Agree entirely, but with this caveat . . . there is something to be gained by reducing/removing false clues and other errors that disrupt the "psycho-acoustics" of listening, and make it all the harder to suspend disbelief and just enjoy the music. It doesn't have to sound like the "best seat in the house" or the best hall in Europe or the best orchestra in the known universe under the best conductor since god created conductors, but if the oboe comes in too loud and front-and-center because some engineer twiddled the knobs you just *know* it's wrong . . . and there goes the illusion.
It doesn't have to be "perfectly accurate", it just has to be substantially believable, without distracting acoustic artifacts that wouldn't happen anywhere in a real performance, and the mind can fill in the details and make it wholly enjoyable. Of course the more experience the mind has with "real", in all its variety, the better the chance that it can take what it gets and "hear" the best of it . . .
This thread needs some humor and a great link.
From (great link): Norman Koren Vacuum tube audio page
[FONT=Arial,Helvetica]Loose screws?[/FONT] [FONT=Arial,Helvetica][SIZE=-1]Sometime in the late 1980's I visited an audiophile in San Diego. When he opened the door he exclaimed, "Norm! I've made the most incredible discovery! A simple modification that really improves the sound— opens it up, sharpens the imaging, and deepens the soundstage. I've loosened the screws on all my chassis. You gotta hear it!" He learned this tweak from a friend of his who had made a small fortune with a "CD enhancer" spray called Finyl. At the time, the ads for Finyl claimed it would increase the number of bits of resolution. [/SIZE][/FONT][FONT=Arial,Helvetica][SIZE=-1]His system sounded terrible.[/SIZE][/FONT]
From (great link): Norman Koren Vacuum tube audio page
[FONT=Arial,Helvetica]Loose screws?[/FONT] [FONT=Arial,Helvetica][SIZE=-1]Sometime in the late 1980's I visited an audiophile in San Diego. When he opened the door he exclaimed, "Norm! I've made the most incredible discovery! A simple modification that really improves the sound— opens it up, sharpens the imaging, and deepens the soundstage. I've loosened the screws on all my chassis. You gotta hear it!" He learned this tweak from a friend of his who had made a small fortune with a "CD enhancer" spray called Finyl. At the time, the ads for Finyl claimed it would increase the number of bits of resolution. [/SIZE][/FONT][FONT=Arial,Helvetica][SIZE=-1]His system sounded terrible.[/SIZE][/FONT]
I have often said that seeking "good sound" is not what one wants from an audio system, its seeking "no aberations" that is the key. The removal of anything that is not right, NOT the addition of something pleasing. They are entirely different things and many people just seem to miss the whole concept.
That's very true Dr. Geddes. IMO I guess. For classical recordings in music halls whose Initial Time delay Gap is over 15 ms, it's very difficult to do in a Living Room.
Dan
Dan
This might be a paradox. But, sometimes (or, maybe in most cases), it'd be a very thin and vague line between "removal" and "addition".
For one who has experienced full scale live concerts, any lackings in frequency range and dynamics in the playback would not be 'believable' enough. In such case, the 'aberration' is the lacking and can't be cured by any 'removal'.
In my own observations, many commercial products cover their shortcomings by 'additions' -- no high, I give you more TING; no low, I give you more BOOM; lacking dynamics, I give you more energy (which I 'store' a lot...).
Of course we may remove all the "illness", but after that, it seems not much left. Then it's time to remove it all. OK, now it works. 🙄
For one who has experienced full scale live concerts, any lackings in frequency range and dynamics in the playback would not be 'believable' enough. In such case, the 'aberration' is the lacking and can't be cured by any 'removal'.
In my own observations, many commercial products cover their shortcomings by 'additions' -- no high, I give you more TING; no low, I give you more BOOM; lacking dynamics, I give you more energy (which I 'store' a lot...).
Of course we may remove all the "illness", but after that, it seems not much left. Then it's time to remove it all. OK, now it works. 🙄
If I was doing high resolution EQ, I'd agree; don't try to bring up cancellations much if any at all, but do hammer down resonances; possibly more than a calibrated mic and pink noise would suggest, due to the psycho-acoustic effects of ringing. Anywhere else I'd say add or subtract anything (bass, treble, reverb, compression, surround sound speakers, etc.) to get it to be the most pleasing. Any reproduction system in a typical listening room is riddled with mechanisms that reduce actual overall fidelity, so is any recording process. I take it from there, and go after what sounds best to me. In the end, it's about enjoyment, nothing else.
Let's stick with your suggestion of having many speakers playing many channels and see if it leads to an imitation of a guitar or flute.
If you learn the reverberation character of a room in order to better hear things in that room, you might say you are "subtracting" the reverberation from the source. Even just talking theoretically, can you still give your ears the same pattern of sound and reverberation as the guitar or flute using lots of speakers? I don't think so.
And when you move your head a few inches, the real flute will still sound like a flute (because your brain is subtracting the room in order to maintain that perceptual "constancy" and to keep your brain from exploding) but your synthetic many-speakers wave-front will not.
When you are listening from down the hall, it is a fabulous mental accomplishment to know you still aren't hearing a real flute:
Maybe that is really the most intriguing question - how come when you are down the hall, you still don't think you are listening to a flute? For sure, your speakers can play the recording flatter and otherwise more perfectly than your ears can tell. You could even equalize for "flat down the hall." The general room sound is the same (although - as I am emphasizing - different in specifics for the speakers and the flute). And all these debates about directivity index are irrelevant when you are down the hall, eh.
I've been using the room for 10 years. I cannot structurally modify it. I've learned the room's lessons and i don't particularly like them. Cut short, the room's not big enough for the sound to be satisfactory - reverb's not long enough. Earl Geddes pointed out at some point the advantage of a larger room is the larger amount of attenuated delayed and decorrellated sounds it produces compared to small room such as mine. (If I've got his point wrong, I'm sure he'll say).
The characteristic sound of the instruments or singers is produced mainly by the front speakers at considerably higher SPLs than sound produced by surround and other auxillary speakers. Either they reproduce the signal given accurately, or they don't. My main speakers do a very good job. They have clean, neutral sound, great dynamic capacity, and well defined directional characteristic. First reflection is not a huge problem. My problem, which is different from the one you pose, is producing a satisfactory synthetic performance space. So I don't agree with your rhetorical answer to your rhetorical question.
If I am replacing my room's acoustic environment with that of a larger performing space, then the speakers providing the new environment will be playing at very low SPLs appropriate to the imaginary distance, be rolled off to imitate HF losses due to attenuation at that distance, and be delayed beyond any time that might allow for psychoacoustic fusion with signal from main speaker. I'm not sure you are making a relevant point.
Don't think so. Stereo is an illusion. This is an attempt to create another one and step in and out of it. We do it reading a novel, listening to a good story teller, going to a play. Any undamaged human can do it.
Actually, I'm listening to the reproduction of the sound of a flute. The more precise that is, and the more realistic its setting, the more satisfactory is the illusion.
Don't want to equalize "for flat down the hall". What I want to hear is the music in the manufactured performance space when listening from down the hall. From POV of the illusion down the hall sounds more realistic because there are even more decorrelated sounds produced.
Room is the same as what? The whole point of the exercise is to modify room's original acoustic with another.
DI's aren't irrelevant in context of the listening room. Down the hall it's a matter of relative SPLs from main speakers and envrionmental ones. Right?
I am at fault for supposing that something which seemed very obvious to me was not obvious to you.
A source, say, an instrument like a flute or guitar, emits sound in its unique and characteristic and historically-evolved 3D (or 4D) pattern. What reaches your ears is that pattern plus the room acoustics. A speaker emits sound in a very different pattern... and why I earlier mused that perhaps a mono single-driver OB or ESL-dipole might better simulate an instrument, or just as a woody, multi-resonant 55-year old Karlson-15 can sound more like a cello than my trick Dayton-Wright ESLs.
Sure, the synthesized wave-form of a rosin-stroked string can be quite perfect. That is the dimension you would call "a violin". But that is only one-dimensional. Anybody ever mistake a synthesized instrument for a real one, even when Mr. Moog's wave-form is perfect? Like with wax-cylinder testing, no question it can't be mistaken for anything else - but that's a pretty easy criterion to achieve.
A recording played on speakers of an orchestra playing in Carnegie Hall is equivalent to listening to the sound "through an open window and down the hall", not being in Carnegie Hall. If you want "realistic" sound, you need headphones and proper mic'ing, at least as start toward realism.
Last edited:
You seem pretty confident that you're right, so if you have more examples, evidence, or references to back up your assertion that turning your head 30 degrees away from a stereo speaker pair universally does not result in an apparent change in tonal balance of the treble, please do so.
I've tried your test in a "normal room" with speakers of "normal" treble directivity for years, it's one of the standard things I check in a new speaker/room set-up, so I already know what the result is - for those typical situations you are correct, there is generally little if any perceived change in treble balance, particularly in smaller rooms.
However you then make the extrapolative leap that this must always be the case, and that it is therefore due solely to the ability of the brain to "tune out" or ignore perceptual changes in frequency response that are caused by the HRTF.
This is where we disagree. My argument is that a lot of the reason why you can turn your head around without noticing much change in treble balance in this "normal" room and speaker configuration, is due to the pervasive ambient treble field in the room which is coming at you from many directions - no matter which way you turn your head there is a roughly uniform treble field, particularly in a smaller shoebox shaped room.
The direct to reflected ratio may change somewhat for each ear as you turn your head, but provided the speakers have a "good" power response in the treble, (and what represents good is open for discussion, as seen elsewhere in this thread) the average treble level won't vary a lot, and can thus be "tuned out" by the brain - particularly when it is consciously aware of the fact that the head is turning and that some HRTF related changes in response are to be expected..
You very confidently offered a personal, sighted experiment of your own as indisputable fact that Dave must be wrong. Your exact words were:
"Dave, I suspect this is mostly misdirected theorising, because if it were true our perception of the sound tonality coming from a speaker would significantly change just by rotating one's head 30 degrees while listening. It doesn't. Even a 5 second test sitting in front of one's hifi shows there is nothing significant going on."
While agreeing that's what generally happens I offered a counter example which when the test circumstances are changed somewhat (vastly lower treble power response, but still flat on axis, in the same room) this no longer holds true.
I asked if you have tried a similar test but you haven't responded to this point, nor have you responded to the suggestion of trying it outdoors. So I ask again, have you done this test with a driver with extended but very directional treble in a small/moderate reverberant room, and/or have you listened to your speakers set up outdoors and done the same test. Have you even ever set up your speakers outdoors before and attempted a careful evaluation of stereo imaging or tonality ?
As to your point about this being a "sighted" experiment, I can only assume from the tone of the comment that you are implying it should be some kind of blind or double blind test, and that if it's not the result is invalid.
Complete nonsense. If you stop to think about it you see that not only is it impossible to construct a blind or double blind test for the example I set forth, (comparing the two speakers with vastly different treble power response to see how it effects head rotation treble effects) it doesn't make any sense anyway.
The whole point of blind tests is to screen out listener preference or anticipation effects when trying to detect near-threshold changes between device A and device B. The point is to try to prove a statistical "no change" where a listener may have erroneously favoured one device over the other had they known which was active at any given time.
It's not a tool that can be used when the differences are vast and it's obvious just by the character of the sound which device / set-up is currently in operation.
Yes, I could set up a test where I got a computer to switch between the same speaker (full range driver crossed with ribbon tweeter) and the full range driver by itself without physically moving any thing, complete with applying two different EQ curves so that in both cases the on-axis response was as flat as possible.
So what ? Within seconds of listening to any familiar piece of music I would immediately know which of the two configurations was in operation due both to characteristic differences in the treble between the two drivers, and the vast change in treble power response between the two configurations. Trying to do this test blind is a complete waste of time.
Or, were you suggesting I should turn my head with my eyes closed ? 😀 Unfortunately my brain still knows which way my head is facing whether my eyes are closed or not... 😉
So if something isn't observed by PhD's in an expensive lab it's not valid ? Right..... 😉 I guess the majority of people on this forum better give up and find a new hobby then and leave everything to the big boys ?
How exactly there can be a "physical reality" to a phantom stereo image I can't see, the thought of it is quite laughable when a phantom stereo image is formed in your mind from two physically disparate speakers.
So it's impossible to shelve down the treble slightly in a mix until it sounds more balanced ? Sounds reaaaaaaally difficult to me 😉
True enough, but if you have your speakers at a vastly different angle to 30 degrees (+/- 10 degrees) you're not going to get a decent stereo imaging experience anyway, so a slight variation in the treble is the least of your worries. Anyone who cares about imaging won't stray too far from this range.
Not true. You're mixing up two different issues here. The first is Dave's original suggestion that flat on-axis is not correct for a stereo speaker placement. This is independent of placement of the speakers in the room, (other than the angle from the centre line of the listener to each speaker) and independent of the wall surfaces and room treatment.
Dave's issue relates only to the direct sound from the speakers not the reverberant field because the reverberant field is not coming from the same 30 degree angle. Regardless of what else is going on in the room if the speaker is at 30 degrees to your ear the apparent treble of the direct sound will be elevated by the HRTF. Whether that will be perceived as a change in tonality or "tuned out" is uncertain.
The second issue which I first brought to the discussion was the idea that the much increased treble sensitivity of the HRTF at much wider angles of ~60 - 90 degrees may be one reason why a flat reverberant treble field usually sounds too bright. Although both are HRTF related, they are two entirely different discussions, so I'm sorry to see them getting mixed in together all the time.
Really ? So if the on axis response is all that matters why even bother with a wide dispersion at treble frequencies if you're going to listen from the sweet spot ? It shouldn't matter, by your reasoning, right ?
Clearly you've never listened to a large full range driver in any somewhat live room or you would know just how much the power response of the treble does affect the perceived treble balance. The same driver can sound well balanced in the treble outdoors or in a very dead room, and sound almost completely dull and devoid of treble in a live room with the same EQ setting.
As much as 10dB of variable slope boost in the treble can be needed to achieve a satisfactory balance in a live room, but now the treble is far too strong in a dead environment. (And it still doesn't sound "right" due to lack of room reverberation in the treble) Clearly it matters, and it's one of the biggest drawbacks of larger full range drivers - they just can't cope with changes in listening space acoustics without sometimes dramatic EQ changes.
Besides, I didn't say that reflected sound always dominates direct sound, obviously it doesn't in large well damped rooms, but in smaller fairly live rooms it often can.
Nor was I talking about the entire frequency range, we are specifically discussing the treble here, and the HRTF provides a large boost in treble sensitivity as we go out towards 90 degrees, so that the rooms power response in the treble may have an unduly large effect on the perceived tonal balance in the treble, as the side wall reflections may begin to dominate in the treble even if they don't in the lower frequencies. That was the thrust of my argument.
What you seem to have missed is the fact that I'm not here to recite text books on "proven" facts in audio, threads like this are a discussion of some of the less well defined and agreed upon issues in audio, and although many people would like to believe things like "flat is best" and leave it at that, the reality is vastly more complex.
There are some aspects of this issue which have not been adequately explained by the "experts", key among them being why a wide dispersion and flat power response through the treble so often sounds too bright, and why such a speaker sounds balanced in a dead room but can become too bright in a reverberant room even though it's power response is generally still falling a bit.
Even people like Linkwitz seem to be struggling to explain the cause of this issue, hence the modifications to the Orions which prompted this entire thread, and his attempt to rationalize the modifications based on "flat on-axis is not right".
Just because in the course of discussion and debate I put forward a proposition or theory on why some observed phenomenon might occur does not automatically mean I believe it to be the truth and a concrete proven fact - I have plenty of opinions, hunches, and intuitions in audio which aren't based on concrete facts, and I offer them into the discussion as thoughts, ideas, and alternative viewpoints. I don't try to state them as facts.
With more discussion some of those ideas get discarded, others may be reinforced, and in the process the conversation is moved forward in unexpected directions so that people, including me can sometimes learn new things from it.
I also sometimes play devils advocate and put forth ideas that I think are interesting and feasible but which I don't actually believe to be the case specifically to stimulate discussion and thinking outside the box by all the participants.
If all we ever do is follow existing doctrine, what forward progress and new insights can be found ? Contrary to the beliefs of many people, scientists and researchers included, many of the great, fundamental discoveries in science were made by accident when looking for something else, and the theories of "how" were worked out after the fact. If those scientists were adhering strictly to what was already known and "proven" whole new fields would never have been discovered.
Some people like to stick with what the text books and researchers say and that's fine, but I like to try and explore beyond that, particularly in a discussion as interesting as the one in this thread 🙂
Last edited:
Ben and Frank.
Sorry you've never heard it done right. It isn't easy, or 100%, but it is possible. And in stereo, no less!
Sorry you've never heard it done right. It isn't easy, or 100%, but it is possible. And in stereo, no less!
Ummm, "complete nonsense" - that's quite a strong rejection of blind testing and that is quite an exceptional point of view. I predict that when you die, you will be condemned in hell to repeat Psych 101 over, and over, and over....
My experience is that the engineers who resolutely think their minds are the most objective sensors are, in fact, the people most susceptible to false illusions.
Save the eternal torment. Read a book about perception and repent before it is too late!!!!
(I hope your religious sensibilities or beliefs in an after-life are not offended by my little effort at joking.)
Last edited:
Hi Frank
Yes, you have that correct. In a large room there is a large time delay from the direct sound to the early reflections and this gives the ear the chance to make sense of things like direction and tembre without the room interfering. Then when reverberation does set in it is much more uncorrelated to the direct sound than it is in a small room such that this sound is heard as a pleasant "ambiance" effect that does not confound the initial judgements of the direct field.
To the rest of the arguments I would say this: Stereo is not perfect, we all know that. That it can be used by competent artists to convey a masterpiece is also without contention. Everything is a composite of these two facts. Allowing that (seemingly rare) artistic team to reach me in my living room in the way that they intended, to the greatest extent possible, is the only true goal IMO.
Last edited:
You would do well to avoid quoting me out of context. 😉
If you read the whole text of what I wrote instead of just the bit you've quoted you'll see that I'm rejecting blind testing as unsuitable for this specific test scenario, because there is no way for it to be truly blind.
There is no way to hide which of the two configurations are active because they are so vastly different in sound character that it is immediately recognisable which is active.
Blind testing in this case does not remove any bias, because I know which configuration is active whether I switch it myself or let a computer do it in secret, therefore it does not provide any additional benefit or accuracy.
Blind testing is only useful when comparing two very similar sounds that have no or very slight differences, to statistically screen out personal preference expectation from sighted testing.
I don't have anything against ABX type testing for certain types of tests where it's appropriate (for example comparing performance of amplifiers, audio codecs etc) but this is an inappropriate use of blind testing protocols - I suggest you read up more on what they are and aren't useful for.
- Status
- Not open for further replies.