 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
145.64.134.241
In Reply to: RE: This is why the live reference, attending live concerts, is paramount. posted by Todd Krieger on October 20, 2010 at 22:04:49
>It’s not so much the “sound field” that is of benefit from the live reference, but training the listener on timbral and dynamic characteristics of the real thing. A definitive reference that helps one migrate toward that sonic ideal with his home sound reproduction.>
Meyer shows frequency response measured at different locations in the hall for celli and first and second violins. Takatsu shows IACC (interaural cross-correlation coefficient) at various location in the hall.
The former relates to timbre, the latter to soundstage.
It would appear that also parameters such as clarity (C80) and lateral energy fraction are affected.
http://www.concerthalls.unomaha.edu/discussion/clarity.htm
http://www.concerthalls.unomaha.edu/discussion/spacious.htm
>Or in other words, if one hears an oboe on a recording, how would he know whether it sounded like how it was supposed to if he had never heard an oboe in the flesh?>
Even if you know how the real instrument sounds, the recorded sound reproduced via loudspeakers will sound different because of the very different radiation behaviour:
[1] Caussé et al., “Perceptual evaluation of a programmable radiation pattern loudspeaker simulating a real instrument”, 127th Meeting of the Acoustical Society of America 1994
[2] Caussé et al., Radiation of musical instruments and improvement of the sound diffusion techniques for synthesized, recorded or amplified sounds”, ICMC Proceedings 1995, p.359
[3] Misdariis et al., “Perceptual effects of radiation control with a multi-loudspeaker device”, Acoustics ’08, Paris, France
[4] Otondo et al., “Perceived influence of changes in musical instrument directivity representation”, Proceedings of the Stockholm Music Acoustics Conference, 6.-9. August 2003 (SMAC 03), Stockholm, Sweden
Once you realize these fundamental shortcomings of 2-channel stereo and the fundamental shortcomings of the usefulness of live concert hall experience and sound of musical instruments as reference, you realize that one should go to concerts for the pleasure of attending the concert rather than using the concert to train one's ears. YMMV, as usual.
Klaus
Follow Ups:
"Meyer shows frequency response measured at different locations in the hall for celli and first and second violins. Takatsu shows IACC (interaural cross-correlation coefficient) at various location in the hall."
These kinds of measurements do not relate to timbre, etc. They relate to objective physical properties, not subjective perceptions which are the result of extensive processing by the brain. The human hearing system is not a measurement device. It correctly perceives timbre so that the type of object creating a sound can be accurately recognized under a wide variety of acoustic conditions. This ability has obvious evolutionary benefits, e.g. recognizing food, mates, and threats.
Reproduced sound will not sound different if the recording/playback is tonally flat and undistorted and the correct acoustic environment presented. If the acoustic environment is warped, then the brain will make an automatic compensation and one of the ways this will manifest is that the tonal quality will sound incorrect. If the instruments fit comfortably in a typical listening room, there will be no great problem in reproducing them accurately with first rate equipment given an appropriate recording.
If one wants to produce and reproduce recordings that sound like live concerts one must go to concerts and lots of them. One has to become familiar with many different venues, seating locations, and ensembles and how these interact acoustically. If all one wants to do is babel about how it is impossible to reproduce the live concert experience, than none of this effort is needed. Success will never be achieved by looking at a bunch of measurements. Measurements are useful only when equipment is on a design bench or repair shop.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
> > Meyer shows frequency response measured at different locations in the hall for celli and first and second violins. Takatsu shows IACC (interaural cross-correlation coefficient) at various location in the hall. < <
> These kinds of measurements do not relate to timbre, etc. <
“Timbre is that attribute of auditory sensation in terms of which a listener can judge that two sounds similarly presented and having the same loudness and pitch are dissimilar” (American National Standards Association, New York, 1960).
The response curves at the different locations show differences of up to 5 dB, if you had two loudspeakers having two of these different curves as on-axis response you would not doubt for a second that they will sound different.
IACC relates to spaciousness, hence to soundstage.
> It [human hearing system ] correctly perceives timbre so that the type of object creating a sound can be accurately recognized under a wide variety of acoustic conditions. <
My daughter has a new flute, and I still have mine which is about 40 years old. You’d immediately recognize each of the two as flute, you’d also note immediately that they sound different. In the above mentioned paper Meyer addresses different seating arrangements: “ So the sound of the first violins is brilliant and very vigorous, while the sound of the second violins is a little dark and poor.” “The difference in timbre between the first and second violins (in the European seating arrangement) may be advantageous in passages where the same phrase must be be played by both groups alternatively.” No doubt you recognize both groups as violins, yet they sound different.
Meyer further elaborates: “ The greatest difference between the two groups (first and second violins) is measured by the microphone near the first violins, because the masking of the second violins by the musicians of the first violins is strongest here. The first violins are relatively loud here, especially in the lower frequency range, where they have no directivity. The higher-frequency sound does not reach this place in the angular range of preferred sound radiation, nor by a reflection from the ceiling, so the spectrum at this place has quite a different character than at other places in the hall.”
> Reproduced sound will not sound different if the recording/playback is tonally flat and undistorted and the correct acoustic environment presented. <
It must and will sound different because many musical instruments radiate different portions of their operating range into different directions, whereas loudspeakers don’t make such directional distinction, and because human hearing is directional, meaning that it makes a spectral difference whether the same frequencies come from the front, or from the sides, above or from behind. Searchers at French IRCAM (Institute for research and coordination of acoustics and music) have found that there is a perceptual difference between an instrument and its reproduction via loudspeakers and that when the directivity of the instrument is simulated by a dodecahedron loudspeaker arrangement, the perceptual gap is closed. This is the reason why directivity behaviour of instruments is analyzed, as to include the correct behaviour in acoustic modeling of performance venues (auralization).
> If all one wants to do is babel about how it is impossible to reproduce the live concert experience, than none of this effort is needed. Success will never be achieved by looking at a bunch of measurements. Measurements are useful only when equipment is on a design bench or repair shop. <
Without science and measurements you would not be able to hear a single note of your favorite music in your home, unless of course you were rich and had Karajan and Wiener Philharmonics come to your home and perform. If you reject science and measurements you should do that also when you GP uses both for diagnosis when you’re seriously ill, or when your audiologist wants to install the hearing aid with frequency response tailored to your particular impairment. If you think live concerts are useful as sonic reference for judging audio gear, be my guest. I no longer do.
Klaus
> The response curves at the different locations show differences of up to 5 dB, if you had two loudspeakers having two of these different curves as on-axis response you would not doubt for a second that they will sound different.
Yes, but as Toole points out, the brain is remarkably good at deconvolving the contributions of the room and the original source. Comb filtering of a signal fed to a loudspeaker doesn't sound at all like comb filtering from natural reflections. The brain seems to integrate multiple arrivals to evaluate timbre, and to use the comb filtering to judge the nature of the acoustic space and position of the source instead. In fact, reflections seem to increase the ear's sensitivity to resonances in the source, by IIRC more than 10 dB.
So, in practice, frequency response distortions in a loudspeaker will be heard as such, whereas frequency response distortions introduced by a room may not be, depending on their nature.
Also, arguably, there are factors in the perception of sound in a concert hall that are invariant with respect to the sort of response variation to which you refer. If there weren't, we wouldn't be able to tell immediately whether we were listening to live or reproduced music. Arguably, some loudspeakers do a better job of reproducing the cues that suggest that we are in a concert hall. OK, knock out the "arguably" since some loudspeakers clearly do. So while we can't create "you are there" realism, we can get part of the way there.
> > The response curves at the different locations show differences of up to 5 dB, if you had two loudspeakers having two of these different curves as on-axis response you would not doubt for a second that they will sound different. < <
> Yes, but as Toole points out, the brain is remarkably good at deconvolving the contributions of the room and the original source. <
The point I wanted to make is that the sound at different parts of the concert hall is different, so there is no such thing as a live concert hall reference sound. If you read those parts of Meyer's, "Acoustics and the performance of music" (Verlag Das Musikinstrument 1978) that relate to the sound field at different part of the hall, with reference to the different instrument sections, it becomes quite clear from text and measurements (frequency response, octave level models, sonagrams, octave level analyses) that there are perceptible differences. The reference is hence a moving target.
Klaus
Sure, but wouldn't the most accurate loudspeaker then be the one that reproduces the perspective on the recording? Which in most good recordings is an orchestra perspective, actually more like the perspective from an imaginary balcony over the front of the stage.
In my book the reference, in the sense of faithfulness to the original, is indeed the recording, not the live event. But whatever your reference is, the technical goals for loudspeakers should be clear: flat response, correct time alignment, low distortion, live-like SPL possible.
How are orchestras recorded? Close mike for every instrument, one mike for each instrument section, one or more mikes at different positions in the audience, what? Is the final mix representative of a particular seat in the hall?
Klaus
It varies, from a single stereo pair flown above the front of the orchestra to individual mics for the instrument sections and soloists. While it's possible to screw anything up, in general, the simpler the mic technique, the better and more natural the recording.
You can't put conventional microphones where the audience would be, it doesn't sound right. The microphone doesn't "hear" what the ear does.
So in the first case the mike would capture the sound similar to the sound at the conductor's place? No one in the audience knows how it sounds at the conductor's position, and worse, no one in the audience knows how it sounds at a position above the front of the orchestra.And in the second case the result is determined by what knobs the sound engineer turns and how he turns them, being similar to no position in the hall at all? No one knows how it sounds at this virtual position.
So what's the use of taking live concert as reference when no place in the audience is close to what the recording technique delivers?
Klaus
Edits: 10/26/10
I think you're making it out to be more difficult than it is. On a good pair of speakers in a reasonable acoustic, you'll hear pretty much what the microphone records. If the instrument is recorded from 2" away, it will sound like it's coming from the plane of the speakers. If it's recorded from 30' away, it will sound like it's coming from, if not 30', then 20', as it would in real life, since the brain tends to misjudged distance.
For a clear example of this effect, listen to the cow bell recordings on the Stereophile test disks. You can hear the distance of the cow bell as it moves from the front to the back of the stage. Different microphone techniques are used and while they produce different perspectives, it's a matter of precise location.
In practice, naturally mike recordings do a fairly good job of putting the orchestra on a simply miked recording where it should be -- at a distance from the listener, And they maintain, approximately, overall balance and timbre. The speakers aren't reproducing the sound from below, but from the front, and this reduces though it doesn't entirely eliminate the effect of the unusual microphone perspective.
In my experience, multi-miked recordings are less successful at producing an illusion of reality. They rarely sound natural, either in tonal balance or spatiality.
Sure, nothing can create a completely convincing illusion, particularly in the case of two channel stereo since two channels can't produce the sense of envelopment one hears in a concert hall. But the best recordings, played on the best speakers, have an uncanny ability to transport the listener into a different space.
Howdy
I'd like to complement you on your ability to stay on topic and keep your posts both useful and reasonable. I strive to do as well but often fail :)
In spite of the previous paragraph I'd like to offer a slightly different experience:
I have an amazing number of well recorded surround SACDs which give a much more realistic (in the sense of being there) experience than this thread might lead one to believe. (Don't get me wrong I also have the a lot of ping-pong MC recordings and completely artificial recordings: some fun, some not.) To me it obviously more realistic to hear the orchestra clearly in front of you while at the same time hearing applause all around you... I know it sound like being there because I have listened to a lot of live performances as well as participating in bands, choirs, etc.
In my case I simply have well engineered equipment and minimal processing: no time alignment, no room correction, no diversion of bass to a sub. I just put four speakers up at equal distance from the sweet spot on as big of a circle as would fit in my room. There is no issue of soundstage depth or width, you are just there...
-Ted
Thanks, I've been in so many Internet food fights over the years that I go out of my way now to keep things reasonable.
Which, however, isn't necessary here because I agree with you 100%. Two channel stereo can be very impressive, but there's no way it can transport you into the concert hall. At best, it's like peering through a window.
Also, as J. Gordon Holt used to point out, you can't get the tonal balance right on all material with two channel stereo. Flat response works well for small ensembles, but you need a 2-4 dB downward slope for orchestral material.
I seriously envy you your SACD collection! Multichannel is on my list, but I have so much to do just to reconstitute my two-channel setup that it's going to be a while before I can swing it. Maybe when I get my permanent speakers -- I've been using a pair of MMG's as temporary speakers since I can't fit my old Tympanis in my new room, and I'm thinking that I may end up using the MMG's for surrounds, if they fit.
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 10/24/10
> 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:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]