If you try to reproduce brass orchestra on realistic loudness, it may happen, otherwise why?
(Maybe this does not answer your question, but there are some studies i know)
In english language, this is a thesis i found recently:
"MODAL COUPLING OF DIRECTIONAL SUBWOOFERS IN RECTANGULAR ROOMS"
A Thesis in
Acoustics
by
Philip Feurtado
https://etda.libraries.psu.edu/paper/18841/17442
The thesis is based on simulations of low frequency sources in
small rooms.
In short below schroeder frequency:
When compared to a monopole and a dipole source, a cardioid tends
to have less variance in frequency response due to a change in
position of the source.
That is because a given set of modes can be excited in pressure and
velocity maxima as well, which is not the case with monopole and dipole
sources.
Nevertheless, the orientation of a cardioid source has to be aligned in
the right way, to be able to excite a mode in a velocity maximum.
see e.g. pages 40ff, 56ff (Conclusions)
Regarding measurements there should be at least one paper
by Lampos Ferekidis:
Authors: Ferekidis, Lampos; Kempe, Uwe
Affiliation: wvier, Germany
AES Convention:116 (May 2004) Paper Number:6110
Publication Date:May 1, 2004
Subject:Loudspeakers
"The Beneficial Coupling of Cardioid Low Frequency Sources to the Acoustics of Small Rooms"
___________________
There has been a discussion here lately involving standpoints by
Earl Geddes and John Keskovsky, which made me thinking about
a cardioid's and a dipole's properites due to (smoothness of) room
mode excitation partly relying on the fact that they are simply
distributed sources (involving a dipole separation):
Also multiple monopoles (driven in phase electrically) can be used
to lower variation in frequency response in different listening positions
or to make the response more smooth in one dedicated position.
Typically the separation of sources has to be aligned with the range
of wavelengths that is subject to "modal balancing".
Last edited:
Yes, but did not David Griesinger tell us that if the
in phase component of the stereo signal and the out of phase component
is used to drive a different set of modes this will enhance envelopment ?
http://www.davidgriesinger.com/asa05.pdf
I mean: If there is a difference in L-R signals, is it really the best
to throw that away ?
In order to pure smoothing of room mode excitation using a given
number of subwoofers (2..3) this seems the best solution, sure ...
But no "spatial" information in stereo signals below typical living
room schroeder frequencies ?
Last edited:
Yes, I think it's the best solution in acoustically small rooms because
1) envelopment is a property of late reverberation (>80ms) and there is no late reverberation in acoustically small rooms
2) spatially distributed level differences caused by room modes translate to interaural phase differences which can deliver spatial cues that might conflict with those delivered by the recording
Last edited:
But there is late reverberation on some recordings, if you think
of organ recordings in churches e.g.
.... which would call for not exciting lateral modes at all ?
But we need all modes anyway for "modal smoothing" ... can we
afford to leave the lateral modes out in a small room ?
When listening near the center between the side walls, would you try
to supress especially excitation of the lateral uneven 2,4,6 ... x lambda/2
modes ?
Last edited:
Hello Anatoliy
Well it depends on what you listen too. I like progressive rock so a live Genisis show can have quite a bit of energy in the first octave from the bass pedals as an example. If I drop on a DVD of a live show it can have substantial bass peaks in that range. The system can hit 120db from 25-80hz between the 2 stereo subs and the paired LFE subs. I have 4 15's handling that range, 2 JBL 2235 for the stereo L/R subs and 2 JBL 2266's for the LFE.
Also for movies where you need to be able to hit at least 105db figure +10db in the LFE for reference level. Typically I rarely listen that loud. Usually about -10db to -6db from reference depending on the movie but I can if I want.
Rob🙂
Exactly that is what we're trying to reproduce. If we add some kind of "envelopment" utilizing the listening room then we add something that is NOT part of the recording. It's like utilizing strong early lateral reflections for added spaciousness. It might sound more realistic or pleasing but it's not part of the recorded content.
I think the modal behavior of a room is too complex. We can't just simply "leave out" certain modes.
Because many rooms are not shoeboxes, but if so
e.g. 2 in phase monopole woofers one at each sidewall,
cannot excite uneven 1,3,5 ... x lambda/2 lateral modes:
Those should produce the highest lateral pressure gradient
in the mid of the room.
Same for a woofer at a midpoint of the front and/or rear wall:
Difficult to excite the uneven lateral ones. One could even
minimize for lateral gradients due to the listening seat
for the higher order modes, if lateral gradients were
identified to be "potentially detrimental".
Using dipoles aligned with the front-rear axis, you can avoid
exiting lateral modes at all, in some rooms this is (with proper
placement) sufficient to get a balanced frequency response at
the listening seat (not so in a shoebox room usually).
Below Schroeder we know the modes by name in a small
living room so to speak ...
e.g. 2 in phase monopole woofers one at each sidewall,
cannot excite uneven 1,3,5 ... x lambda/2 lateral modes:
Those should produce the highest lateral pressure gradient
in the mid of the room.
Same for a woofer at a midpoint of the front and/or rear wall:
Difficult to excite the uneven lateral ones. One could even
minimize for lateral gradients due to the listening seat
for the higher order modes, if lateral gradients were
identified to be "potentially detrimental".
Using dipoles aligned with the front-rear axis, you can avoid
exiting lateral modes at all, in some rooms this is (with proper
placement) sufficient to get a balanced frequency response at
the listening seat (not so in a shoebox room usually).
Below Schroeder we know the modes by name in a small
living room so to speak ...
Last edited:
Griesinger is talking about *improving* the *perception* of recorded content that specifically includes acoustic envelopment at lower freq.s . Expressly content that has random in-phase fluctuation.
To better reproduce random in-phase fluctuation found on a *signal, he proposes driving the room in a specific (if likely domestically unrealistic) manner relative to the listener's "forward direction", as well as acoustically treating the room (and/or utilizing some very targeted equalization).
This is NOT *adding* a condition to *create* envelopment. In this respect it's not at all like "utilizing strong early lateral reflections for added spaciousness".
The only time he "dips" into *creating* envelopment is with respect to concert hall's that lack it by utilizing their LARES system - which is a sound reinforcement system (..not a stereo reproduction system). (..page 15.)
Ironically, the near-field dipole bass setup you've got will do a better job of reproducing random in-phase fluctuation (..at least for that one listener), than what he proposes. 😉
*Note that he also mentions that most signals (of at least acoustic recordings - specifically concert hall recordings) use recording techniques that do capture random in-phase fluctuations. It's also quite likely that some processing for non-acoustic recordings contain this as well (at least to some extent).
Just like a Bose 901 "improves" spaciousness. It's adding something that is not in the recording in order to make it sound "good". The added spaciousness is a property of the listening room and not a property of the recorded space. It's added to each and every recording regardless if appropriate or not.
There are better ways to make stereo sound more spacious and enveloping based on content that is actually in the recording: Neo:6 or PLII. That's what I use.
How so?
Last edited:
Perhaps not, but there are also "properties of the recorded space" that are not captured by the microphones, not least being the level and direction of reflected sound that the listener might hear in the hall. Replacing that missing information with reflections in the listening room is not as artificial as you make it out to be. And in any case unless you choose to live in an anechoic chamber those reflections are going to be there.
Use them.
How is artificial reverb any better or more "natural" than listening room reflections (which are based entirely on "content that is actually in the recording")? Especially since at least some of those reflections are going to be there no matter what you do?
It's a fool's errand to build a anechoic listening room and then be forced to add synthetic reflections to make it sound natural and real again . . .
Discussing "sounds good to me" is still nothing I'm interested in.
Please read up how Neo:6 and PLII work. They don't add artificial reverb. Here's a third alternative: http://decoy.iki.fi/dsound/ambisoni...for Multispeaker Stereo (TRIFIELD)_Gerzon.pdf
Can you reference a good discussion of Neo:6?
At any rate, from what you are saying all rooms are bad since they all alter the original recording. Only an anechoic chamber would not. Isn't this a rather unrealistic position.
At any rate, from what you are saying all rooms are bad since they all alter the original recording. Only an anechoic chamber would not. Isn't this a rather unrealistic position.
I didn't invent stereo 🙂 Basically, yes, stereo should work best under anechoic conditions. But, then there's the perceptual side which isn't well understood because it's not well explored.
An example: The less reflexions the more the issues with phantom center images become audible. Reflexions fill in response notches generated by stereo's interference field. So early reflections help mitigate that problem. At the same time early reflexion can have detrimental effects. Simple solution is to add a center speaker like the guys at Bell labs proposed some 80 years ago...
Of course one could argue that the room has to be part of the stereo system but where are the meaningful standards that would make such a proposition work?
Last edited:
As far as i understood matters, we were currently not disussing
"Reflections" above the Schroeder frequency of the listening
room (and inside the listening room).
But how to deal with L-R differences due to phase (and amplitude)
in a stereo signal below the Schroeder frequency of the listening room.
However, interaural pressure gradients at low frequencies
- in a modal behaving room below Schroeder -
might translate as phase information, because that is what spatial
hearing is based on at LF (<800Hz ILD is not a reliable cue for localization).
This was Markus' point in suggesting computation of a mono signal
below Schroeder and then only care about modal smoothing ... e.g.
by using multiple subwoofers and the mains.
Nevertheless AFAIK even unter "natural" conditions, there is no
localization possible, for pure sine tones approx <80Hz.
For frequencies that low, the point would be "spaciousness" not
localization IMO.
Are stereo L-R phase/amplitude fluctuations below Schroeder really
just "prone to misinformation" and have no value in getting
an impression from a recording venue ?
When using pure intensity stereophony that might be the case.
I listened for a long time using 2 dipole subs in the left and right
bottom edge of the room, and was able to run them in mono or
in stereo (simple summation S=L+R).
With some organ recordings of larger venues, i preferred the stereo
setup, not because any differences in "localization" or "tonality",
but the spaciousness of the larger venue was reproduced in a
far more realistic way.
Maybe a "sounds better to me" kind of argument and
also maybe relevant for only a small fraction of recordings
... surely.
But i do not feel, L-R stereo signal differences are just "waste" at
frequencies below Schroeder, because they cannot contribute to
localization ( ... which i do not doubt for a second).
They can make the impression of a recornding venue more realistic
(in spaciousness and depth): There is something transported through
the recording which simply cannot be originated in the listening room,
"it's a kind of magic" if you like.
The question to me is: If you have e.g. 2 Subwoofers, is it better
to use them for (1) "modal smoothing" solely (and feed them with the
same signal) or do we want to afford (2) stereo below Schroeder.
Rationality says (1), when listening to those recordings mentioned
my heart says (2).
"Reflections" above the Schroeder frequency of the listening
room (and inside the listening room).
But how to deal with L-R differences due to phase (and amplitude)
in a stereo signal below the Schroeder frequency of the listening room.
However, interaural pressure gradients at low frequencies
- in a modal behaving room below Schroeder -
might translate as phase information, because that is what spatial
hearing is based on at LF (<800Hz ILD is not a reliable cue for localization).
This was Markus' point in suggesting computation of a mono signal
below Schroeder and then only care about modal smoothing ... e.g.
by using multiple subwoofers and the mains.
Nevertheless AFAIK even unter "natural" conditions, there is no
localization possible, for pure sine tones approx <80Hz.
For frequencies that low, the point would be "spaciousness" not
localization IMO.
Are stereo L-R phase/amplitude fluctuations below Schroeder really
just "prone to misinformation" and have no value in getting
an impression from a recording venue ?
When using pure intensity stereophony that might be the case.
I listened for a long time using 2 dipole subs in the left and right
bottom edge of the room, and was able to run them in mono or
in stereo (simple summation S=L+R).
With some organ recordings of larger venues, i preferred the stereo
setup, not because any differences in "localization" or "tonality",
but the spaciousness of the larger venue was reproduced in a
far more realistic way.
Maybe a "sounds better to me" kind of argument and
also maybe relevant for only a small fraction of recordings
... surely.
But i do not feel, L-R stereo signal differences are just "waste" at
frequencies below Schroeder, because they cannot contribute to
localization ( ... which i do not doubt for a second).
They can make the impression of a recornding venue more realistic
(in spaciousness and depth): There is something transported through
the recording which simply cannot be originated in the listening room,
"it's a kind of magic" if you like.
The question to me is: If you have e.g. 2 Subwoofers, is it better
to use them for (1) "modal smoothing" solely (and feed them with the
same signal) or do we want to afford (2) stereo below Schroeder.
Rationality says (1), when listening to those recordings mentioned
my heart says (2).
Last edited:
But it doesnt, there have been perceptual effects observed like inhead localization, similar to reproduction via headphones.
But anyway, i am more interested in ideas for the range below Schroeder
...
Who observed those effects and have they been documented? I only know the anecdote Toole presents in his book about an ambisonic listening test done in an anechoic chamber.
Tom Danlay always encourages people to drag their speakers outside and listen in order to experience what stereo can do.
Stereo works best under whatever condition the mastering/mixing was done. If that was done in an anechoic chamber then maybe. However if it was done in a control room with a degree of normal room reflections (although typically less than an untreated living room) then that is how it will sound "best" or at least closest to the engineers artistic intention.
If you let go of trying to achieve absolute "reality" (what you hear in your room sounds exactly like the live instruments in a real venue recorded and played back raw) which is not possible with two channels, to trying to achieve hearing what the engineer heard in the control room ("artistic intent") then your goal becomes a lot more achievable.
An analogy - if you display a violet flower on a TV screen should we be disappointed that the colours don't look like the real violet flower held next to the TV (since the violet colour is outside of the colour space of the TV and CAN'T be reproduced faithfully) or should we be happy that our TV is calibrated as closely as possible to the colour specs of the TV standard and thus most accurately duplicates the artistic intention of the director when the colour grading was performed in post production ? We see what they intended without further embellishment ?
If you're referring to the "stereo dip" at around 2Khz then I don't see how a side wall reflection fills this in. The delay to the first side wall reflection is well beyond the fusing time of the original signal so will not alter our perception of tonal balance significantly - 2Khz is high enough that we are perceiving the tonal balance based on the first arrival not steady state response.
The only time it would alter our perception of balance ("fill in" the hole) is if you're well past the critical distance into the reverberant field - but you won't be getting a good phantom image well back past the critical distance anyway so that's a moot point. Why design for such a non-optimal listening position.
The first stereo dip at around 2Khz can be compensated for with a small bit of EQ - in well-imaging speakers that get their final "voicing" by small tweaks based on listening (rather than pure design by measurement) you could argue that the designer is inadvertently (or knowingly) compensating for the stereo dip as part of the voicing process - if you know what angle of separation the speakers will be used at the frequency and depth of the dip is fairly predictable as peoples heads don't vary in size that much.
If you're referring to the stereo interference field above about 3Khz affecting phantom channel imaging, (phantom stability with listener movement etc) I think that comes down largely to two things - the typical insufficient angular separation between the two speakers (HRTF crosstalk >3Khz drops dramatically if you go from a typical 45 degrees of separation to 55-60 for example) and cabinet diffraction from the tweeters causing polar ripple in the treble for small listener position offsets. (The latter seriously degrades the phantom channel image and stability)
Sufficiently separate the speakers and eliminate diffraction from the tweeter (make it a true point source with smooth off axis response) and phantom image trouble largely goes away - even in a fairly dead room with damped side walls. My best listening room had curtains along both side walls at the speaker end so was fairly "dead" at high frequencies (the other end was live) and phantom channel imaging was excellent, much better than when the curtains were pulled back.
Simple solution that is great for centre channel dialogue, but that adds as many problems as it solves for music.
Sure you can produce a musical instrument dead ahead without any interference effects but if you have a TV the centre speaker is either too high or too low. (Pinhole projection screens aside) But more importantly what happens to the instrument that you're trying to image as coming from half way between centre and one side ?
The answer is that you're back to two speakers (centre and right say) both producing the same sound at different relative volume/phase/whatever, but this time you have half the angular separation than you had before (centre to right instead of left to right) which makes your high frequency "stereo" crosstalk MUCH worse.
The fact that there aren't any standards is a real shame, that doesn't mean they are impossible to implement, its just a commercial chicken and egg problem not a fundamentally unsolvable technical problem. There is a realistic, optimal window for reverberation time in a "small" room for example, that alone would be a start. Most living rooms are well above the optimal reverberation time, especially for wide dispersion speakers that prevail.
Last edited:
Especially when using the NTSC format (Never The Same Color).
... OK, jokes aside.
Stereo interference as seen by a microphone, stereobase 2m, distance 2m.
Curve 1: Center (should be flat, but isn't due to simulation artefacts)
Curve 2: 2,5cm side shift
Curve 3: 5,0cm side shift
Curve 4: 7,5cm side shift
(See picture1 below)
ILD front vs. 30-Degrees
http://www.analogue.org/network/images/ILD.gif
Attachments
Last edited:
- Status
- Not open for further replies.