Diffusing 1st reflections of speakers that measure great on and off-axis - instead of absorbing

[Duke]

Thank you very much Bjorn for taking the time to write an in-depth and educational reply.

There isn't to my knowledge an objective definition of early and later reflections. But normally we're talking about early reflections till 10-12 ms area and late after that.

Ten milliseconds is the figure I arrived at experimentally about twenty years ago, and is also the figure Earl Geddes uses. Acoustician David Griesinger comes up with this same figure, though in a different context (concert hall and lecture hall acoustics and psychoacoustics):

"Transients are not corrupted by reflections if the room is large enough - and 10ms of reflections free time is enough."

The RFZ or ISD (initial signal delay) gap of that graph is around 18 ms. Where you see the sudden rise in level, is where you have the termination of the ISD gap.

The graphs are confusing to me. In the second figure you posted, it looks to me like the Direct Sound arrives around 8 milliseconds, and that the strong onset of reflections is around 18 milliseconds, and that the part labelled RFZ only lasts for about 10 milliseconds. I don't expect you to troubleshoot my misunderstanding, but it's not clear to me that the RFZ gap is 18 milliseconds.

For example the termination of the ISD gap is very important. The level will decide how much liveliness of space you achieve and how well you can cover earlier arriving audible reflections. While treatment will attenuate specular reflections, it's common to still have some minor ones that's audible. A strong termination of the ISD gap tricks the brain to overlook these besides giving a boost of energy/liveliness which is very pleasing. The purpose of the ISD termination is also to remove the localization cues of the later arriving energies and to reinforce the localization cues of the direct energy.

THIS is FASCINATING!! So if I understand correctly, we want a STRONG rather than gradual onset of diffuse reflected energy to clearly demarcate the termination of the ISD gap. Over in the world of concert hall acoustics the "scale" is different but I think the same general principle applies:

"Envelopment is perceived when the ear and brain can detect TWO separate streams: A foreground stream of direct sound, and a background stream of reverberation. Both streams must be present if sound is perceived as enveloping." - David Griesinger

Implied is a time gap separating the direct sound and the onset of reverberation... in a small room, the latter being presumably analogous to the onset of reflections at the termination of the ISD gap.

But you are the first one to point out to me the "a strong termination of the ISD gap" plays a critical role. THANK YOU!! IF it matters a lot that this strong onset of reflections be diffuse instead of specular, one implication is that you need a lot of channels to do multichannel right because just having a small number of rear-channel speakers mimics specular reflections, at least as far as the arrival directions go.

I may have more questions, but let me do some homework first and see if I can answer them for myself. [edit: My homework took an expensive turn; I just bought D'Antonio and Cox's book.)

 

[Duke]

What if you deliberately added into your main signal a 6 ms delayed signal...to purposefully overwhelm inadvertent room reflections with intentional fake ones?

My understanding from Toole's book as that perceptually the worst possible direction for reflections to arrive from is the exact same direction as the direct sound. Such reflections tend to be perceived as coloration rather than spaciousness. The same reflection (with the same delay) arriving from a significantly different direction is far more likely to be beneficial, if I understand correctly.

 

[March Audio]

Actually, it is possible to define direct - reflected - diffuse. And in the answer lies the clue to understand the difference between small and large rooms:

Direct is easy - sound only from the speaker, it has pressure amplitude and direction (Exact properties depends on the radiator).
Reflected is sound reflected 1 or more times from a surface, sound that has pressure amplitude and still has direction.
Diffuse sound is sound that has reflected from several surfaces, mixed together so the resulting sound field has pressure amplitude, but no direction.

There is a difference between discrete reflections and diffuse reverb in that the former actually has direction, the latter has no direction. And for a diffuse field to develop, it requires a space large enough to create enough reflections from different angles at different time.

Since the diffuse sound has no direction, it does not affect the direct sound in the same way that reflections with direction does. Then we also understand why it is necessary to attenuate reflections in a small room.

I think @j_j point was that the perceptual effect is rather more difficult to define.

Also from that perspective when is it "diffuse enough".

As he says it's a continuum and to talk interns of large and small is simplistic to say the least.

 

[patate91]

Seems to me "direct" is pretty straightforward, and it seems to me that, along at least one continuum, reflections can range from specular to diffuse. I welcome correction if I'm mistaken.



In the size rooms we see in home audio, what are the relevant perceptual breakpoints?



Nothing against J_J, but I have a hard time following him, and regarding room acoustics he's already told me "dead is my preference." Dead is not my preference so at some point, even if I am able to follow J_J, our paths will probably diverge unless he changes my mind. (I'm not challenging him to do so, but if it happens, it happens.)

Here is a splicing together of Bjorn's thoughts from two different posts:

"The clarity, detail or "resolution" is brought to a completely another level when early arriving specular reflections are strongly attenuated... The goal is to have a lateral late arriving diffuse tail which... yields spaciousness and envelopment without obscuring the recorded signal. You achieve both accuracy with a high degree of clarity, intelligibility, localization and correct tonality combined with a spacious and enveloping experience."

This is where I think he and I have some common ground. I'm pretty sure he knows more than I do so I hope to learn from him, even if I'm still stumbling over the meaning of terms here and there.

I think at this point you could post some rew measurments (with all the time domain related informations).

 

[Plcamp]

My understanding from Toole's book as that perceptually the worst possible direction for reflections to arrive from is the exact same direction as the direct sound. Such reflections tend to be perceived as coloration rather than spaciousness. The same reflection (with the same delay) arriving from a significantly different direction is far more likely to be beneficial, if I understand correctly.

Then maybe using the satellites of a 7.1 decoder could choose the direction from which drivers (perhaps even holographically) deliver the delayed signal...why should it not be possible for a dsp to figure that out...and correct the dispersive imbalance for critical listening, instead of movie special effects?

 

[Duke]

As he says it's a continuum and to talk interns of large and small is simplistic to say the least.

Sometimes we can infer from the context which part of the continuum we're talking about.

If we read that home audio listening rooms are "small", I think we can infer that they are small relative to other types of acoustic spaces (classrooms, music venues, theaters, auditoriums, stadiums, etc.).

And if we read that so-and-so has a "large listening room" in his house, I think that can infer that is it large relative to most home audio listening rooms, but not in relation to other types of acoustic spaces.

So perhaps the terms "large" and "small" are simplistic, but in some contexts they provide sufficient information. In my opinion.

I think at this point you could post some rew measurments (with all the time domain related informations).

Could you tell me why you want me to post raw measurements with all time domain related information? Is there a particular point you want me to prove?

Then maybe using the satellites of a 7.1 decoder could choose the direction from which drivers (perhaps even holographically) deliver the delayed signal...why should it not be possible for a dsp to figure that out...and correct the dispersive imbalance for critical listening, instead of movie special effects?

Sounds good to me. I'm not going to do it, but you can!

 

[patate91]

Sometimes we can infer from the context which part of the continuum we're talking about.

If we read that home audio listening rooms are "small", I think we can infer that they are small relative to other types of acoustic spaces (classrooms, music venues, theaters, auditoriums, stadiums, etc.).

And if we read that so-and-so has a "large listening room" in his house, I think that can infer that is it large relative to most home audio listening rooms, but not in relation to other types of acoustic spaces.

So perhaps the terms "large" and "small" are simplistic, but in some contexts they provide sufficient information. In my opinion.



Could you tell me why you want me to post raw measurements with all time domain related information? Is there a particular point you want me to prove?



Sounds good to me. I'm not going to do it, but you can!

I'm curious about the data from a "good" reflecting room. Your position is about enveloppement, it should shows in measurements.

It's not about proving something, it's about sharing and maybe learning.

 

[Duke]

I'm curious about the data from a "good" reflecting room. Your position is about enveloppement, it should shows in measurements.

It's not about proving something, it's about sharing and maybe learning.

Thanks for clarifying.

I'm not a room acoustics guy, nor do I have a "good" room myself. So at this point I really don't have anything to share.

I'm a speaker designer, interested in "where the goal posts are" for small room acoustics, in case there's a window of opportunity to make speakers a bit more room-friendly.

For instance, let's revisit this image, which is apparently the actual energy-time curve for a "good" room:

Eyeballing this through the lense of a speaker designer, what might be desirable is a radiation pattern which minimizes the floor and ceiling bounces as well as early sidewall reflections, so that we don't have any strong reflections from those areas within the time gap labelled "RFZ". If the speaker can do this much on its without needing dedicated room treatment at those first reflection zones, arguably the speaker/room interaction may be a bit better. This implies a narrow radiation pattern may be desirable. (Treating the wall behind the listener will probably still be quite beneficial because those reflections cannot be minimized by driver radiation pattern alone).

Something else implied by this graph is that a surge of well-energized, diffuse, fairly late-onset in-room reflections is beneficial, which implies that a narrow radiation pattern may NOT be desirable if it results in insufficient energy going into those desirable later reflections.

So an ideal "best of both worlds" approach may be a wide-pattern speaker + appropriate acoustic treatments at all applicable first reflection zones. This implies a dedicated listening room.

The market segment I want to work is people who have less-than-ideal listening spaces (like mine!), including those wherein dedicated room treatment may not even be feasible. In such a room, I think a fairly narrow radiation pattern which inherently minimizes the floor and ceiling bounces and early sidewall reflections may be preferable. Not that there wouldn't still be worthwhile benefit from appropriate room treatments, but hopefully their use would not be as "necessary" with relatively narrow-pattern speakers.

 

[Kvalsvoll]

I think @j_j point was that the perceptual effect is rather more difficult to define.

Also from that perspective when is it "diffuse enough".

As he says it's a continuum and to talk interns of large and small is simplistic to say the least.

I am not so much into semathics and definitions, and the world is rarely strict yes-or-no.

But there are significant differences in the reflected energy (can I say that? Can't call it "reverb" because then someone will come by claiming there is no such thing - which is at least partly correct.) in a small room like my listening room, compared to a large concert hall. In the small room the "reverb" will be dominated by reflected sound with direction, where as in the large hall most of the energy has no direction.

This means that what you see in a RT60 number for a small room, is not the same as what you see in a large room.

 

[March Audio]

Sometimes we can infer from the context which part of the continuum we're talking about.

If we read that home audio listening rooms are "small", I think we can infer that they are small relative to other types of acoustic spaces (classrooms, music venues, theaters, auditoriums, stadiums, etc.).

And if we read that so-and-so has a "large listening room" in his house, I think that can infer that is it large relative to most home audio listening rooms, but not in relation to other types of acoustic spaces.

So perhaps the terms "large" and "small" are simplistic, but in some contexts they provide sufficient information. In my opinion.



Could you tell me why you want me to post raw measurements with all time domain related information? Is there a particular point you want me to prove?



Sounds good to me. I'm not going to do it, but you can!

Without measurement data assumptions can be dangerous

 

[patate91]

Thanks for clarifying.

I'm not a room acoustics guy, nor do I have a "good" room myself. So at this point I really don't have anything to share.

I'm a speaker designer, interested in "where the goal posts are" for small room acoustics, in case there's a window of opportunity to make speakers a bit more room-friendly.

For instance, let's revisit this image, which is apparently the actual energy-time curve for a "good" room:

View attachment 76371

Eyeballing this through the lense of a speaker designer, what might be desirable is a radiation pattern which minimizes the floor and ceiling bounces as well as early sidewall reflections, so that we don't have any strong reflections from those areas within the time gap labelled "RFZ". If the speaker can do this much on its without needing dedicated room treatment at those first reflection zones, arguably the speaker/room interaction may be a bit better. This implies a narrow radiation pattern may be desirable. (Treating the wall behind the listener will probably still be quite beneficial because those reflections cannot be minimized by driver radiation pattern alone).

Something else implied by this graph is that a surge of well-energized, diffuse, fairly late-onset in-room reflections is beneficial, which implies that a narrow radiation pattern may NOT be desirable if it results in insufficient energy going into those desirable later reflections.

So an ideal "best of both worlds" approach may be a wide-pattern speaker + appropriate acoustic treatments at all applicable first reflection zones. This implies a dedicated listening room.

The market segment I want to work is people who have less-than-ideal listening spaces (like mine!), including those wherein dedicated room treatment may not even be feasible. In such a room, I think a fairly narrow radiation pattern which inherently minimizes the floor and ceiling bounces and early sidewall reflections may be preferable. Not that there wouldn't still be worthwhile benefit from appropriate room treatments, but hopefully their use would not be as "necessary" with relatively narrow-pattern speakers.

Thanks for your reply.

I think that one day or another you'll have to measure your speaker designs in a couple of non treated rooms. The data will certainly be usefull for a lot of people.

 

[March Audio]

I am not so much into semathics and definitions, and the world is rarely strict yes-or-no.

But there are significant differences in the reflected energy (can I say that? Can't call it "reverb" because then someone will come by claiming there is no such thing - which is at least partly correct.) in a small room like my listening room, compared to a large concert hall. In the small room the "reverb" will be dominated by reflected sound with direction, where as in the large hall most of the energy has no direction.

This means that what you see in a RT60 number for a small room, is not the same as what you see in a large room.

I think this whole conversation came about from not so much semantics but an extreme interpretation of the definitions and scenarios where it occurs. As discussed at the beginning of the thread no-one takes a single frequency RT60 number and bases treatments on it.

Will the "reverb" be perceptually dominant? How big a room before it isnt perceptually dominant? This question isnt semantics.

I would defer to @j_j at this point as its his area of expertise.

 

[Kvalsvoll]

I think this whole conversation came about from not so much semantics but an extreme interpretation of the definitions and scenarios where it occurs. As discussed at the beginning of the thread no-one takes a single frequency RT60 number and bases treatments on it.

Will the "reverb" be perceptually dominant? How big a room before it isnt perceptually dominant? This question isnt semantics.

I would defer to @j_j at this point as its his area of expertise.

That would be very welcome (@j_j , I saw that video presentation from another thread, and found it very valuable.), if he still reads this thread. Knowledge about how hearing actually works, and how we perceive sound, is of course important for a better understanding.

Differences in properties of the sound field is that the small room has reflected sound arriving at quite high level very early in time, and that sound also has direction. The large space has a relatively low level of early reflections, and mostly diffuse, later arrival sound, which decays at a much slower rate.

 

[Kvalsvoll]

If the speaker can do this much on its without needing dedicated room treatment at those first reflection zones, arguably the speaker/room interaction may be a bit better.

My experience is that it is necessary to have to those desired properties in a speaker, to be able to achieve a very good and realistic reproduction. And such speakers will also work much better in normal, horrible rooms.

 

[Duke]

Without measurement data assumptions can be dangerous

Okay, let me try again.

IF terms like "large" and "small" are "too simplistic", as you say, what do you propose as alternatives?

And if assumptions are dangerous without measurement data, how much measurement data do we need before using adjectives to describe the size of a room?

In other words, I DO NOT THINK we need to wait for somebody to produce data before we use words like "large" and "small" when referring to acoustic spaces. Nothing against acoustic data, but we don't ALWAYS need it to have a conversation.

 

[March Audio]

Okay, let me try again.

IF terms like "large" and "small" are too simplistic, as you say, what do you propose as alternatives?

And if assumptions are dangerous without measurement data, how much measurement data do we need before using adjectives to describe the size of a room?

Again, for this I would tend to defer to @j_j for specifics. He is the expert in this area. My main issue here has been with the extreme black and white interpretation that has been taken by some (one), when the reality is far more subtle.

 

[Duke]

Again, for this I would tend to defer to @j_j for specifics. He is the expert in this area. My main issue here has been with the extreme black and white interpretation that has been taken by some (one), when the reality is far more subtle.

Well I thought I was having a conversation with you, but okay.

 

[localhost128]

But you are the first one to point out to me the "a strong termination of the ISD gap" plays a critical role. THANK YOU!! IF it matters a lot that this strong onset of reflections be diffuse instead of specular, one implication is that you need a lot of channels to do multichannel right because just having a small number of rear-channel speakers mimics specular reflections, at least as far as the arrival directions go.

LEDE requirements for the termination (sharply delineating the effectively anechoic ISD-gap) are "not be below -12dB from Ld (direct signal)". this can actually be quite difficult to do in small rooms as diffuse returns that high in magnitude will require sufficiently broadband diffusers - that then in themselves have greater minimum distance requirements which must be aligned with the chosen ISD length. this is usually accomplished with nested diffusers (a large array constructed of smaller/individual diffusers that has a much lower design frequency than the individuals). there could also be a haas kicker, which induces a high-gain specular reflection from a large (wrt wavelength) flat/planar panel installed at the rear side-walls and angled towards the listening position - arriving at the termination, or slightly later (nested within the diffuse returns).

 

[Duke]

Thanks for your reply.

I think that one day or another you'll have to measure your speaker designs in a couple of non treated rooms. The data will certainly be usefull for a lot of people.

Hmmm. I don't have the ability to do energy time curves, at least not yet, so let me think about it.

 

[localhost128]

Actually, it is possible to define direct - reflected - diffuse. And in the answer lies the clue to understand the difference between small and large rooms:

Direct is easy - sound only from the speaker, it has pressure amplitude and direction (Exact properties depends on the radiator).
Reflected is sound reflected 1 or more times from a surface, sound that has pressure amplitude and still has direction.
Diffuse sound is sound that has reflected from several surfaces, mixed together so the resulting sound field has pressure amplitude, but no direction.

There is a difference between discrete reflections and diffuse reverb in that the former actually has direction, the latter has no direction. And for a diffuse field to develop, it requires a space large enough to create enough reflections from different angles at different time.

Since the diffuse sound has no direction, it does not affect the direct sound in the same way that reflections with direction does. Then we also understand why it is necessary to attenuate reflections in a small room.

++. and also that in Small rooms, no reverberant sound-field exists to mask those reflections.

hence why Large rooms with reverberant sound-field are treated with absorption placed statistically (randomly) throughout the room to "bring down the RT60 times" as the sound-field has equal/probable energy flows in all directions simultaneously (no direction, as you state) - versus Small rooms that lack a reverberant sound-field and instead we utilize time-domain analysis to identify specific (localized) boundary points incident of a deemed-destructive indirect, discrete, focused, specular reflection and place absorption (or other treatment) accordingly.

and thus also why analysis of porous absorbers must be understood in the context of the lab/test environment vs that of where it will be utilized, as stated earlier in the thread. many data specs are measured based on reverberation chamber method (while also only being valid down to certain lower freq cutoff), but when placed in small room the angle of incidence of the reflection can easily be determined - so the data is not directly relatable.

Manfred Schroeder did a lot of work in this space.

"Small" rooms (from the standpoint of how they are defined in acoustical engineering), deal with modal resonances and focused (sparse) specular reflections - all of which are localized energy with time-arrival and vector/direction properties. just as there is a transition from modal to specular (due to wavelength size vs a particular dimension), there is a "transition" region in Large rooms from specular to diffuse/reverberant.

Blackbird Studio C is one "Small" room that goes to great lengths to emulate a "diffuse field" (via the 27tons of PRDs on the walls), yielding an increadilby dense, reflection-rich, "random" response - albiet at remarkably low gain, as the returns are -30dB down from direct signal due to the spreading of energy and inherent losses (esp edge diffraction) from 2D RPGs.