|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
Without being a speakerbuilder and not knowing the fundamentals behind the art, my layman's impression is that most world class imagers are always pretty narrow (10 inch or less). But then I wonder, some of the $10K plus speakers are pretty wide but yet they image so superbly. So then I wonder, how big does the width of a speaker really factor into the imaging ability. Is it the components and the crossover that do the trick? But if that is the case, again why are the narrow speakers such good imagers?
Follow Ups:
Some speakers are narrow for the tweeters and wider for the woofers. Ariel 10T's come to mind as do Dunlavy Millenniums.
in regard to ribbons. If a point source is ideal for imaging, then how can the tweater of something like the Newform 645, which is 45" tall, image at all?
A point source isn't necessarily better than a line source - it still depends on the execution of the design. For example, a point-source approximating speaker with poor phase reponse and/or system resonances will not image as well as a line source system with good phase response and freedom from system resonances.But let's compare apples with apples - let's say we have a perfect point source and a perfect line source. A point source will probably give you a bit better image localization if you sit in exactly the right spot, but a line source might very well sound more realistic. Let me explain:
First of all, a line source is not going to have reflections off the floor and ceiling like a point source will. So right away we eliminate that source of unwanted reflections.
Second, a true line source is much more tolerant of variations in listener height - you can listen standing, sitting, or lying down and it still sounds right. I like to get up and dance (when no one's looking, of course), and that's when I appreciate a line source the most.
Third, a line source creates a more realistic feeling of acoustic space. This is because the sound field set up by a line source is more like what you would get at a live performance, where you are probably fairly far from the performers. In the tyical room, you might be 10 feet from the speakers. As you move around the room, the volume changes noticeably. With line source speakers, the volume is much more uniform throughout the room. This is much more like the soundfield you get say 40 feet from the performers at a live concert.
If you have the misfortune of listening off axis, again a line source speaker will give you better soundstaging because the volume of the farther speaker will fall off more slowly.
A hybrid that seeks to blend a point source woofer with a line source mid/tweet ribbon or panel is a delicate juggling act. In my experience the most successful hybrids incorporate some sort of level control to best match the relative levels of the drivers.
Maggies are probably the most common line-source approximating speakers. If you get a chance to listen to a large pair do so - the sound field they generate really does have a very different feel from what you get with point-source approximating speakers. And if set up right, they can image and soundstage quite well.
Not everyone will agree, but most will, that the Newform 45" or 30" ribbons image wonderfully. I don't own Newforms but am getting ready to build a DIY based on this ribbon. A friend of mine recently purchased a pair of R645's and the ribbons sound and imaging are simply stunning. Listen for yourself and you be the judge.
> If a point source is ideal for imaging, then how can the tweater of something like the Newform 645, which is 45" tall, image at all?I would have to agree that they don't [image] very well at all. There are line sources that do image very well, but that's another topic.
I think Duke pretty well described the physics and psych-acoustical effects.And narrower speakers do seem to have an edge in imaging. But there are some notable exceptions and not just at the high priced end. The Boston line back in the late 70's and early 80's were moderate priced, had very wide baffles, and imaged very well.
In addition to the Haas effect, there is another bit of physics. Due to defraction effects, the baffle edge or grill can bend the sound around the cabinet, giving rise to directional effects. Speakers with the drivers offset to one side will tend to have a polar radiation pattern shifted to the near side. Roughly equivalent to toeing the speakers toward that side. This effect can and has been used to affect the imaging. Both by designers, and by knowledgeable people in installing or setting up systems. And again, this is independent of speaker cost.
Let me share my experience.
Although my speakers (ATC SCM20) images quite well, all the sound stage appears 'between' the speakers. I used to experience a wider soundstage from other speakers I had. This is confirmed by what Neil Gader of TAS heard of the same speakers.However, after reading ATC manufacturer comment for that piece of review in TAS, he said the grill was actually designed to counter the cabinet edge diffraction and intended to help imaging. NG prob listened to this speaker without the grill, so as myself and any other self respecting audiophile.
So I tried putting the grill on, can't believe what I heard. The soundstage now expand to at least 2 feet from each of the speakers.
Try A-B-A comparing again to make sure that I wasn't imagining things. I wasn't.So for wider baffle speakers, try attaching the grill. You might lose the ultimate transparency, but you might not be listening to what your speaker are capable of.
happy listening!
Hi Mesh,Your question is a good one, and I agree with your observation that in general narrow speakers do out-image wide ones. Let me try to explain what I think is going on. First a little background...
One of the most fascinating properties of human hearing is a time-gate effect that uses the first .62 milliseconds of an impulse to determine the direction of the sound source, and then largely ignores directional cues from anything arriving after that first .62 milliseconds. This phenomenon is called the "Haas effect", after Helmut Haas, the man who first described it. This .62 millisecond time interval corresponds to the maximum time it would take a sound impulse to travel the roughly 8.5" around the head from one ear to the other. Repetitions of the original signal arriving after that first .62 milliseconds (i.e. reflections) contribute primarily to the perceived timbre, although distinct lateral reflections (off a bare side wall, for example) can cause image shift.
As an aside, let me give you an example of the Haas effect in action: You and I are at opposite ends of a fairly large, reverberant room - in a museum, for instance. You close your eyes as I speak to you. Now, because of the distance between us and the reverberant nature of the room, probably over 95% of the sound that reaches your ears from my mouth is reflected sound. Yet because of the Haas effect, you ears ignore directional cues from those reflections and even with closed eyes you can instantly and unerringly point your finger right at me.
Okay, back to speakers. My point is, that first .62 milliseconds (corresponding to 8.5") is critical to imaging. Any reflections off screws, nearby drivers, or the edge of the cabinet serve to smear the image. Cabinet edge diffraction is usually the worst offender. The most severe smearing of the image would come from reflections or diffractions from discontinuities close to that 8.5" distance (or .62 millisecond cut-off point), while a discontinuity closer in wouldn't smear the image as much. So, either a very narrow speaker, or a speaker with a very wide baffle (like two feet across or so) would theoretically image the best.
It is possible to design a curved or bevelled-edge baffle that largely eliminates image-smearing diffraction. The best example of this approach that I know of is the old Snell Type A, one of the finest speakers of its day, and the first speaker to really show me how much fun first-class imaging can be.
Very few speakers are wide enough to escape having a cabinet edge within 8.5" of a driver. More than likely, a fairly wide speaker that images exceptionaly well either uses bevelled cabinet edges, or has sound absorptive material on the baffle, or uses drivers whose radiation patterns are directional enough to pretty much miss the cabinet edges.
Of course there's much more than baffle geometry involved in designing a speaker that can image well - phase response, alignment of driver acoustic centers, driver behavior (particularly in breakup modes), grille frame, enclosure resonances, mechanical rigidity and even tonal balance all play a significant role.
Hi,A question related to baffle design: would it then be best to make a cylindrical cabinet with the driver at one end? I assume there will be back-waves so perhaps a conical/spherical opposite end would be called for?
I am contemplating removing the drivers from my KEF Q15.2 (reflex cabinet) and building a smaller sealed cabinet. What basic rules should one follow as regards the front baffle and dimensions? The really nice thing about these KEFs is the concentric tweeter, so it calls for good cabinet design to take maximum advantage.
Thanks in advance,
Adnan
I wish it were as simple as minimizing baffle size. But, there is another factor that figures in, and that is the response step where the speaker goes from essentially hemispherical radiation (at short wavelengths) to essentially spherical radiation (at long wavelengths). There will be a 6 dB step-down in on-axis response, and the designer has probably taken this into account.If you really want to go all out in building new boxes, I'd suggest you keep the front baffle pretty much the same size but add large bevels around the edges, and round off the sharp corners. If it was me I'd probably build tall, narrow transmission line boxes with bevels on top and on the sides at least up near the drivers. Transmission lines have the cleanest midrange due to their relative lack of internal reflections back into the cone.
Best of luck!
Umm...why exactly do you want to replace the cabinet of the KEF Q15.2? I ask, because I am sure that you must have considered that the driver(s) in question are optimized for the box that they are in and yet, you seem to feel that you can gain something by placing the drivers in a different cabinet.A small box like that of the Q15.2 will not have big problems with diffraction effects. The small surface area of the box is actually ideal for avoiding these types of problems. The Q15.2s already image very strongly by controlling treble dispersion, essentially by mounting the tweeter in a shallow-horn, which acts as a wave-guide and which also matches the tweeters dispersion pattern to that of the woofer at the crossover point. This is why the crossover between the drivers is so seamless in a dual-concentric design.
My advice: leave the drivers in the cabinet. That design offers very little room for improvement.
Hi,The Q15.2, as a ported design, attempts to strike a balance, as so many bookshelf speakers seem to try to do, between small size and low-frequency response. On this model, there are slight colorations in male vocals that can be easily picked out.
KEF's center channel speaker, the Q95c.2 uses the same driver in a closed box of half the volume as the Q15.2. The result is that the Q95c.s does not go as low as the Q15.2 but vocals and imaging are even better, in fact, the coloration practically disappears.
I called KEF and they admitted that the Q15.2 has that problem and also agreed that the sealed box is superior. The sealed box offers a -3db point at ~80 Hz, which is perfect in an HT application with a subwoofer.
I would use the Q95.c all around except for two issues - the cabinet shape is unappealing for anything but TV-top use and KEF, in their infinite wisdom, have chosen to price it the same as a pair of the Q15.2s.
I'd like to make custom cabinets that blend in well with my room and having made cabinets before, the mechanical aspects don't worry me. Are you suggesting that I keep the front baffle dimensions the same and just make the cabinet half the depth?
I wouldn't attempt it unless I was fairly sure that I'll see an improvement. The concentric design with its superb imaging and dispersion is what attracted me in the first place.
Regards,
Adnan
I'd be curious to know what you think the source of the coloration is. If it is spurious resonance from the port, you should definitely consider plugging up the port, before replacing the entire cabinet. Do you believe the volume of the cabinet to be too large for the driver in question? Unlikely, but if you feel this to be a problem, you can partially fill the cabinet (perhaps with sand) until you acheive the desired volume.Replacing the cabinet seems like an interesting experiment, but in this case I think that you will end up with a no improvement or perhaps even worse colorations.
Hi,The cabinet volume is probably fine for a reflex design. If I were to simply stuff the port, the cabinet will be too large for a sealed design.
I'm not sure exactly where the colorations are coming from. It is most likely the port but there may also be improvements possible by adding bracing and using Black Hole 5.
The good thinag about this project is that if the new cabinet fails to offer an improvement, I can always reinstall the drivers in the original cabinet and call it quits.
Regards,
Adnan
In addition to a port resonance, you might also be dealing with a coloration (null?) produced by the interaction of the driver and it's mid-range backwave (perhaps the cabinet needs more internal damping) or perhaps it is a boundary effect caused by the speaker's proximity to nearby boundaries. I agree with your final point. No harm in trying. Would be curious to know if the experiment is a success.
The AR302 discussed below is a wide-baffle stand-mounted speaker. Set up as supplied with naked cabinet edges, imaging is rather vague (as with other wide-baffle speakers). Once an absorbtive material has been applied to the cabinet edges (in this case felt), imaging is more in line with that of small minimonitors; images are placed with much more precision within the soundstage. Clearly, treble dispersion and particularly the re-radiation of treble (and upper mid)frequencies from the baffle edges has a strong effect on imaging.
Perhaps it has to do with the price... I've been told that a typical speaker contains 20% of it's retail price in materials.. yup a $2000 set has $200 worth of electronics and lumber in it. Rest is overhead, marketing and middlemen costs. This is a common ratio, not soley restricted to Audio.
It's apparently much easier and cheaper to build a decently imagining 'narrow' speaker, than a larger one... judging by what's on the market.
A decent larger system involves significantly costlier drivers and x overs, a more robust and more carefully designed enclosure, as well as the engineering to make it all come together, but for your extra $$ you do get a larger richer soundstage and greatly increased low end to boot (usually ;-) No real surpise is there?
Hi Folks,While this opener is quite simplistic, it does illustrate a fundamental truth. As Duke mentioned, the tendency of a baffle to direct radiated energy into the forward hemisphere is both frequency-dependent and baffle-size dependent. At higher frequencies, where the baffle dimensions are comparable or large in comparison to the radiated wavelength, the baffle will function well to restrict radiated energy into the forward hemisphere. At lower frequencies, where the baffle is small in relation to the radiated wavelength, the radiation becomes essentially uniform into full space. This is the so-called "diffraction step" phenomenon - which results in an asymptotic 6dB change in SPL (3dB from a change in efficiency, and 3dB from a change in directivity) measured in the forward hemisphere.
It is true for transducers as well as for baffles, that the shortest dimension of the surface under consideration tends to dominate the speaker's spatial dispersion of energy. For this reason, narrow speakers tend to have a high frequency of transiton between half and full-space radiation, rather regardless of the speaker's height. That's why even very tall but narrow speakers (such as the Pipe Dreams, to name but one of an inumerable pedigree,) have a high 2-pi/4-pi transition frequency.
But now to get back to narrow speakers, and their imaging properties: one consequence of a high half-space/full-space (i.e., 2-pi to 4-pi steradian space) transition frequency, is that much of the speaker's power spectrum diffracts around the enclosure and is radiated to the rear of the speaker. Such energy typically encounters a reflection from a rear surface (such as the wall behind the speakers) within several milliseconds of emission. This reflected energy reaches the listener's ears and is perceived in combination with the speaker's direct radiation by the ear-brain mechanism, as an apparent net increase in the depth (and possibly width when one considers tertiary and higher-order reflections) of the recreated soundstage.
The overall effect can be quite remarkable, especially if the rear surface is a "good" reflector, such as a bare wall. Of course, the interference between the direct and rear-reflected sound will create a comb-filtering filtering effect too, which has been shown in and of itself to have "interesting" soundstaging properties in many cases.
In this respect, narrow speakers are similar to bipolar or dipolar speakers: much of their imaging characteristics can be attributed to their pronounced rearward-directed radiation.
As an aside, another factor which may contribute to the advantage of a narrow baffle, is that it's easier to design a "quiet" narrow baffle than it is a wide one. Secondary radiation from the vibration of the baffle itself tends to degrade the speaker's impulse response, which in turn makes for poorer "microdynamic" acoustical performance, which yet in turn tends to degrade the precision and specificity of the recreated soundstage.
Hope this helps - best regards,
MAP
This post is made possible by the generous support of people like you and our sponsors: