|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.145.45.236
In Reply to: RE: Measurements are not equivalent to perception posted by KlausR. on October 23, 2010 at 23:00:32
Record playback is it's own art form, and recorded music is a thing distinct from (though obviously related to) live musical performance in a natural setting. All I have to do is listen and this fact becomes obvious to me. And while I came to this conclusion by listening (not by reading about it), I do not consider myself to be unscientific. But I will say that the gap between live sound and recorded sound played back through two loudspeakers seems to close as the musical event portrayed becomes simpler and/or smaller in size. Orchestral-sized events are the worst "live reference" one could possibly choose if one wants to fine tune one's system, in the hope that our loudspeakers will one day sound like live music. And while I am not rich enough to own (or even audition) the finest and most expensive loudspeaker systems in the world, I doubt that the very best loudspeakers are essentially better than the more affordable ones at blurring the distinction between "live" and "recorded" sound. The best loudspeakers might distort the recording less than the cheaper ones will, but they can never compete with or replace live music. Loudspeaker playback can sound real enough to suspend disbelief at times, but for the most part it is it's own strange thing.
Edits: 10/24/10 10/24/10Follow Ups:
> But I will say that the gap between live sound and recorded sound played back through two loudspeakers seems to close as the musical event portrayed becomes simpler and/or smaller in size. Orchestral-sized events are the worst "live reference" one could possibly choose if one wants to fine tune one's system, in the hope that our loudspeakers will one day sound like live music. <
I dunno. Some loudspeakers are clearly better at reproducing large ensemble works than others. Such as large line sources. And anything that can play loud, deep, and clean, with a uniform on-axis and polar response. Sure, we can't completely reproduce the concert hall experience, particularly with two-channel stereo, but we can produce a remarkable sense of being transported into another and larger acoustical space, one in which the walls of the listening room seem to melt away.
I think it was Earl Geddes who said that when the performance venue has about the same size and acoustic characteristics as the playback room then there might be a chance that reproduced music sounds like live.
> I doubt that the very best loudspeakers are essentially better than the more affordable ones... <
At blurring the distinction between "live" and "recorded" sound or otherwise. Loudspeakers have different radiation behaviour as compared to musical instruments so the perceptual gap will remain open, regardless of price.
As for the question if the very best loudspeaker are better than affordable ones, the answer depends on who you ask! Floyd Toole has obtained objective criteria for subjective loudspeaker quality. He said (some 5 years ago or so) that loudspeakers meeting these criteria can be made for something like $800/pair, you'd need to spend more for deeper bass. If I look at high-end loudspeakers in any price range, many of them perform poorly, according to these criteria.
We all know from experience that loudspeaker playback doesn't sound like the real thing, and science simply gives us some possible explanations.
Klaus
The problem with Toole's theory is that AFAIK there aren't any $800 loudspeakers that sound as real as the best ones. Or if there are, I wish someone would let me know, so I can buy them. I don't see how there could be, because AFAIK the technology to make a state-of-the-art loudspeaker at that price doesn't exist.
Toole has tried to correlate objective measurements with subjective listener preference. He found that a particular behaviour of amplitude response (on and off-axis) is preferred in blind listening tests. What he says is that there is no need to spend big money in order to obtain this preferred behaviour, hence the price figures indicated.
What is state of the art in loudspeaker design? In Toole's book its flat response on-axis and smooth behaviour off-axis. To that add a clean waterfall plot and correct time alignment. That whole package possibly cannot be made for $800, especially if on top of that you want high SPL with low distortion.
Klaus
That's exactly the problem I have. As far as I know, it isn't possible to make a loudspeaker that covers the whole audio range, has smooth on-axis response and good dispersion, and plays cleanly at a level that's adequate to reproduce all performances of un-amplified acoustical instruments at that price. Also, while Toole's criteria are good ones insofar as they go (although as I think he himself points out, it's an open question whether time alignment is necessary to the reproduction of music as opposed to test signals), I don't think they're the only criteria that affect loudspeaker performance. For example, state-of-the-art imaging and soundstage depth are difficult to achieve: arguably only a few esoteric line sources do it, or come close, without compromising maximum output. Similarly, there are issues of non-linear distortion, diaphragm breakup, and room interactions. In the absence of acoustical room treatment, dipoles and cardioids seem to have certain advantages. The upshot is that state-of-the-art reproduction seems to remain the province of large, expensive, and esoteric devices -- ribbons, line sources, large electrostatics, and dynamics that use exotic cone and cabinet materials and construction and servo woofer control. What's more, there doesn't seem to be a loudspeaker, no matter how elaborate or refined, that's state-of-the-art in every respect.
Not at that price range, no. But you don't need to pay astronomical prices either. In my opinion the speakers of the link below come close to the ideal. They use waveguides and DSP, no magic ingredients.
Time alignment: whilst not absolutely necessary, it makes a difference, even to the tin ear that I am :-)
"State-of-the-art imaging", what exactly is that and how do you know that the speakers are delivering those particular goods? What is the reference when judging imaging?
Soundstage depth: this would relate to distance perception in rooms. Nielsen in AES paper 3069 references some of the relevant literature. His introductory comments are quite interesting: "Sound reproduced through a normal set of stereo loudspeakers or a pair of headphones often lacks the impression of depth. This may be due to improper recreation of the oroginal sound field at the listener's ear. However, it may also be because the perception of depth by the hearing is not as good as we expect it to be. A reproduction cannot be better than the reality."
State-of-the-art reproduction: it would appear that only wave field synthesis is capable of recreating the original sound field, but that's a technology I'm not at all familiar with and anyway, there seems to be a problem in terms of home user acceptance.
Klaus
Blumlein refers to the reproduction of depth in his original stereo patent, and I've seen a stereophonic demonstration film he made in the 30's in which the speaker's distance from the microphone is reproduced to spectacular effect. That affect won't occur to any significant degree with a monophonic recording. So while I think it's probably the future of sound reproduction, wave field synthesis isn't necessary to create a sense of depth in the stereo sweet spot, beyond the crude approximations of stereophony, anyway (MS stereo is really first order WFT with a single sample).
Room reflections seem necessary for this. In fact, according to Toole, first-order Ambisonics in an anechoic chamber didn't provide any sense of front-back localization at all! The same is true of headphone reproduction of conventional stereo recordings, which also lacks any contribution from the HRTF. That being the case, it's not surprising that it doesn't reproduce front-back perspective. That can be introduced in headphone listening with head tracking and real-time room and HRTF emulation -- see the latest issue of Stereophile for an interesting example.
To the best of my knowledge, very little is known about the psychoacoustical basis of depth perception. Toole's book, for example, essentially says "We have no idea." The reproduction of depth has something to do with recorded ambiance and listening room reflections, and it seems to be best reproduced by loudspeakers with uniform polar frequency and, according to one source, phase response.
AFAIK, to reproduce depth, two things are required: recorded reflections, and the ability to reproduce it in the listening room at an angular separation from the loudspeakers. Headphones and anechoic chambers don't satisfy these criteria. Neither does a loudspeaker with non-uniform frequency or phase response -- the issue here is apparently that the brain must be able to relate the reflections to the original source to interpret them as reflections or, perhaps, that spectral distortions introduced by the reproducer or room interfere with the brain's ability to localize the sound and gauge the acoustic on the basis of comb filtering and the HRTF.
The speakers in your link measure beautifully, and I'm all in favor of the use of DSP in loudspeakers. IMO, the high end industry is hopelessly stick-in-the-mud when it comes to the application of new technology. But, and it's a big but, I don't think a handful of measurements can completely characterize the audible performance of a speaker. The measurements aren't comprehensive enough and don't entirely reflect behavior in an actual listening room, and even if they were and they did, our understanding of psychoacoustics is too primitive to correlate the measurements completely with what we hear.
"I think it was Earl Geddes who said that when the performance venue has about the same size and acoustic characteristics as the playback room then there might be a chance that reproduced music sounds like live."
Exactly. It works great with one or two instruments in a large living room with full range speakers that have very flat on axis and power response The fine tuning required adjusts balances the remaining errors to a certain extent, e.g. by making minute changes in microphone position and aiming one can balance things off to the necessary tolerance.
If one moves about in the listening room one gets a different response from the live instrument. Once the differences between live and recording are down below this threshold then there won't be significant differences between live and recorded. Indeed, that may have been part of the "trick" when I did my piano experiments, since the reproduced piano was about 8 feet to the right of the real piano and along a wall, not in the corner. So of course it sounded slightly different, but so would the actual instrument if it had been moved. One needs to reach a threshold where the willing suspension of belief can take place. If one is overly obsessive one will never reach this point. However, if one has this kind of personality it is likely that one finds most activities frustrating and unsatisfactory and the solution for such a miserable life is not to be found in equipment, technology or other material possessions.
It goes without saying that the designers of the microphones and speakers (and other equipment) used science and measurement in their development and manufacturing process. However, science and engineering can only go so far and that is the point where the art of recording and reproducing music begins.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
The way I see it, different types of loudspeakers specialize in different types of radiation patterns and they exist in order to cater to the personalized preferences of different types of listeners. There may be loudspeaker designs that qualify as "best overall", but the "jack of all trades, master of none" ideal may not appeal to everyone. Certain types of loudspeakers seem to capture the radiation patterns, dynamics, and timbres of certain types of instruments better than other loudspeakers do. If you like to listen to those types of instruments then you are free to specialize. And of course, certain loudspeaker designs are more expensive to produce than others are. Your choice of "favorite loudspeaker" might depend on what you consider to be objective criteria or it might depend on the kind of music you like to listen to... and, how much you want to pay! More expensive loudspeakers are definitely required if we want the deepest and most accurate bass, regardless of the topology. But I think many of us (consciously or unconsciously) accept the fact that as we optimize our systems for certain aspects of playback (or even for "best overall" level of playback), we simultaneously sacrifice playback capabilities in certain ways.
Edits: 10/24/10 10/24/10
Radiation patterns of loudspeakers and musical instruments have absolutely nothing in common. A picture says more than 1000 words, they say, so I have prepared an overview (Word file, for the time being in German) of patterns of different instruments, taken from literature I was able to locate. Feel free to drop me a mail off-board for a copy.
As Toole has found it's not so much the type of pattern that matters as how well the response curves behave. Good behaviour is not a question of money, but of good engineering. Geddes speakers are an example.
Klaus
Absolutely nothing? Or, is it a question of degree? No loudspeaker mimics the radiation patterns of musical instruments exactly but certain loudspeakers would seem to come closer than others in some instances, and sometimes that edge in similarity is what tips the scale in that speakers favor in those instances. The large radiating surfaces and dipole dispersion patterns of my old Magnepan 1.6 speakers had nothing to do with how realistic piano music sounded when I used them for this purpose? Are you saying that in-room frequency response was the only reason that piano music sounded so lifelike with these loudspeakers? I guess it's possible that frequency response is the only thing that matters in every case, but somehow I find that hard to believe.
Edits: 10/24/10
I don't think that there's any evidence that frequency response is the only thing that matters. And a lot of evidence that while it's extremely important, it isn't.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: