Lynn, it is possible to create the feeling of an orchestra sound "big sound" with the tone and air movement. It just takes the right recording, lots of drivers and real fine tuning of the horns and room. i agree it is impossible with the typical high end system. You need a system that has no compression that is detectable and one that will project perfect tone and scale - stereo gets you close but to give the full size use delayed upper, rear (with phase manipulation like Circle surround) and deep bass channels. My current system doesn't pull it off quite as well as a couple of other systems I've built because I need to get better power response from the main channels. In the sweet spot it is very close but I know from previous systems my 1K up response is too narrow so I am going to different midrange and treble horns.
No offense, as we are indeed struggling with this as well. These are the very same questions that has been raised within the our small group here as well. In fact, while we have a good Discussion Paper on the practical side of achieving results - how to get those results consistently. We are learning what to look for in drivers, design execution and so on, so we are up and away on that front. But of course we are not satisfied with that - these guys wouldn't be and not myself. There are questions that needs to be answered. And we are employing people of caliber here with mathematical abilities better than mine - and perhaps getting some kind of mathematical model.
So what is it?
What If I was to say that you can make a speaker system that can be driven from any source Z, that the box alignment stays in tact, that the slopes of the crossover doesn't change - all locked in? That would be at least a bit interesting, but what happens when you listen to it is what makes it very interesting.
We don't have all the answers - but got people here working on it.
I hear you, it was not our idea to use those words - alas Dave Wilson beat us to it. We are not comfortable with the language either.
But interested in your 'skirt' comments re digital playback - to the point that the skirt could mask a large 'clinging rise' (Paul Miller showed this in an exotic [expensive requirement] measurement) relative (more than +60dB, kid you not) to and above the noise floor and that kind of ULF noise is something Allen and I discussed, as one of the major source of 'digital sound' and that the other was post-DAC slew rate problems. But getting a bit OT.
I will indeed show your comments to the guys in our group - and they do tally with the reluctant use of language that some us have - hence 'noise' and 'jitter' are only provisional, it's only because of Dave Wilson who calls his crossovers as 'anti-jitter' and is in market place, these terms can stick even if not entirely valid - a case of once out of the box, hard to contain.
No, I don't think so, as that is already very low to start with. We both perceive 'phasiness' quite easily, and I am pretty sure I understand by what you mean (ever played with M-S recordings and gone too wide?) and indeed I can find it totally intolerable. But not entirely dismissive of some motional feedback too, if those phases don't sync and that can certainly cause effects that bounces back into the system - just put even a lightest finger on any diaphragm connected to a motor system and the Z plot has changed. Air and any, for want of a better word, turbulent air, will reflect back. Indeed box resonances can do this as well. But this is now only because that I am alert to what may be reflected on the electrical side. Esa Merilainen has some interesting stuff on motional feedback in his 'Current Driving' book.
But no, this affect points to something related in the source being voltage. It is related to current phase angles, that is our view.
When drive voltage and current gets out of sync, then what is the product?
For example, one of our guys did a series-crossover two-way and noted that it just sounded cleaner, in fact better than many other designs he had done, really sweet. Others have made positive comments about series crossovers, there must be a reason. But he didn't do any more as he found them a pain to do.
With another two-way design he did, with a classic rising Z up to 1-2KHz peak, used a conjugate filter (LCR) to flatten the Z so it was largely flat in band from about 200 Hertz up, which likewise flattened the current. It was confirmed it had it had zero affect on the response. Then he and a co-worker were able to blind test each other and both able with ease to hear when it was in or out.
In those two cases above, I believe there is a common denominator. They flatten the current phase angle requirements of/from the amplifier as you flatten the Z. A voltage source can produce any current phase angle without constraints, not really a good thing, it is the load that is the boss - but a current source will always produce current of zero degrees irrespective of the load - as the series source Z is high and the load Z is low. Add a series resistor to a driver and it will go flatter - and a current source is just a big resistor.
Then we have what Wilson does: It is not hard to figure out what he does - and a friend of mine, Edgar Kramer, has owned both his Sasha and now his Alexis, both with 'anti-jitter' crossovers. He says that in comparison with early Wilson models he has owned (and reviewed some) that there is a lack of hash and what he was described sounded very much what we were hearing and a significant lowering of the noise floor. Hash is what hash sounds like - and an obvious increase in clarity and fluidness and indeed stability of imaging etc.
Edgar then interviewed Dave Wilson (the interview is on YouTube) and Edgar brought the subject up and Wilson decided to keep mum about it - and Edgar didn't press him and I haven't heard anybody press Wilson either. But there is no doubt he is treating the same sort of thing and knows the crossover is a big part of the problem. He is doing what he does because he has heard it and so have we - this is not a phantom issue.
I had lunch with Edgar before the Wilson interview and told him I had figured out how Wilson was doing it. I don't know if I should reveal it here, but I will say that it wastes a lot of power and flattens the electrical phase angle in a rather crude way. You may be able to figure it out from that...? I will leave that up to you. But a knowing eye over Atkinson's published Z plot makes it pretty clear. In fact, I have used the same trick as well. So this adds to the case book - and the clues.
We notice that flattening out the current phase angle of the voltage source, keeping current travelling as much as possible in a straight line, and here 1st order will indeed have an advantage. Across the Voice Coil, any parallel components/network should be avoided - they will alter current versus frequency (resistors are less problematic as the draw straight current) and besides, you want the current go through the VC and be constant.
You aim end up with a very flat Z across the board and likewise the current phase angle will also be near flat as can be.
This is an actual measurement of the new Elsinore Mk 6 - published only yesterday:
All this is of course only just interesting, until you actually listen and then realise you have stumbled, sort of, and by some intuition and gut instinct, across something that audibly translates into something significant. Like, you play it for people and they go w-o-w. No other words needs to be spoken, it is that obvious.
If we were to drive this speaker system with a current source, and you can, the Z plot would stay the same, but the current phase angle would flatten even more and become 0 degrees. The final response will stay constant as it will be determined by Theivenin's Theorum and will even lock in the mechanical high-pass filter of the bass alignment above 20 Hertz, and also, lock in the slopes of the crossovers, so that both mechanical and electrical filters are locked in and retain their slopes.
What I am saying is that the Elsinores can be current driven. That is some party trick - but think about it, if it behaves the same whether fed current or voltage, then are we not getting some current drive benefit from a voltage source?
I hope that intrigues you as much as it does me. There is a logic to that.
Here we have it simulated in SoundEasy.
Red is voltage and Green is current.
I can assure you it translates perfectly into the real world with my actual 40 Watt transconductance amplifier, 100mV in and 0.25A out, with output Z = 270 Ohm.
Note that Thevenin's Thereom is applied here - when the Z is equalised, the response will follow - and that raises a lot of questions as to what happens to Q - including Qe being eroded by Re+X, where X is the Z that is seen looking backwards - even DC resistance in large series chokes can be tuned out - if desired, so this has legs.
Now we must come back to the topic re big sound and it does have that.
Yes, there is the lack of grain that you only hear when it's not there - this is that hash that Edgar was describing. But that is not all.
For a start, the sound totally escapes the confines of the box - and creates a soundstage of such stability, able to focus small and the suddenly when source requires it, to expand as if there was no limits - and an ease that just totally engages you to actual instruments playing. It is not a subtle effect. Yet small instruments stay their comparable size, tight and focused at the same time.
I really suspect that we are achieving some kind of pseudo-current drive - it is current through through the VC that should be tracked, it is current that creates motion in a motor, and not the voltage across the VC. I believe that the current easily gets corrupted. You do raise an interesting question - if to be categorised as a distortion, then what is the category - and it doesn't show up with static signals - everything looks normal.
Time for bed.
Cheers, Joe
Last edited:
And the higher price alternative is...?
(Note that I don't want to misunderstand your brief comment. Also note that the original question was aimed at Mr. Rasmussen, sorry for any confusion.)
Zobels aren't unknown in loudspeaker crossover/balancing network use.
Last edited:
These two statements seem to be in-congruent, because of course a series network is a network based on parallel components with the driver/vc.
BTW, this is interesting.. but it deserves its own thread. (..Moderators, hello?)
Kudos!
That one works well.
I think it is great that you're developing systems to identify the means and cause the desirable performance aspects.
Is it off topic? But this is definitely related to the topic in a wider sense.
I now know a number of designers who avoid Zobel networks. Because they say they can hear them. I used to use them as well and it takes a philosophical change of thinking.
Any parallel network will stop what the VC ought to do, that is the voltage across VC should be a function of current. But we are using voltage as the source to drive it. That means that the voltage across the VC is 'regulated' and the current in the VC now becomes a function of the 'regulated' voltage. We then are is a state of hope that the current will be, for want of a better word, clean.
What happens now, with a network that changes Z with frequency. It then makes any hope we have impossible.
Yes, we have inductance (and other things, but let's stick with that for now) and the rising Z is actually inductive EMF, where F is force and to be understood as voltage, V. The Z goes up as a result of a force generated by the VC. If we put a current through that, the voltage will go up if driven by I, but the current will stay the same through the VC. We solve one problem, but create another. We solve one if we go over to V driving, but then taken our eyes off the ball elsewhere. But then combining this problem with Zobels and worst of all, the crossover becomes the biggest baddie of them all.
All I am saying, there are things going on here that needs to be looked at.
Click on this link, this is a crossover that tries to deal with these issues, the Elsinore Mark 6.
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-174.html#post3882471
Cheers, Joe
.
Not sure I can agree with that. Speaking for myself, when listening to any variety of 'popular' music I want to achieve an involving, even compelling reconstruction of the sound intended by the band/producer. That's generally a studio construct, as even 'live' CD's are heavily manipulated before issue, and taken more from the board mic feeds than from mic's in the hall. Voices, acoustic instruments, etc. within that mix need to be reproduced with fidelity. Reproducing PA sound is NOT my goal and I suspect that holds true for most other listeners out there. I suspect that the goal for bands and sound men is to get their PA to sound somewhat close to their studio sound!
The difference from purely acoustic music is that there is no direct reference to the studio sound, while we can get closer to that reference with purely acoustic music. IF we're lucky enough to live in a city where unamplified acoustic music is routinely available.
Bill
The problem with live acoustic music is that the hall/venue has many reflections and modes. Even the best halls in the world sound very different from the floor row 14 and 2nd balcony left side! I hardly ever get a seat from the sweet spot! Sure, I don't want to sit on the balcony at home listening to my classic recordings!
Thanks for your response. It isn't always apparent from the outset what posters mean just by reading their short responses. I was not aware of the "Elsinore Project" before encountering this thread.
While there seems to be a default focus on crossover electrical performance, there are a couple of issues that bring me back to the OP's question: "why do some speakers sound 'big'?" Instead of focusing first on the electromechanical properties of some assumed loudspeaker architecture, I would first focus on the loudspeaker as a black box (including driving electronics inside the black box), then understand the psychoacoustics of placing that back box into a room and dealing with human hearing characteristics that affect the sound of "big". Focusing first on some piece inside that black box seems to me to be diversionary to the original question. YMMV.
One of my comments above was related a piece of the "inside the black box problem": getting the phasing right on any crossovers involving multi-way loudspeakers. Nothing more esoteric than that statement. Eliminating crossovers and "ways" is another way to avoid these issues mentioned. Note that there is a trade in doing this and it includes at least increased modulation distortion, which seriously detracts from the "big" image.
However, modulation distortion is much lower for horn-loaded loudspeakers than is for the same drivers operating as direct radiators. I find it interesting that there is so much default focus on small box direct radiating loudspeakers for that reason.
There are other performance areas relating to "big": controlled directivity (implying flat FR across the auditory spectrum within the radiating sector of the loudspeakers), and the natural dynamic range of the loudspeaker without accompanying distortions.
Chris
Speaking out of turn, I'm not sure that Lynn wasn't saying what you are saying, i.e., some people are using a reference other than live acoustic music for their preferred "sound" of their home sound reproduction system. Correct me if I'm wrong, but I believe that you are saying that you are one of those individuals that don't use a live acoustic sound reproduction standard, i.e., not amplified.
I have to agree with Lynn: I can't imagine using anything other than live acoustic music as a reference. I don't believe that I'm alone--a guy by the name of Klipsch also had the same reference while designing and implementing his loudspeakers. Perhaps it is because of similar background in terms of music education or even just listening experiences, or perhaps it is because that any other standard used results in a "slippery slope of audiophilia" from which Tower of Babel scenarios immediately arise: if we cannot agree on a reference for sound reproduction systems, then not much constructive agreement exists at all for forums like this one...
Perhaps.
What KEF models? I have seen quite a few KEF measurements, but none like that.
.
Last edited:
Been around for decades even before then and haven't seen any such an example. Please look again, this is near flat from about 25 Hertz and upwards and I might, up to 100KHz and beyond - I am sure I would have noted it.
.
Back when i worked for a KEF dealer they only had one model that sounded "big." It was the KM1 that used non-hi end big motor drivers, a decent amount of bass surface area and active drive. The other models targeted at the high end home audio market sounded very much like the typical hi end junk sold, (vandersteen , B&W, wilson, Theil, ect) maybe a bit smoother since they didn't use off the shelf drivers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.