Twilo: I agree with you that the Mirage bipolars, while very popular in their day, were not completely satisfying speakers. A couple of things you could potentially experiment with in an opposed/twinned set-up would be slightly shelving down the volume level being directed to the rearward-facing pair, or mirroring a small toe-in angle front-to-back, both in order to fine-tune the imaging/soundstaging. |
I have always grooved to the sound of wide-dispersion speakers. I grew up in a household where my father's final listening room speakers were a pair of classic Allison: One towers, with their V-shaped front baffles that sported a dual compliment of their ultra-wide dispersion dome midranges and nipple-shaped tweeters angled 45 degrees apart, and whose flat backs were designed for on-front-wall placement to achieve a virtual infinite-baffle loading of the speaker/room interface. Although they didn't compete with today's speakers in terms of resolution and precision, their presentation was a kind of aural widescreen that energized the whole room to a degree (and even at low volumes) which you just don't hear from most speakers. I'd love to hear the new iteration of this landmark 70's design that was brought out a couple of years ago with renewed involvement from Roy Allison...
I find the sound of dipolar panels like Maggies, Soundlabs, etc. very attractive, until you try setting them up in a realistically-sized listening room ; conventional box-style monopolar radiators are just plain easier to place and extract optimum sound from in many instances. Bipolars can be another story becaue you don't run into as many phase-related difficulties, but I've never been overly impressed with most of the designs I've heard that achieve this radiation pattern through the use of separate, opposed drivers. Maybe something more along the lines of designs like the MBL Radialstahler, Ohm Walsh, or that forthcoming B&O superspeaker, which radiate through 360 degrees from single drivers lacking vertical baffles.
BTW, although I can like the sound of bi-, di-, and omni-polar speakers, I don't necessarily agree with the 'live music' argument against monopolar speakers in theory. Yes, live music is radiated in all directions (although not equally), but I don't view this as being directly analogous to the home playback situation. I am philosophically of the 'you are there' school, as opposed to the 'they are here' school. To be 'there', it is a good idea to minimize the reverberant contribution of the listening room to the reproduced acoustic, which means controlled-dispersion speakers. I see the job of the loudspeakers as being not the inverse of the original musical instruments and performers, but as the inverse of the microphones which recorded them.
The end result is that there often seems to be a sonic dichotomy between the sense of envelopment and presence, whether 'correct' or 'artificial', that you can get from speakers which directly radiate to more than just the frontal direction, and the sense of precise focus and scaled perspective you can get from monopoles. Both have their virtues, but in my listening rooms monopoles have always been the practical dictate. My present choice of Thiel speakers represents a monopolar design having wide dispersion for the breed with very even off-axis response tapering (at least in the horizontal plane - improving the vertical plane uniformity would seem to be one advantage of Thiel's newer coaxial designs) and a low-diffraction cabinet. Actually, I'm not sure that my idealized speaker design wouldn't be a monopolar line- or point-source, the former of which is represented by the Wisdom closed-back narrow planars, and the latter by the Cabasse tri-axial 'eyeballs', neither of which I've had an opportunity to hear, to my sorrow (not that I could really afford 'em anyway). Of course, one could always try installing a pair of Quad ESL's into a wall that's been cut-out to form a simulated infinite-baffle between two relatively small listening rooms... |
How about breaking it down to the theoretical fundamentals: What would be best if it could be perfectly realized in physical implementation (an impossibility)?
>A completely omnidirectional, evenly radiating pulsating sphere (or point, if you prefer)
>Same as the above, but a monopolar hemisphere only, maybe wall-mounted to simulate an infinite baffle constituting a listening room boundary
>A true monopolar, laser-like (parallel-focused) 'ray of sound' eminating from a point-source (actual or simulated) and aimed directly at one of the listener's ears (what would the difference be between this and a set of headphones?)
>A 360-degree radiating, cylindrical line-source with zero vertical dispersion
>Same as above, but a 180-degree cylindrical half-section only ('monopolar' line-source, analogous to the sphere/hemisphere example above), again maybe wall-mounted to simulate an infinite-baffle
>A dipolar version of of any of the above, such as a Quad ESL is to a simulated pulsating sphere
>None of the above: The ideal radiation pattern should be an exact inverse of the recording microphone's 'acceptance-field' pattern, whatever that may be (in realistic terms, this kind of thinking could only even begin to apply with a very tiny minority of recordings actually made, due to prevalent recording methods)
>None of the above: Given the preceeding, stipulate that you'll never be able to standardize and optimize the recording process to conform to some idealized encode/decode protocol with an equally-conforming playback system, and thus there is no one 'correct' ideal radition pattern possible, so you should just work with whatever sounds good to you in your room
>It doesn't matter, as long as you listen inside a perfect anechoic chamber, maybe in a multi-channel setup |
Twilo: The "RA" in RA Labs stood for Roy Allison - see my post at the top. I don't know if it was Dr. Allison himself who displayed those speakers in that back-to-back set up, but in any case, he didn't invent the configuration, as many hobbyists have played around with it through the years, often using similarly inexpensive small speakers. The opposed-twinned set up is not a cure-all, but it does let one investigate bipolar sound without needing special speakers. |
John: Apparently not! :-) I've forgotten more particulars about setups I listened to within the last year... |
Some thoughts:
>All speakers bounce sound around the room, in varying ways and to varying degrees, and that includes so-called 'point-source' monopoles. We need to be careful to properly distinguish among our terms: point-source, line-source, limited-dispersion, wide-dispersion, mono-, di-, bi-, and 'omni'-polar radiators, etc.
>No speaker is truly a point source, and neither is any microphone (and many do not try to be). A point source is a theoretical construct and cannot be achieved, only imperfectly simulated. But in theory, a true point source would be omnipolar. However, whether this would actually represent some kind of ideal receiving or radiation pattern is not necessarily a given, although it is often casually portrayed that way.
>The ear/brain does not function as a strict analog to a microphone. For one thing, the ear/brain can distinguish between a lot of the direct and reflected sound. What you hear in your listening chair is not what a microphone would record situated in the same spot.
>By the same token, the recording process is profoundly inadequate to capture what a person at the original event might hear.
>No matter what type of microphone or speaker is employed, we never even get within drive-by distance of the ballpark regarding a symmetrically complementary encode/decode record/playback process. The relationship is asymmetrical, arbitrary, unknown, and in practical terms unknowable.
>By the same token, even if the speakers could somehow function as a precise inverse of the recording microphones, since the listening room acoustics will always be an arbitrary imposition overtop of what can be captured of the performance space environment, we must acknowledge the incongruity in promulgating both that the speakers should mimic the behavior of the mics on one hand, and yet that those same speakers should attempt to eliminate or reduce the influence of the listening room acoustics on the other. You cannot really do either one very well, but you most certainly cannot even begin to do both simultaneously. (And arguments going to the supposed ability of either speakers or mics to somehow embody the physical properties of either instruments or singers is hardly worth commenting on, so fundamentally misguided is the idea.)
>We cannot lump together various di-, bi-, or 'omni'-directional speaker designs in rhetorical opposition to 'monopolar' or 'point-source' designs. The most popular variety of non-monopolar radiator is probably the dipolar planar kind, and this type of radiator can have less reflected sound than traditional dynamic/box speakers (not to mention bi- or omni-polars), due to its simulated line-source behavior limiting floor and ceiling reflections, and its 'figure-8' side-cancellation behavior limiting sidewall reflections, while the remaining front and rear wall reflections may be easier to treat without 'overdeadening' the entire room.
>An anacheoic chamber will not make a good listening environment primarily because recordings are not mixed and mastered by people operating in anacheoic conditions, and well-designed stereo speakers will take into account the fact that they will not be used in anacheoic conditions. If recordings and speakers *were* made to be listened to in anacheoic chambers, we would perceive the inadequacy of stereo to provide convincing reproduction and prefer some sort of well-implemented scheme involving more channels, coming from more directions (with the artificial exception of recordings whose original performance space was an anacheoic chamber as well).
>While limited-dispersion loudspeakers may represent one kind of virtue for obtaining accurate home playback, they can only do so for a single listener in a single listening spot. In the real world of homes and people, speakers having broad, even in-room power response will often be more practically enjoyable.
>The advantages of planar speakers are not solely defined by their radiation patterns; there is also the issue of eliminating box enclosure distortions. In the case of full-range electrostatic panels, there is the elimination of crossovers and multiple, frequency-divided drivers. You pick your poison - there is no one perfect solution.
I could probably go on, but I'll lay out for now. For the record, I use dynamic, box, monopolar, multi-point speakers intended to have relatively broad, even dispersion and low difraction, and to sum with minumum phase and time distortion at the optimal listening position (they are Thiels). This approach, like all others, has its advantages (some of them purely practical, some of them quite possibly purely theoretical) and disadvantages - and also like all others, fails in the end to achieve a realistically convincing portrayal of the actual thing. |
Summitav wrote: "Those who state that live recordings will not sound 'real' in an anechoic environment have not listened (properly) in that environment."
You are quite correct that I haven't had the opportunity to listen to anything, much less a good stereo, inside of an anechoic chamber. That's a chance that's tough to come by for most of us. I have no doubt that the experience would be revelatory in many ways. (I have listened, and worked, in studio control rooms where recording, mixing and mastering are done, know that these are not anechoic environments but rather controlled environments, and have prefered using ones - and gotten better results - where the monitoring options are not limited to just the nearfield.)
But I'll still stick by my contentions A) that a stereo system would sound best in an anechoic chamber if the speakers (and the recordings played through them) were designed with that as their intended environment, and B) that a well-implemented multi-channel scheme would sound more naturally convincing in that environment than would stereo (there's nothing sacrosanct in theory about limiting ourselves to 2 channels as some sort of ideal paradigm for sound reproduction, it's just much simpler to do well than a higher number of channels).
Those statements imply some corollaries:
>That speakers intended for home use will sound better if they are not designed solely on the basis of anechoic measurements, but take into account more typical listening room acoustics.
>That a well-implemented multi-channel scheme could also sound better than stereo in the home, but also that this would not only be highly dependent on the efficacy of the recording process used, but on closely controlling things like dispersion and room acoustics as well. Or in other words, the room properties, or distortions, that can actually make 1- or 2-channel reproduction go down more easily as a subjective matter, will become more problematic as we continue to add channels and speakers. The more we do to try and supply some semblence of the 'real' recorded performance space acoustic, the less we will be able to tolerate overlaying the arbitrary and unrelated listening room acoustic as a kind of a ameliorative substitute. |
Muralman1: The radiation pattern and phase relationship you describe is dipole, not bipole. |
As others have stipulated, the listening room's added reflections are distortions. But we need to always keep in mind that "what the mics heard" was insufficient to represent what a human listener at the event would have perceived, and that what they did hear was judged largely based on the engineer's experience of monitoring and mastering in control rooms that are neither performance spaces nor anechoic chambers, and about what type of finished product will sound best played back in a typical home listening room through typical home speakers. A recording is not an objective 'verite' account of what happened, but a subjectively molded account derived from certain common assumptions intrinsically embedded in the production and reproduction processes - one of which is that there will be some liveliness to the listening room.
This also goes to my point about a speaker's radiation pattern and the acceptance pattern of the mics not necessarily being commensurate, but quite possibly incompatible, if we want to minmize listening room reflections by limiting dispersion (and often even if we don't). The simplistic assertion, made by some speaker manufacturers, that speaker radiation must somehow mimic microphone acceptance is further undermined by the implicit, but false, assumption that all recordings will be minimally mic'ed and that we can even know the acceptance patterns of the mics used, much less their positioning. For recordings assembled from multi-mono multi-mic'ing, and of course for electrical direct-injection into a recording console, the relationship simply doesn't exist at all. A more pertinent relationship might be that of the radiation pattern of the mixing and mastering monitors to that of the home speakers, but of course that can't be a consistent thing. In the final analysis, the best way to assess speaker radiation pattern must be subjective auditioning in the room in which they will be placed, using the kind of music which will be played through them, by the person who will be doing the listening.
About Karl's point #2, although it's true to a certain extent that if the listening room could match the recording venue, then the overlayed refections would be more "consistent", they would still be just as distorting. And the problem would be worse the larger and livelier the two rooms were. However, for smaller and relatively 'dead' rooms, I think we have some evidence in favor of this hypothesis, however unimportantly. Many of us will own some of Rudy van Gelder's vintage Blue Note jazz recordings (mono or stereo), hundreds of the earlier ones of which were taped right in his New Jersey home's furnished living room, employing minimal mic'ing. Despite the facts that these recordings are bandwidth-limited, afflicted with dynamic distortions, and subject off-axis and less-proximate instruments to premature roll-off, nevertheless many can display a scary amount of the familiar quality of sounding especially 'real' when listened to from another part of the house outside the listening room, where their flaws are not only less obvious, but where the direct and reflected room sounds have melded into one indirect sound deviod of specific spatial cues, regardless of the speakers used. In this narrow sense, these records can rather easily exceed higher-fidelity concert hall or studio recordings. My theory has long been that this is precisely because they were recorded in an acoustic environment presumably very similar to most people's playback environments. Try it sometime... |
To me a true binaural recording uses a dummy-head mic, although again, in the end the playback system positioning remains a rather arbitrary and subjective thing by comparison. |
