great question John.

Nice post Jon. But do you really think the only characteristics about a cable that matter are measurable in quantifiable terms? For example, we know that though 2 amps may both spec very well, in the end, our ears choose a particular one for it's "Character".

Lex

[JonMarsh]

Hi Lex,

No, I don't think the *only* things that matter are readily quanitfiable. However, I think a lot of things *are* quantifiable, (even for cables) and a good designer will make every effort to do so, and to understand the interaction with other components in the chain.

So far, in my experience, most of the things that impact a cables "tonality" are measurable in one way or another with the right equipment. And, because of the way the human ear works, I've found that although you can't reliably discern a few tens of a dB on a sine wave comparison, when you consider the affect on music (or even pink noise) over broader frequency regions (which is more in alignment with the affects that cables have) these small differences are audible. This is something I found out when designing RIAA preanmps in the seventies, and ended up settling on a two stage topology with partial passive interstage equalization, and some active equalization in the last stage. The best sounding design followed the RIAA curve with high precision (0.05 dB) from below 20 Hz to beyond 50 kHz. Larger errors than this could almost always be identified reliably by listeners, *even though the speakers were no where near this flat*. Weird stuff, huh?

In the real world, it's a trend for a long time in a "high end" system to try to pick components synergistically- matching the strengths and weakness to each other. This seems reasonable, until you make a more subtle examination of how and why those "strengths and weaknesses" occur. If a solid state amplifier has a bit of glare in the presence region (not at all uncommon), you may be drawn to matching it with speakers that are slightly reticent in that range- which is also a common shortcoming in many cone type speakers, because of the nature of popular driver sizes and crossover designs.

Does this solve the problem? Well, it masks it, or ameiliorates to a degree. But, the forward characteristic of the amp is results from IM and THD, which aren't reduced by using a speaker with a slight dip in the response. Similarly, the dip in the response (particularly the integrated room power response) comes because of the narrowing dispersion of the midwoofer, and may be exacerbated by off axis crossover lobing. (classic example: Egglston Works designs) (Also, the majority of 6.5" two ways with conventional drivers and crossovers (2.5 kHz and above, typically)on the market). So, the problem really hasn't been fixed, it's just been kludged around.

An important conclusion out of this is to use measurements with a high correlation to what we hear, (Jensen's spectral contamination measurement is a good example) and to develop measurement methodologies which inform design decisions.

This is probably why I like "Stereophile" magazine better than "The Absolute Sound", because at least the former provides some objective feedback and measurements, and in the case of speakers, CD players, and amplifiers, these give some additional insights to the knowledgable reader.

Best regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[John Holmes]

"An interesting topic to start a Saturday morning on."

Every now and then I get lucky!

"Great question John."

Thank you sir!


On to the subject. Sorry Jon, I'm not as technically gifted as you, thus some of your response is over my head. But what I did comprehend, you give more weight to a cable being able to perform this ( time alignment) over a speaker?

Next, did you cover the possibility of one canceling the other? If you did I apologize, I missed it.

"Regarding minimum phase (time aligned) behavior in speakers, I built time aligned speakers in the 70's with first order crossovers, and very good reproduction of wide band impulse waveforms, such as a 1 msec pulse (which requires bandwidth and phase alignment beyond 1/10 and 10X the nominal pulse freuqency (1 kHz). However, it's possible to achieve that coherent response only in a relatively small window on axis. This is true for modern speakers using this approach, such as Theil, Vandersteen, and Dunlavy labs. Now, these first order crossover have significant drawbacks, including non-flat room power response (due to off axis crossover lobes), and relatively high distortion, because of the wide band response demands on the drivers, and the likeliehood of operating a driver well beyond it's pistonic region. What I've found over the years, is that I got better correlation for high quality sound reproduction by using drivers within their pistonic region (no cone breakup modes detectable, not even in impedance curve, the more senstive method), having very dead, insert cabinets (non-resonant), and selecting drivers with minimal inherent distortion, due to balanced motor design. Then, the rest is voicing for combination of axial response and very smooth off axis response (large sweet spot, flat room power response). This has, in my opinion, yeilded the best results, especially for the money, and considering the dynamic range of modern recordings. I'm harldly a trailblazer in this opinion, though- just take a look at Revel's philosophy- founded by Kevin Voecks, formerly of Snell Acoustics. He also used to believe in the mantra of pulse reproduction in speakers. Now, time alignment of the transfer function in the crossover region is very necessary, and often the easiest way to get that is by physical time alignment of the acoustic centers of the drivers in one plane- but it's not the only way. DCM Time Windows, anyone?"

Does the above statement apply to speakers which use a 2nd order x-over as well, such as THX speakers?

Thanks,




"I came here, to chew bubble gum and kickass. And I'm all out of bubble gum!!!" My DVD's

[JonMarsh]

Hi John,

Sorry if I'm being a little obtuse or verbose at times- that happens before my Saturday morning coffee, you know. I'm just operating on automatic, in a sense.


My point is that by any standard which you might apply to speakers, standard cables (even zip cord) are much more "time correct" than loudspeakers. That doesn't mean that cables can't have audible colorations, just that their behavior doesn't include significant phase shift or group delay compared with what speakers do. When people talk about phase shift or group delay in cables, they're usually referring to at most few tens of degrees at 20 kHz or so- an area where the human ear doesn't reliably respond to phase. Where the human ear is quite sensitive to phase seems to be up to about 3-5 kHz. Standard speaker cables have no appreciable phase shift in that region.

So, then the question may come up about time shift between the lower frequencies and the upper? This is true, but simple inductance is a good indicator of this. And again, compared with speakers, cables are pretty dang good. I think where you will most often here the affect of cable characteristics, is in octave or more wide ranges of subtle frequency response differences. Look for things like major differences in inductance and resistance, to start. They do exist, as my comparison above shows.

So, to answer as directly as possible, any tolerably decent cable will *not* destroy the time relationships of the drivers and crossover response in a minimum phase time aligned speaker (say, one of those affore mentioned Theils, Dunlavy's, Acoustats, etc.). That DOESN'T
mean that the cable can't or won't contribute audible coloration- but it won't be due to making the kinds of changes to the signal that a non time-aligned speaker does.

Now, when you talk about phase/time response affecting transient clarity, I do have to agree, but in my opinion, the biggest culprit in modern systems (after poor speaker designs, especially those using drivers outside their pistonic range) is high feedback electronics. Or maybe any electronics with feedback loops. I'm getting more and more radical about this in my old age.

Right now I'm enjoying some of the cleanest, detailed, and most natural, focused sound I've ever had in my home system. And there isn't an opamp or conventional feedback amplifier anywhere in the signal path from my CD transport to the speakers.

SCD777ES Player/Transport



MP-DAC III, zero feedback design


Marchand passive preamp



Ayre V-5 Zero feedback solid state map

Now, only one kind of speaker is "time correct", or minimum phase- that is, either a full range electrostatic, or a multi-way dyanmic system using first order crossovers, or special construction's like Eric Baekgard's Link driver, where a 1st order bandpass driver is used in conjunction with a 2nd order bass and treble wired *in phase*, not the normal out of phase connection necessary for 2nd order Butterworth or Linkwitz-Riley (all pass) networks. These can be minimum phase at some summing point in front of the speaker.

ANY multi-way speaker, regardless of the crossover type, should be "time aligned" in the crossover region, ideally. That is, the acoustic centers of the drivers should be lined up in the same plane. With conventional speaker construction, that often isn't the case, and so you have to compensate by "playing games" with the woofer and tweeter filter transfer functions, so that the phase is shifted enough to compensate for the time difference. The more the acoustic centers are offset, the harder that is to do.

Whether the speaker has a 2nd order filter, 3rd order, 4th order, or whatever, the principle is the same- in principle, the acoustic centers *should* be in the same plane, then a "classic" filter function can be used , with good results. When you have to add phase compensation in the network, there's usually some accompanying amplitude problem, maybe just an extra few dB bump or ripple, but there will be something. Vance Dickason's Loudspeaker Cookbook gives a lot of good examples of the tradeoffs there. Certain crossover types are easier to compensate in this way (such as the 4th order Linkwitz-Rile) because the phase transitions are complementary between the two drivers, and occur over a relatively narrow frequency range. A fourth oder L-R crossover *should* use drivers with time alignment of the acoustic center, and if it does so, will sum very flat on axis, with a smooth phase shift and hand-off between the drivers. BUT, it won't reproduce a pulse waveform, because of the changes in phase over signals in the crossover region. Is that a more audible problem than the difficulties "time aligned" speakers have considering their total room power response, distortion, etc? Only the prospective purchaser can answer than. It's their ears, after all.

BTW, THX has nothing to do with this. THX is a set of standards for basic capabilities for cinema sound, and for THX marketed home systems. Doesn't mean much of anything in my experience, except that it will play at least X loud, over Y frequency range. Nothing to do with minimum phase response, "time aligned", or anything else like that. It doesn't specify crossover type or time coherency, etc.

I'm not saying THX is bad, just that it's not particularly relevant, other than setting some minimal standard for cinema and HT. Unfortunately, too minimal, I think, judging by some of the bad sound in THX theaters, and poor DVDs with THX certification.

End of Sunday's rant.

Have a nice weekend, John- I hope I answered in a way that's useful to you.

Best regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[LarryB]

JonMarsh:

Thanks very much for a highly informative post (even though I understood only about 1/5 of it ).

Larry

P.S. What do you think of Jim Thiel's approach of putting the tweeter essentially inside the midrange?

[JonMarsh]

Using concentric drivers the way Theil, Tannoy, and Kef do is a good design principle- the biggest advantage being that it will elminate vertical lobing in the crossover transfer function as you go above and below the driver axis. (assuming the usual configuration of mounting tweeter and woofer in a vertical line).

The potential pitfall is horn loading and diffraction effects from the cone. So, you don't have the same freedom of design for the main cone- because typically you'd like to make it reasonably shallow. The depth of the cone angle will affect how wide a dispersion you get from the tweeter at different frequencies.

This brings up the whole issue, then, of how do you chose to optimize the power response of the system, that is the direct, early arrival on axis sound, versus the integrated room response- which varies by frequency. There is a lot of difference of opinion about controlled radiation angle versus wide angle versus omnipolar versus bipolar versus dipolar- at the least, things aren't boring, that's for sure!

For any given type (conventional box, dipole system like Magneplanar or dipole dynamic driver speaker systems such as Sigfriend Linkwitz has designed) you have to understand the issues for the particular speaker and the preferred room characteristics and positioning to get the best results. A sprited discussion of those pro's and cons could fill quite a few pages here!


Best regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[John Holmes]

Thanks Jon. I had a great weekend! I trust that you have/are enjoying yours as well.

As usual, a very informative response. Yes, you've hit the nail of what I was asking.

Thank you.

My big reason for asking was, I got to thinking about the cost of the speakers that we enjoy so much. We spend the money for said speakers because, we fell in love with the sound. In many cases, we heard these speakers on standard zip cord. Then for whatever our reasons, we decide that we would like to "tweak" the sound. For mid to hi-fi models, most of the exspense (let the manufacture explain it) is R&D. Okay, I can buy into that. But if that is true, how can we improve (with cable) what's been optimized to began with?

Now let me state for the record, for several reasons, I do believe and know that cables make a difference. I have heard it. I am just curious.




"I came here, to chew bubble gum and kickass. And I'm all out of bubble gum!!!" My DVD's

[JonMarsh]

Hello John,

You've brought up some good questions, particularly, the affect that the equipment the speaker designer uses to do the final "voicing" has on the correlation of subjective results. In a very real sense, if you want to hear *exactly* what the designer does, in principle you should be using the same gear and anciliaries.

Most of the classic AR speakers were voiced using tube amps, and the interaction of the highish output impedance of those amps with the driver and crossover impedance *was* part of the voicing of the speakers- if you ran them with solid state gear of that day, things got a bit funky for more than one reason! Frequency response changes were introduced compared with how the speaker was originally voiced.

Most comtemporary designers will use amplifiers with reasonably low output impedance, so at least you don't get variations in frequency response from that.

You're right that the cables can be part of the network seen in series with the speaker and affect the response. How much, and how well it correlates with the original design is hard to know. I have to feel, though, that the closer you can get the cable to resemble a short piece of busbar between your amp and speakers, the better off you are. Minimize inductance, minimize resistance. That's one argument for locating amps close to the speakers.

So, though the cables can be a impact, what about the room? When I was working on "transient perfect" speakers in the 70's, our main listening room was a full blown LEDE room (Live End, Dead End), which was considered a superior acoustic design for reducing room effects while still having a fairly natural far field acoustic. It's still used a lot these days- particularly for recording studios- but how many home listeners have a setup like that? Room issues and setup will in many cases be a bigger concern- not that you can't still hear the affect of the cable, but tuning the speaker position in the room, as well as the listening position, and killing comb filtering from conflicting early reflections (side wall, typically) may pay off with far greater results than a kilobuck set of cables. But once you've done all that, then it's time to look at the smaller factors.

Then too, you have to consider your starting equipment. If you're trying to work from the basis of a Bose surround box or Sony's HT in a box, (or my daughter's Cambridge Audio Sub/Satelite system) all this detail stuff is swamped out by shortcoming of the original design. But as you get into good speakers, hopefully enough attention was paid in developing them that they benefit from fine tuning the associated components- whether it the cables, the room, the electronics, whatever. I've never seen any good component that wouldn't benefit from improved neutrality and transparency elsewhere in the system!

I was also wondering if you had some specific cable or speakers in mind with regads to your question- or the claims of some manufacturer. Just curious.

Regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[John Holmes]

Jon,

I fully agree, the room and proper positioning will yeild the most dramatic results from your speakers. Before I moved a few months ago, I used some early reflection treatment with decent results. That room was much smaller than the one I use now for the HT. It also was dedicated while this one is not. The big plus with the new locations room has been, it has enough room for my fronts to be 40" from the front wall and my closest side wall being 64".

I also have noticed that the up-line equipment can play a huge role in the final product that we hear.

Some of my questions were from the issue of already owning decent speaker cables. Many times, an individual may have purchased nice speaker wire for it's sonic signiture to mate with "that" specific speaker. But what happens when you buy new speakers? I did this exact thing! I purchased speaker cable to deal with a very specific issue in that "dedicted room" with a paticular speaker. Now that I've moved, the same rules do not apply. However, without any thought, I just assumed that this being a so called "quality cable" (and not to mention it's cost), it should go back to it's duties.

The reason that I did not mention any brands were, I felt that it was not as relevent to the bottom line. I figured that many of us are in the same boat and may not have given it any thought. Nor did I want any prejudice to cloud the issue.

I should also add that, I am very happy with the sound of my HT. I did put it together with the thought of HT first, music second. But, I would be lying if I said, that the numerous post about these wonderful music soundstages that you guy's speak of, didn't peak my interest to what I may not be hearing. I feel that my system does very well on music, but I have never been able to pick out every instrument's placement within the stage. Two or three and the vocals maybe on a given recording. So, I am just kicking some ideas around as to why this mabye so. Maybe I just don't have the ear for such! I use the equilateral method of placement as well as have my tweeters at the proper height. I have also played with countless measures of toe-in. And no matter what speakers or other equipment that I have used, the best that I can say is, my vocals are dead center and my sound is clean. My soundstage is huge. I'm just not able to pinpoint instruments like others have written about.




"I came here, to chew bubble gum and kickass. And I'm all out of bubble gum!!!" My DVD's

[JonMarsh]

Hi John,

One way to look at this, is, if you're happy with your general tonal balance, you've got a "huge sound stage", and hopefully a reasonable size sweet spot or listening area, why worry?

There are things that can be done to improve imaging specificity. Some of it may fall under the pain in the *ss category for you, depending on your room size and esthetic consideration.

For starters, if you're setup primarily for home theater, your speaker positioning may not be optimized for stereo imaging. Even Mike Knapp over at Home Theater Talk found that just speaker positioning could make a big difference, by flattening the balance and distribution of room nodes, and minimizing early reflections that conflict with the imaging cues in the program material. A good starting point for a stereo speaker layout is provided at the Cardas site:



This may not be a practical setup for you, because of other domestic consideration, but try it as an experiment- just to see for yourself how it works. I've personally found that farther out from the back and side walls is mandatory for strong recording specific imaging, but of course, it can mess up the family room or living room from the viewpoint of SO.

Another issue is room treatment. I mentioned live end/dead end (LEDE). The end of the room where the speaker are should be somewhat the dead end. In particular, you want to minimize side wall and ceiling reflections which arrive shortly after the main sound. This causes comb filtering of the frequency response, and smears the "time characteristics" (a lot more than any cable!)

This evening I'll pull up some RPG room optimizer runs and post them for you, and show how it identifies areas for selective treatment. Basically, from your seating position, anywhere on a wall or ceiling if you put a mirror, you sould see one of your speakers in it, that's a potential problem area.

If you think your room is acoustically dead enough as it is, do this test: walk out near the speakers (or have a friend do this) and clap your hands. Not continuously, just once or twice. Listen closely- you;'ll probably hear a slap echo of the sound from the walls and/or ceiling, and you may hear a flutter echo as it decays away, bouncing back and forth between the walls. This is your room's *own* acoustic, and if left to itself, it will tend to overwhelm the acoustic on the recordings you have- whether natural or synthesized.

The less the influecne from your room, the more you'll hear into the recording- for better, or for worse! Note that this exposes the fallacy of the Bose direct/reflecting concept- if you create an acoustic using multiple room reflections in your own listening room, you can't reproduce the acoustic space on the recording.

Of course, that brings up another issue- what kind of acoustic space is on the recording? Multi-channel pop recordings are often pretty much pure synthesized, and often not very well. Good jazz and classical recording should be much better- I have some recordings of Holtz's "The Planets" that can raise goose bumps on the right system setup. Even a synthesized pop recording can do some very interesting things- just listen to the James Taylor "Hourglass" SACD or CD.

Last, regarding speakers, amps, CD players, etc., any part of the chain that adds it's own footrpint (glare, hardness, resonance, etc) detracts from imaging becuase it obscures the original program material and nuances which define imaging. Bose Sub/satelite systems may be the ultimate offendors in that category, thought the Cambridge Audio Sat/sub setup my daughter has isn't far behind (she and I built a nice set of speakers for her, to use at her mom's home, which her mom appropriated for the living room. And now they're destroyed, because of a mold/fungal infestation in that house because of improperly installed Pergo flooring. No, I dont want to talk about it! )

All for now,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[John Holmes]

"And now they're destroyed, because of a mold/fungal infestation in that house because of improperly installed Pergo flooring. No, I dont want to talk about it!"

Okay, I won't touch it! :B


Holy CRAP! The detail in music is back!!!

I applied the formula to the width which moved the speakers from 6 ft. apart to 7'11" apart. I cannot believe how I can now "see" instrument placement within the soundstage. What else that has surprised me, the vocals now not only are centered, but depending on the recording have additional placements through out the presentation. As if I was at the show. I also notice some instruments seem to be behind me and to my sides...this is great!!!

Still, I have to figure out a way to bring them further into the room as, they are still only 40" from the front wall.

I put in a movie to test the new positions. It seems to work okay for this too! So far I've used no toe-in.

I plan to play with it a little more but, man oh man, am I happy with the results.

Jon, thanks so much for giving me your time and countless effort. My respect for 2ch is literally 10x! I still can't believe how simple it was to fix. This was the detail that made me buy these speakers!

Thanks again, I'm going to get emotional with my (new) system. :B




"I came here, to chew bubble gum and kickass. And I'm all out of bubble gum!!!" My DVD's

[JonMarsh]

Cool!! :B

-Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[Bruce]

Jon,

Do you have any experience with speaker placement and listening positions on opposite long walls (Immedia method?).

I've found it gives the main speakers plenty of breathing room from the side walls, plenty of room between L&R, and tends to make back wall reflections disappear when your listening position is within 18 inches of that same back wall.

I have a room that's 19' x 13' x 8-12' cathedral ceiling and this method works extremely well. Great wide soundstage, great individual instrument imaging, and the soundstage even appears to go 15 feet deep (in back of the speakers).




Bruce
____________________________________________

[JonMarsh]

I have read about it, but I've never had a room setup that it would work optimally for me. At one time I was using that layout in my current home, but it limited the audience size for the projector. You also have to be careful about speaker positioning and listenera positioning relative to the back wall; in both cases, you're on boundaries, which lifts the low end; this is OK depending on the speaker characterisitics, and assuming no one has to listen *away* from the rear boundary wall.

Regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....