The Objectives of a Loudspeaker in a Small Room

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
In the "Beyond the Ariel" thread, we got a bit sidetracked on a fundamental issue which, I think, deserves its own discussion- what are the objectives for design and engineering in the first place?

As much as people like to discuss the "sound" of capacitors or the tradeoffs of moving mass and high Xmax or dynamic range, the elephant in the room is stereo itself. We start with a HUGE approximation- we sample sound pressure versus time at two points (or at more points, then condense the sampling to two points), then we stick a couple of boxes or panels into a room of questionable reverberation totally unlike the original acoustic space. Out of all this, we expect to have our ear/brain sense something that reminds it of live music in a different space.

So notions of "accuracy" and "fidelity" come prepackaged with enormous constraint, and everything we do from that point is just manipulation. As engineering apes, humans are very good at manipulation, but here the target is much less clear. We get to a fundamental question about which there is very little agreement- what should a speaker in a small room do (in an engineering sense) to keep up its end of the illusion?

In a Platonic sense, one could say, "The loudspeaker should replicate the original soundfield at two points in the listening room space, the listener's earholes." A lovely notion, but impossible under normal circumstances. Even more so when one incorporates the notion of localization by small head movements.

The excellent research at Harman indicates that the illusion is best created with a flat on-axis response, and a polar pattern that rolls of smoothly with increasing angle, and this is how I design point-source speakers. Yet, how does one explain the enduring preferences for dipoles and omnis?

In any case, for me, this fundamental question of goals is an interesting one, so I'd like to encourage the "experts" to discuss the pros and cons of different approaches, hopefully with some actual data and theory that goes beyond hand-waving and bald assertion.
 
A good start, but after I commented on the objectives in a small room, it dawned on me that we should agree on the objectives of a sound system first. In this regard I have often found a perplexity set of differeing deffinitions. No wonder its so hard to agree on a design approach when we can't even agree on the end result.

I wrote a white paper on this topic some months ago when a good way of expressing my objective came to mind. I have attached this for review and comments.
 

Attachments

  • aquariums.pdf
    15.8 KB · Views: 568
Interesting analogy Earl, but IMHO, the glass is the playback source-amplification-speakers-room, the speaker-room dominating of course. What lies beyond the glass is the recording, which in many cases would look like some seriously mutated fish in muddy water.

SY, I would like to see Harman style panels, trained listening to live, unamplified music, to do DBT's with Toole's, Earl's, Danley's Andrew Jones, etc designs vs Linkwitz, Kreskovsky and other seriously engineered dipole designs in typical sized living rooms.
IIRC, Duke Lejune (sp) did some informal tests where Summa prototypes were a statistical dead heat with Gradient Revolutions, a cardioid/dipole hybrid - with less that SOTA drivers - as compared to the BC's.
I'd like to see far more tests similar to those with rigorous scientific controls.

cheers,

AJ
 
I don't buy the analogy, though it's consonant with the reigning paradigm in high-end audio ("If I just reduce the distortion/increase slew rate/get more bits of resolution... the illusion will be significantly closer to perfect."). What you fundamentally have not done with the aquarium analogy is address the problem that the complex 3D soundfield has been collapsed to two (or if we're generous, five) points, all of which interact with a small, highly reverberant box surrounding the transducers and the listener.

And additionally, the peculiar differences between sound and light waves (e.g., you can hear but not see around a corner) also suggest to me that visual analogies are misleading. Try to recreate the appearance of the aquarium against a neutral background in another room with mirrored walls using two light sources with fixed dispersion.
 
Actually the Summa test was done here at my home, with Duke's help. Yes, the Summas and the Gradients were statistically the same. The Summas were rated better in the low end and actually, the ported JBL's had the highest rated low end, but a dismal high end rating from the diffraction horn. Two different driver sets were used in the Summas, TADs and B&Cs. They were not statistically different. There were about 25 subjects from the audio club, the sources were blind.

I should mention that the Summas were about 12 dB more efficient than the Gradients. The Gradients would have about 20 dB less headroom than the Summas owing to the smaller, less efficient drivers. It's much easier to design and build high quality low output loudspeakers than high quality high output ones.

As to the analogy, all analogies, by the simple fact that they are analogies, fail in some regard. To me, and I said that this is the way that "I" look at the problem, the speakers are windows on the source and just like looking through two lenses, ala binoculars, we can get a sense of 3D from only two sources, albeit not perfect.

Multichannel (and I prefer three to two, I use three in my own system) is attractive, but not yet there as far as source material goes. And are you going to buy or build 5 or more real high quality loudspeakers? I would opt for fewer higher quality loudspeakers than many lower quality ones as being the better choice.


And yes the electronics are part of the "glass".
 
gedlee said:
Actually the Summa test was done here at my home, with Duke's help. Yes, the Summas and the Gradients were statistically the same. The Summas were rated better in the low end and actually, the ported JBL's had the highest rated low end, but a dismal high end rating from the diffraction horn.
Two different driver sets were used in the Summas, TADs and
B&Cs. They were not statistically different. There were about 25 subjects from the audio club, the sources were blind.

I don't recall all the details of the test. Is there a link to the results somewhere? Was the music recording selection primarily acoustic or amplified?


I should mention that the Summas were about 12 dB more efficient than the Gradients. The Gradients would have about 20 dB less headroom than the Summas owing to the smaller, less efficient drivers. It's much easier to design and build high quality low output loudspeakers than high quality high output ones.

Yes, that was the interesting part. 2 very different platforms, yet similar results in the same room. Of course, controlled directivity was one (major?) common factor. The Gradient perhaps even more so, with its coincident mid/tweeters vertical control and driver summation.
The fact that the drive units of the Summa were much more powerful and SOTA was not lost on me. I see no reason why the Gradient platform could not use more powerful drivers, like the B&C coax that Danley uses in the SH100 and larger, higher displacement woofers.

Perhaps Duke could be convinced to repeat this test with the same participants, in one or two other rooms, with say the top 3 contenders from the old test with a controlled directivity dynamic dipole like SL's Orion++ system thrown in. It wouldn't be definitive, but perhaps a trend might emerge.

cheers,

AJ
 
AJinFLA said:


I don't recall all the details of the test. Is there a link to the results somewhere? Was the music recording selection primarily acoustic or amplified?


AJ

Unfortunately the results were taken off of my website when I closed down the Summa pages.

The musical selections were quite varied and came from the DLC LIT listening disk.

These tests are very difficult to impliment, which is why so few are done. Our intention at Ai is to do more of them. We have a listening room where we can do blind auditions. The problem at Ai is the lack of competent listeners. The Thai's just don't appreciate good sound. Its not a priority with them. We hope to change that.
 
diyAudio Member
Joined 2004
Well its obvious that what were asking to be done is impossible but I guess everyone realises that already. A good approximation is the best we can hope for.

With regards to recording these large venues using a limited number of aspects/perspectives to capture the sound. I'd say we need a quantum leap forward in sound capture technology to claim to have accurately bottled up the original. There are processing techniques which can extrapolate or rather interpolate the possible data that occured which the mics didn't capture because of their defined positions. But these stereo and ambience expanding techniques are crude at best and should be avoided altogether IMO.

If we want to reach a specific goal within a close set of tolerances ie. original performance is closely followed by the reproduction then I believe only way to succeed is to remove the room equation out of the capture process. You'd do this by creating all music from soundbanks or performances done in anechoic/studio conditions. This greatly limits the number of factors to control and accurately reproduce since the original performance is synthetic just like re-production in your home and your listening room acoustics will share more in common with a mixing studio than a concert hall etc. This idea will likely be held with contempt amongst many music lovers but it does allow the recording and capture process to be a more scientific and repeatable process.

I think a very important part of the equation is removing, as best you can, your own listening rooms acoustic signature. Luckily there's a lot of ways to achieve this but unfortunately each room is a unique case so no universal standard, which anyone and every can deploy, is available.

Digital room compensation, loudspeaker polar response, physical room treatments and placement are the key factors for me. We have no real control of the recording process but these we do have a say over.

All we can do is make a bad situation better.
 
I tend to agree with you about minimizing the acoustics of the recording space, but not of the listening space. I try to maximize the listening space acoustics to yield ambience and spatiousness to the playback. It takes some careful design to do this right, but I find that dead rooms just sound lifeless. I have heard speakers in an anechoic chamber and boy are they lifeless.

If there is a lot of acoustic space in the recording and a lot in the playback room then the net result is confused as one is large and the other small.

If the recording has "acoustic" but not the room then the playback is flat without and feeling of space.

If the recording has minimum acoustic of its own but the playback room has good acoustics (and good speakers) then I find this to be the best situation for creating the feeling that I look for of the music being in the room with me. Others look for the sound system to take them to the recording space, I find headphones best for this, but in the long run I don't find the overall effect that appealing.
 
what a thread! what a question!
interesting enough for me to log in (at last) and try (at least ;) ) to write something in english ;)

IMHO there are three fundamental issues in audio
two of them are electroacoustical (the third concerns retrieval of the signal from the recording medium)

the first electroacustical problem concerns the microphone in a recording space
the second concerns the loudspeaker in a listening space

much is debatable in both, much is not yet understood and much seems to be overlooked

let me draw Your attention to some interesting ideas on the subject
these are so far practically overlooked on this forum, judging from "forum search" results:
http://www.moultonlabs.com/dave_more/nick_batzdorf_interview
http://www.carlssonplanet.com/downloads/uploads/carlsson_ortho.pdf
http://www.linkwitzlab.com/Acoustics and Psychoacoustics of Loudspeakers in Small Rooms.doc
http://www.stereolith.ch/

contrary, perhaps, to the first impression, these concepts have much in common
one should read "between the lines"
 
thanks for the link AJinFLA

very insightfull comments and quite solid theoretical background

I have one important reservation though

the author writes: "by preserving the sidedness of early reflections, the ear does not receive cross room reflections that would otherwise cause perception of the loudspeaker as the sound source"

"perception of the loudspeaker as the sound source" can be a major problem indeed
but not always - much depends on how the loudspeaker is positioned in the room
take corner horns or positiong recommended by Harold Beveridge for his electrostats as an counterexamples (see "the Beveridges": http://pic20.picturetrail.com/VOL22/541589/834883/62208841.jpg )

but the main cause of the problem is an interaural crosstalk and not assymetric reflections as suggested above

the truth is that the loudspeakers are localized as separate sound sources beacuse of interaural crosstalk i.e. because both ears are receiving a sound wave from both loudspeakers when they are positioned in a "stereo triangle"

there is nothing wrong with that "per se"
after all Moulton rightly points that "loudspeakers themselves are perceived in stereo as early reflections of a sound whose direct version we missed"
that's perfectly OK, this is how "stereo" works

BUT the problem arises when perceiving of the loudspeakers as "early reflections of a sound whose direct version we missed" conflicts with psychoacoustics of the listening room, i.e. when the "listening room data", gathered by the brain, are suggesting that they (loudspeakers) shouldn't be THE FIRST early reflections and yet they are so perceived - as the FIRST AND DOMINATING early reflections - because the so called "direct sound" is coming from them

this situation - which is nearly inevitable in a typical "stereo triangle" of typically positioned (i.e. well into the room and away from the walls) loudspeakers in a typical listening room - results in serious limitations in "spaciousness, envelopment, image depth, and image solidity" factors

it is a major problem from the point of view of "achieving realism of sound reproduction" (a "recognition factor")
 
woa at last....
this is the exact subject i have been trying to bring up for a few months ..but nobody has ever really followed
( i will blame it on my very poor english syntax/grammar
:p )


What you are bringing up here now Thomas,
is the main reason why i have taken a big step back in all my audio quest/research ..back to the original goal..

i think that alot of us ( and probably all others :p )
are currently "lost" in our search for the ultimate compromise for reproduction of our sources.

I firmly believe that there is only 1 possible perfect solution (well either perfect or featuring the best compromises ) , but unfortunatly there is a large ( but probable near definite ) number of different source attributes wich make this quest's potential very broad.

So we can only hope to get as close as possible to the solution, within our sources limits.

I have yet to understand the recording process
( my knowledge is quite limited in this area )
wich could help me quite a bit in understanding how the source was intended to be played, so i wil relay on you gurus to point out the major components of the sound recording stage.

So what now to do with the miriad of possibilities...
 
Re: Re: Multichannel

PeteMcK said:
Re: "Multichannel (and I prefer three to two, I use three in my own system) is attractive",

as a matter of interest, what signal goes to the 3rd (centre?) speaker, L+R?

(I've though about trying this for some time).

Cheers,
Pete McK

Yes and no. When there is no center channel, as there is on movie tracks, etc. then it is derived from L+R - 3dB.
 
you can probably tell I'm not au fait with HT stuff, but can someone explain how to generate L+R-3db in a stereo situation?? I'm assuming Ht RECEIVERS OR WHATEVER have some sort of summing function, but how would we do it with stereo amps.

this is becoming interesting:)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.