fwiw there's a waveguide I like for its ease and not sounding "canned" - look at the little cabinet on the left
An externally hosted image should be here but it was not working when we last tested it.
I'm utterly surprised that after 3000+ posts , the excellent Karlson speaker with coaxial developed in 1951 has not been mentioned
Can anyone explain how in this speaker the sound would not be perceived as coming from the floor ? The speaker is aimed at the listening position ? Sound is coming from the floor, right ?
have you ever heard about omnidirectional speakers?A strictly uniform directivity pattern over the entire audible bandwith is technically impossible - even with dipoles.
from MBLs and Duevels (and several others) to Graaf's "FCUFS" as well as the simple "periakusma", there are many diferent ways to get a uniform directivity pattern (on horizontal planes).
It would be more difficult to get a uniform response both in horizontal AND in vertical planes, if that's required. Not necessarily impossible, though...
Interesting. That's exactly how the "periakusma" have been designed. High freq. response is progressively increasing the more "off-axis" you go, that is toward the vertical (while being -obviously- uniform in any horizontal plane).
I guess it all depends on what you mean with "realistic". Of course it's impossible to recreate a physically realistic sound reproduction, that is a sound field which is physically equal (or at least very similar) to the original one. No doubt about that.
But fortunately that's not required. All that's needed is to create a convincing, believable illusion of that. You need not to fool your instruments: you need to fool your senses, your perception! Which is a whole different story. And fortunately something which is not too hard to do.
The only real problem is that we know way too little about our perception mechanisms to be able to do some real engineering on some solid foundations. What we perceive with our senses may be quite different from what our instruments tell us, and we do not have any good enough overall model to allow us to correlate the two things. Empiricism, trial and error driven by (possibly wrong or limited) theories, hypothesis and ideas are still the only way to go, unfortunately.
IMO the biggest error one can make is taking too many conclusions and making too many assumptions based on the little, limited knowledge that we have. Quite the contrary, we should always keep in mind Socrates's words: "I know that I know nothing".
Think to digital audio data compression: as long as engineers "tried to follow their instruments", designing compression mechanisms that focused on trying to keep (linear and non-linear) distortion low, they get only little gain in data rate reduction at the expense of bad to awful perceived performance.
Once that approach have been completely abandoned in favor of more empirical, psychoacoustic based approaches we've got mp3, aac, etc. That gave us significant data rate reductions with very little perceived degradation. In spite of the fact that measured performances (e.g. THD) of something like mp3 would be deemed absolutely unacceptable by just about anyone out there (who would think that an amplifier or even a speaker with a THD which at times can be more than 20% can be almost perfectly transparent to our senses?).
that would be a huge mistake... who really care about that? are you listening to your records to judge the work done by the sound engineer(s) who made it or rather to enjoy the music in them!?
Last edited:
To be honest, I don't think it's a mistake. The product of the musicians and the studio doing the recording, with the assumption that the musicians approve the product that comes out of the studio is the product that we're trying to reproduce.
So, we have two separate factions of audiophile here, one that prefers to "copy" the product that was arrived at and released by musicians and recording engineers, and another who would prefer to disassemble the product and rebuild it to their own tastes.
So, it's subjective. "Tell me the truth." vs. "Give me a beautiful lie that makes me feel better."
Arguing about which is better is a futile and never-ending endeavor, as neither side of the discussion is going to suddenly have an epiphany and change their subjective opinion.
So you are suggesting a method to make reflections non
correlated from the direct sound, am I wrong ?
That would be a method, which could at least "ease" the conflict in that neverending
directvity discussion IMO: Which DI, which kind of frequency dependency should be
the best ...
Separation of direct sound from reflections is a mental task (source separation).
Much of it relies IMO on narrowband (interaural) correlation between the envelopes
of the direct sounds of a source in natural music venue.
Reverb is decorrelated in that interband correlation of envelopes has diminished,
with rising number of reflections.
I see the "reverb on a recording" much like instant coffee:
You need some water (listening room reflections) until you can drink it,
otherwise it is too dry to taste like "reverb".
(btw. reverb in a natural venue is much more than just "delayed copies" ...)
Unfortunetely the recorded reverb is radiated from the front speakers of
a stereo system having the very same
-angle of direct arrival and
-dispersion in the listening room and
-undergoes the same reflection pattern in the listening room like the direct sound recorded
The side reflections in the listening room can make the reverberant portion more realistic
by increasing the energy from the side and thus also increasing IACC.
But you usually cannot avoid to convolve "perceivable listening room cues" into the direct
AND the reverberant portion on the recording. Also most listening rooms only work for
reproduction levels far below "realistic" ...
To avoid those listening room cues using common speakers would mean
- building diffraction/reflection free zones around the speakers horizonzally
- diffusing and damping also floor and ceiling reflections
The best solution would surely be, to reproduce direct and reverberant sound through
different channels. Then also specialized speakers could be used, where the dispersion
pattern is optimized for the given task
But if you have to reproduce both (direct and reverberant sound) through the very same
speaker, you will always have a conflict:
- A very narrow radiating speaker will sound too dry for most listeners.
- But a wide and coherent radiating speaker (true omni) will cause lots of
listening room induced artefacts, when used in an untreated room.
The early reflections will tend to make up a "spacious" but very undefined image.
A substantially different thing you can do, is making a speaker having non constant group delay
at larger off axis angles while maintaining flat group delay on axis, at least as a goal
to me formulated. (Things like "CD" are just goals also ... usually they are not implemented
perfectly.)
That behaviour is supporting perceived diffusivity of the listening room itself and thus supporting
the mental activity of source-reverb separation.
This way you can afford rather wide dispersion - to pour some water into the
"instant powder like reverb" on the recording - without risking audible cues from the
listening room too much.
Surely it is a different approach like "coherent speakers and extensive room treatment":
Because i believe that only few people will do that.
Nevertheless you are free to treat your room even using this concept, there is no exclusion.
Yes i use them and those have also been improved in the last year.
About two dozen listeners have visted my listening room during that time
- all of them for longer sessions - and i have evaluated their impressions
on certain settings, placements etc. I was gaining experience.
_____________________________
Some graphs ...
http://www.dipol-audio.de/model2-messungen-dateien/HorizontaleWinkel_FrequenzgangSchalldruck.JPG
Frequency response
0 Degrees off axis: Blue
15 Degrees off axis: Red
60 Grad off axis: Orange
Of course the frequency response has been modified by now, the drop in highs is compensated
actively >10Khz, but that should not matter in this discussion.
As you can see, this speaker has even slightly widening dispersion >4Khz.
http://www.dipol-audio.de/model2-messungen-dateien/HorizontaleWinkel_FrequenzgangPhase.JPG
Phase response (from same measurement)
0 Degrees off axis: Blue
15 Degrees off axis: Red
60 Grad off axis: Orange
Last edited:
The attempt to "copy" the product is only possible if you also copy the mixing studio acoustics in your room.
This is not possible, because it is not generally know where the product is mixed. And if in a rare case the studio address (and maybe even the room) would be know, you don't know how it sounds in there unless you work there.
So you have to settle with your own additions to the products by listening in your own room, which is of course pretty subjective.
The subject remains subjective no mather how one wiggles around it.
Usual furnishing will lead to absorbtion rising with frequency ...
My "Model2" speaker has widening dispersion above ~4Khz and diffuse
off axis radiation.
Of course this speaker sounds very different from usual 2-way cone plus
dometweeter designs, especially regarding listening room interaction.
But the highs have been rated as being "pleasant" (on pleasant sounds)
but also "realistic". Although this was kind of hard work.
I did not believe before that i would find a FR for the highs meeting
most if not all listeners taste. I am very close to it now and it is also
"in the ballpark" from a measurement point of view.
A person will usually not change his preferences due to the system he
is accustomed to (e.g. for years) within a few hours.
But that said slightly widening dispersion seems to make life easier and
also listener's ratings more comparable.
Last edited:
Not really. I guess it depends on how it's recorded. Some recordings are live sounding, like Chesky or Sheffield Lab types, where the goal should be to "be in the space" that they try to capture, where others are specifically engineered to be "dead", in which case, "bring it into your room" is what I'm after.
In any case, my goal is to do whatever the artist or engineer intends, which is usually accomplished very well by minimalism and transparency.
I usually don't run into issues until I try to force all recordings regardless of how they were engineered into my own pigeon-hole of acoustic desire. It's a lot like the act of EQing subwoofers for a single listening position...you make it worse everywhere else...if you design a system to overlay your own acoustic dream no matter what you're playing or who recorded it, you're going to have a small percentage of stuff that sounds fantastic..and if that's all you listen to, then *fist bump*.
If all you listen to is classical, by all means, over-engineer your room acoustics and loudspeakers with reverberations and delays and refraction and indirect sound until it sounds like a cathedral, but you can't really say that deadmau5 or Daft Punk sounds great in your cathedral.
I would subscribe to that. It seems we need some openness of mind
and even some tolerance for idiosyncrasies of recordings.
The reproduction chain's scope of influence is restricted.
But you can also make a kind of "informal quality definition" out of
the dilemma you sketched: A good reproduction chain should be
quite tolerant due to idiosyncrasies of recordings.
It is the "overstretched set of paramers" kind of chain, which will
behave "bitchy" when confronted with deviations from the norm.
a few musicians actually does that, and very few of them are audiophiles. Most of them are not trained to critical listening of reproduced music, and have basically no idea of what a good audio reproduction can or should be. Not to mention that the same record will never sound the same on different rooms and systems anyway.
wrong. As there is no such thing as "truth" in reproduced audio.
there's no such thing as "Hi-Fi", either. That's just a marketing buzzword, which have been created long ago just to mean those audio gears who did have specs better than a given minimum. That is a large enough bandwidth and low enough distortion not to sound like a bad entryphone or such.
It was all (and only) about a few numbers. Nowadays, even the cheapest portable player easily fit that definition. 
Taking that buzzword literally is just plain nonsense. Fidelity to what? How?
Talking about fidelity to what was heard on the mastering console is meaningless. You can get it only if you can access the original place and monitoring setup, or somehow reproduce it exactly ("room" included, of course). But then you can only get almost perfect fidelity to what have been mastered in a certain place, in a certain time and with a certain setup. Get a record made in a different studio and/or simply monitored with a different setup and you're screwed.
Not to mention the fact that the average end-user have had no chance to listen to the original masters of his records on the corresponding original monitoring setup, thus have no idea of how any given recording was "supposed" to sound like in the first place (that is, he have no reference).
Thus, how could the average end-user be able to judge whether his sistem is really "Hi-Fi" WRT something that he don't know at all?
How could all that make any sense? Of course it doesn't.
It would make sense to speak about "fidelity" (and "truth") if we could get something really close to the original reality. That is if we could recreate an (almost) exact "copy" of the original sound field present in whatever concert hall (or studio, etc) an event was recorded. But clearly that's not possible with existing technology (and possibly never will). Yet we would also need some means to evaluate and "certify" such "fidelity", or that would not make sense either.
Thus, people should simply forget about a fortunate buzzword which created a very unfortunate myth and start thinking about what matters. That is, enjoying music.
We should forget about a meaningless concept of "fidelity" and rather talk about "high quality" audio instead. Where of course "quality" refers to the perceived quality. Which is all that matters. And of course is, no doubts, highly subjective.
Audio reproduction is all about illusion. Fooling our brains into thinking we're hearing something very different from what we are actually listening to (that is real sounds, real instruments and/or human voices rather than an artificial sound field produced by a few electro-mechanical devices).
What we need is making that illusion to be as good as possible. The better the illusion, the higher the "fidelity", if you still want to use that (very inadequate) word.
This may be justifiable and correct for the room, but not the loudspeaker. The loudspeaker must be 'accurate" no matter what, but directivity is still an open question. Then one needs to look at the question of "being there" or "they are here", recreating an approximation of a "concert hall" or playing a two-channel original work as original work. Deciding this last aspect will nail down the design. Is it perfect? - no nothing like this can ever be perfect, but it is the most rational approach to optimization.
The only subjective question then becomes the type of reproduction that you prioritize. I know that you and I look for opposite things. I also know that there is much much more two-channel original works than there are concert hall recordings. But if that's your taste then so be it.
This is incorrect and an extremist point of view. The room is an influence but it is not as if the entire piece of recorded work is dominated by the room. Nuances yes, but major attributes of the art on the recording no. So one try's to get as close as feasible.
Taking your extremist point of view is paramount to giving up "all or nothing". "Its all correct if you like it." is what that approach comes down to because it is all flawed. Which makes an IPhone and insert headphones the supreme audio playback system.
Elias,
You don't believe that cars (as just one example of marketed products) are built according to the tastes of their construction engineers and test drivers. And you don't believe that those cars only work well on the roads, they have been tested on.
Like cars, music recordings are produced and sold with a target group in mind. The listening tastes, habits and facilities of each target group are well enough known. And the opinions or abilities of the chief production engineer are no more called for than the opinions or abilities of the star car designer.
Any beliefs, that the mixing concept of a single person or the acoustics of a single production room have a special meaning in the final product, is romanticizing.
The pure data are the original, and they don't carry any recollection of what had been before.
If by this you mean that it should add or subtract nothing, then I completely agree. Otherwise, I am not sure what you mean. How does one measure "tolerant".
sorry, perhaps I have not been clear enough. In a sense, basically this was also one of my points. Of course I was exaggerating, bringing the idea of "fidelity" to its extreme consequences to try to make it clear how nonsensical it is (if taken "literally").
here we are to the problem, again: close to what? close to something that is completely unknown? (at least to anybody but the few people who made the mastering...).
And, given he has no reference, no idea about how it "should" (how it was supposed to) sound like, how would one know whether his system is actually "close" to that or not?
OTOH, we all know how natural sounds "sounds like". And we are all able to instinctively recognize whether any sound is "right" or "wrong". That is, whether it is (seems) "natural", lifelike, or on the contrary it "has something wrong", is "not like it should be". We can all tell that by hart.
Faced with a sound from an unknown source, in most cases just about anyone is able to tell whether he is hearing a "real" sound or a reproduced one. Making it impossible (or at least harder) to tell apart a reproduction from "a real thing" is what I would call "high-fidelity" or, much more properly, "high-quality" audio reproduction.
Yet that have basically nothing to do with how the recording was "supposed" to sound like, thus have nothing to do with "truth" or "fidelity" in that sense.
like it or not, indeed it is. Normally, people listen to music 'cause they enjoy it. Whatever makes someone enjoy it the most is the best (for him). Sic et simpliciter.
Why should I listen to something I don't like, or that I like less than something else?
Are we in the business? Should we listen for our own pleasure or to "analyze" how a recording have been made?
Having a choice, I would choose whatever I like best according to my own ears. Nothing else matters (cit.). I couldn't care less if I'm told that's wrong because of some nonsensical "fidelity" or "correctness" considerations.
I bet that just about anybody else in fact does the same (unless so strongly conditioned to do otherwise. But that's the power of marketing, faith and belief... and it's a whole different story).
Last edited:
well... define "accurate". With respect to what?
as far as this takes into account human perception, that is if we try to optimize (the complete system) with respect to that, yes. Optimizing any part of the system with respect to some more or less arbitrary specs (as is most often done) is all but rational...
well... define "accurate". With respect to what?
as far as this takes into account human perception, that is if we try to optimize (the complete system) with respect to that, yes. Optimizing any part of the system with respect to some more or less arbitrary specs (as is most often done) is all but rational...
Accurate, as in your subjective opinion does not matter. Measurements trump opinions.
I have heard this dogma oh so often. But in the end, nothing but accuracy has ever been shown to have wide appeal. The rest is just "It sounds good to me so it is good." That's fine for you, but I am after something more global and universal. Something which is not just this weeks fad like handbags.
Exactly. For some reason it seems that very few people approach the 'business' of sound reproduction from this angle, which is why it typically sounds "wrong". And in many cases very, very wrong.
The trouble is that speaker people are trying to bend and mold sound damaged from its passage through the earlier part of the reproduction chain, to make it sound better, which always ultimately fails in some areas. The sensible way of achieving the end goal of convincing sound is to consider every part of the replay system as being critical, favouring none above the other.