A Broad Discussion of Speakers with Major Audio Luminaries

[Mnyb]

If you have been following my posts or have read my book(s) you would know that the direct sound dominates the impression of sound quality above the transition/Schroeder frequency. Directivity - i.e. early reflections - are significant in mono and stereo listening, but much less relevant in multichannel systems because the desirable recorded reflections dominate the listening room reflections. If you have chosen well-designed, neutral, loudspeakers EQ above about 500 Hz is not likely to be needed, and if done badly can degrade good loudspeakers. For the Nth time: room curves are results, not targets for EQ.

All that said, low frequency room resonances must be addressed in any mode of listening, meaning bass management, two or four subwoofers, and discrete EQ. All this has been discussed in my posts over the years, and is in my books.

Sorry to be sounding "commercial" by mentioning my books, but practically everything I write in this forum and others is in them - and more. It just would seem like a worthwhile investment in relevant knowledge. BTW technical books earn very little money - and I don't need it, so my involvement with this forum is purely a teaching exercise for me in my retirement. It bothers me that so much effort is expended in unsubstantiated chatter, when there is real knowledge that would elevate the discussion to a more interesting and rewarding level. Some of you forum members are already there, I know - and thank you for your continued support. The upcoming 4th edition should be a fairly definitive statement of where the audio industry stands, and there are some fresh observations on how we hear what gives us pleasure, along with state-of-the-art knowledge on headphones by Sean Olive. End of commercial message

Yes I actually adhered to your advice got myself decent speakers KEF LS60 and two subs , only use EQ under ca 200Hz to take care of some obnoxious room modes ( cutting peaks , no trying to boost nulls ) . Some adjustments to taste a mild bass shelf and some treble trimming.

 

[ex audiophile]

Yes I actually adhered to your advice got myself decent speakers KEF LS60 and two subs , only use EQ under ca 200Hz to take care of some obnoxious room modes ( cutting peaks , no trying to boost nulls ) . Some adjustments to taste a mild bass shelf and some treble trimming.

What are you using for eq?

 

[Mnyb]

What are you using for eq?

WiiM or ROON at the moment sub-xover in the speakers treble in the speakers

 

[amirm]

Sorry to be sounding "commercial" by mentioning my books, but practically everything I write in this forum and others is in them - and more. It just would seem like a worthwhile investment in relevant knowledge.

The books are essential for any audiophile enthusiast. I do however find a lot of value in your forum posts. They sometimes tell the same story with a different nuance that makes the message "stick." It also takes a lot less motivation to read a post than a book. So I hope you continue both!

 

[CinamonRolls]

The key to a truly compelling listening experience is properly set up multiple subwoofers that tame room resonances. Bass is about 30% of the overall impression of sound quality.

I read a famous multi-sub proponent (the one who built large waveguide two-ways) say multi-subs OR decorrelation filters. I'm not sure what decorrelation filters are but I think it might be reverb applied to the low frequency band. If that's accurate - maybe I totally misunderstood him - and digital processing power is increasing are we going to see a day where we can add decorrelation filters to our mains and get closer to multi-sub approaches?

I'd still want multi-subs for the output but if you only have two mains it would be nice if a signal processor could tame room resonances.

 

[Duke]

I read a famous multi-sub proponent (the one who built large waveguide two-ways) say multi-subs OR decorrelation filters.

I'm under the impression that he never actually came out with the decorrelation filters he was working on at one time. That would have been very interesting. Did I miss something?

 

[Floyd Toole]

I read a famous multi-sub proponent (the one who built large waveguide two-ways) say multi-subs OR decorrelation filters. I'm not sure what decorrelation filters are but I think it might be reverb applied to the low frequency band. If that's accurate - maybe I totally misunderstood him - and digital processing power is increasing are we going to see a day where we can add decorrelation filters to our mains and get closer to multi-sub approaches?

I'd still want multi-subs for the output but if you only have two mains it would be nice if a signal processor could tame room resonances.

Decorrelation is disorder - i.e. each source of low frequencies is in a random time sequence with others. It almost certainly will reduce coherent coupling to room modes which is good, but it will change the time domain structure of the bass. The purpose of multiple subs is to coherently couple energy to modes but in a fashion that attenuates specific sets of room resonances. With these attenuated, the recorded bass is heard without the detrimental effects of standing waves. Using subs means that the main speakers can play louder with lower distortion. For a single listener or a small number, near-field subwoofing is an alternative or an addition.

 

[Heinrich]

Decorrelation is disorder ...

Was it Dr Griesinger who attempted to use bass as a subjective indicator for big rooms? Phase, hence a manipulated (made up) reflection pattern would suggest spacial qualities?

 

[Floyd Toole]

Was it Dr Griesinger who attempted to use bass as a subjective indicator for big rooms? Phase, hence a manipulated (made up) reflection pattern would suggest spacial qualities?

David was most interested in concert halls, and in such large spaces even long bass wavelengths have space to propagate and be reflected from large surfaces. The delays generally responsible for impressions of envelopment are of the order of 80 ms and more. Small rooms are in a different category entirely. In the 3rd edition of my book, the topic of "stereo bass" is discussed, as this is the closest one can get to replicating something that might resemble events in a concert hall. David was among others who promoted the idea that stereo bass was a requirement, and at Harman (where he worked, as I did) a demonstration was set up to test the notion. I titled the section in my book "Stereo Bass: Little Ado about Even Less" with apologies to Shakespeare. That about sums it up, because substantial effort, including generating special signals to reveal differences showed that any audible spatial differences attributable to stereo bass occurs at frequencies above about 80 Hz. Others have found similar results - so bass managed subwoofers sacrifice nothing spatial, but multiple subs are huge advantages to the timbral quality of bass when room modes are tamed. Todd Welti wrote a paper summarizing the topic: Welti, T. (2004), “Subjective comparison of single channel versus two channel subwoofer reproduction”, 117thConvention, Audio Eng. Soc., Preprint 6322.

 

[JPA]

David was most interested in concert halls, and in such large spaces even long bass wavelengths have space to propagate and be reflected from large surfaces. The delays generally responsible for impressions of envelopment are of the order of 80 ms and more. Small rooms are in a different category entirely. In the 3rd edition of my book, the topic of "stereo bass" is discussed, as this is the closest one can get to replicating something that might resemble events in a concert hall. David was among others who promoted the idea that stereo bass was a requirement, and at Harman (where he worked, as I did) a demonstration was set up to test the notion. I titled the section in my book "Stereo Bass: Little Ado about Even Less" with apologies to Shakespeare. That about sums it up, because substantial effort, including generating special signals to reveal differences showed that any audible spatial differences attributable to stereo bass occurs at frequencies above about 80 Hz. Others have found similar results - so bass managed subwoofers sacrifice nothing spatial, but multiple subs are huge advantages to the timbral quality of bass when room modes are tamed. Todd Welti wrote a paper summarizing the topic: Welti, T. (2004), “Subjective comparison of single channel versus two channel subwoofer reproduction”, 117thConvention, Audio Eng. Soc., Preprint 6322.

Do you have an opinion on Double Bass Arrays, where multiple subwoofers on one wall of a room generate a bass signal, and other multiple subwoofers on the opposite wall absorb the signal? It appears to work well in theory and simulations, but I wonder if home theater seating (especially 2nd row seating on a riser) in the real world would disrupt the smooth flow of sound waves enough to reduce its effectiveness.

 

[Floyd Toole]

Do you have an opinion on Double Bass Arrays, where multiple subwoofers on one wall of a room generate a bass signal, and other multiple subwoofers on the opposite wall absorb the signal? It appears to work well in theory and simulations, but I wonder if home theater seating (especially 2nd row seating on a riser) in the real world would disrupt the smooth flow of sound waves enough to reduce its effectiveness.

These systems work, ideally best in perfectly rectangular, acoustically symmetrical, rooms. In the real world there will be deficiencies (risers would be a disruptor), but the principle is logical. It is discussed in Chapter 14 of the upcoming 4th edition of my book, co-authored with Todd Welti. The scheme is only practical for custom installations and requires at least four loudspeakers strategically located in/on each of the front and back walls. Since energy is created and then absorbed, the efficiency is not high, but in such installations cost is not likely to be a consideration. There are more cost effective, less disruptive, solutions that work in ordinary rooms, and that increase overall efficiency - more subwoofers are needed but they can be smaller. Having experienced such systems over many years in different rooms it is difficult to find fault with any of them - the bass is just as one wants, clean, deep and tight, and shared among multiple listeners. Massive single subs are illogical knowing what we now do.

 

[tktran303]

Dear Floyd,

What do you mean by absorbing? Are the woofers on the opposite side of the room out of phase with the ones on the front?

Have you used software, like Treble Technologies, to simulate/model the room, and it's floor, walls, furnishings etc, with the speaker(s) and the listener(s)?

BR,
Thanh

 

[Floyd Toole]

Dear Floyd,

What do you mean by absorbing? Are the woofers on the opposite side of the room out of phase with the ones on the front?

Have you used software, like Treble Technologies, to simulate/model the room, and it's floor, walls, furnishings etc, with the speaker(s) and the listener(s)?

BR,
Thanh

Yes, the idea is that the energy arriving at the wall is not allowed to reflect - it must be cancelled/absorbed to prevent standing waves from developing.

As I understand that model, it is another of the idealized/simplified simulations of direct and reflected sounds arriving at the listening position or area from loudspeakers. The belief is that the sum of all sound components is the key metric to sound quality. Proof of their success is how closely the calculated and measured steady-state sound fields compare. The assumption is that this metric is a reliable indicator of perceived sound quality. It isn't. Two ears and a brain interpret a complex sound field very differently from an omnidirectional microphone. Binaural hearing separates the direct sound from later arrivals allowing for directional perception - localization - and for key indications of sound quality. In addition there are various spatial effects including image size and the very important sensations of envelopment. All these perceptual dimensions combine in complex ways. Only at frequencies below the transition/Schroeder frequency is a steady-state room curve reliably indicative of sound quality. I have discussed this more than once in this forum, and it is all well explained in my books, especially in the upcoming 4th edition. Todd Welti has used room simulation models in developing his multi-subwoofer solutions, which he describes in detail in AES papers and in the 4th edition of my book where we share a chapter.

 

[tktran303]

Yes, the idea is that the energy arriving at the wall is not allowed to reflect - it must be cancelled/absorbed to prevent standing waves from developing.

This was done at least 20 years ago, IIRC.

It results in bass response that is similar to a dipole woofers at the MLP,

I'll speak to someone and see if I can some data to share...

 

[Floyd Toole]

This was done at least 20 years ago, IIRC.

It results in bass response that is similar to a dipole woofers at the MLP,

I'll speak to someone and see if I can some data to share...

"similar to a dipole woofers at the MLP" - not possible. Dipole sources radiate opposite polarity on each side simultaneously - they are "velocity" sources, so they couple energy into the room standing waves best at the null locations, where they are sensitive to orientation. Used intelligently, this can be used to attenuate specific room modes by varying the energy coupled into them. This is, of course, fundamentally different from conventional "pressure" sources. These too can be positioned to vary coupling into modes, but at the high pressure regions of standing waves.

The subwoofer scheme being discussed here radiates sound from one wall, and an inverted version of that sound is radiated at the opposite wall delayed by the propagation time across the room. The theory is that the sound travels only one direction across the room and then disappears. It has no resemblance to the behaviour of a dipole source.

 

[olieb]

I found this discussion coming from a thread on diyaudio.
But these answers got me a bit confused.

These systems work, ideally best in perfectly rectangular, acoustically symmetrical, rooms. In the real world there will be deficiencies (risers would be a disruptor), but the principle is logical.

As I understand that model, it is another of the idealized/simplified simulations of direct and reflected sounds arriving at the listening position or area from loudspeakers. The belief is that the sum of all sound components is the key metric to sound quality. Proof of their success is how closely the calculated and measured steady-state sound fields compare. The assumption is that this metric is a reliable indicator of perceived sound quality. It isn't.
....
Only at frequencies below the transition/Schroeder frequency is a steady-state room curve reliably indicative of sound quality.

Now, does these special kind of source-sink models (DBA-subwoofers) work as intended or are they simplified simulations that do not?
Are they chasing a wrong metric or is the metric not so wrong for subwoofers below Schroeder frequency?
I understand the limitations in terms of room layout, but why would the characteristics of directional hearing play a role here for the performance of sound reproduction in the subwoofer range? If I understand correctly, you just explained that spatial hearing below 80Hz is so limited that bass management does not sacrifice anything in that regard.

 

[Juhazi]

Large speakers with good low extension will help to smooth out bass response, compared to a single sub.
Stereo low bass in music recordings is irrelevant despite it is in the recording, harmonics will make the localization.

My setup is in no way optimized for the room, speaker location is dictated by general interior design ot the living room. Speakers are symmetrically along the long wall. Here you can also see the effect of mic moved 80cm front-back. There is no way to correct that by dsp or multisub. I have tested/measured 1, 2 and 3 diststributed subs in my small HT room and definitely 3 was best in every way without any dsp. Sadly that was just a test and helper "subs" didn't go below 40Hz

 

[Floyd Toole]

I found this discussion coming from a thread on diyaudio.
But these answers got me a bit confused.



Now, does these special kind of source-sink models (DBA-subwoofers) work as intended or are they simplified simulations that do not?
Are they chasing a wrong metric or is the metric not so wrong for subwoofers below Schroeder frequency?
I understand the limitations in terms of room layout, but why would the characteristics of directional hearing play a role here for the performance of sound reproduction in the subwoofer range? If I understand correctly, you just explained that spatial hearing below 80Hz is so limited that bass management does not sacrifice anything in that regard.

Below 80 Hz two phenomena combine to dominate our perceptions, it seems. One is that our binaural systems fail badly at localizing sound at such low frequencies/long wavelengths (our ears are too close together). We do not hear a distinct "direct" sound, so "time alignment" of individual subwoofers is not required. Instead, in small listening rooms resonances/standing waves absolutely dominate bass sound quality, strongly accentuating and attenuating specific bass frequencies and resulting in no two listeners hearing the same sound. The traditional solution has been to use large areas of low-frequency absorbers (bass traps). I may have been the first to publish results of deliberate mode manipulation using multiple subwoofers (AES 1990). Since then Todd Welti and numerous others have offered variations on active and passive schemes using multiple subwoofers to manipulate the energy in specific room modes to optimize performance at specific locations or over specific areas. The source/sink method being discussed is one of them. They all are beneficial, and none of them rely on bass traps, although low-frequency absorption is always an asset if it can be aesthetically achieved. Steady-state measurements are adequate to describe performance at very low frequencies, and because strong room resonances behave as minimum-phase phenomena, the time domain (ringing) is also corrected when peaks in the frequency response are attenuated. There is a generous amount of misunderstanding about this in the audio industry. It has been discussed in earlier editions of my book, and in the literature. Todd Welti and I share a chapter in the 4th edition discussing the subject. Bass in small rooms can be managed, and it must because low frequency performance accounts for about 30% of our overall assessment of sound quality. We all listen in small rooms. This problem is sufficient to invalidate, or at least cast doubt on most casual subjective evaluations.

 

[olieb]

Thank you for the reply.
So this is what I take away:
Below ≈80Hz

• we „hear“ only the spl presented at the ear(s) [no spatial cues]
• room modes can be described by steady-state terms and measurements
• they are minimum phase phenomena and can be EQed correcting the time domain in the process
• multiple-subwoofer schemes work and that includes DBA

But would you agree/disagree that this bass array idea is different from most other „mode manipulation“ concepts as it tries to „erase“ the effect of low frequency modes altogether?
If it works (and that is a big IF of course) the sound would not propagate by means of low frequency modes but would transform the rectangular room into a duct with more or less fully absorptive termination.
One difference would be that the sound field is position independent to a high degree and mode decay would be nearly absent.

 

[Floyd Toole]

Thank you for the reply.
So this is what I take away:
Below ≈80Hz

• we „hear“ only the spl presented at the ear(s) [no spatial cues]
• room modes can be described by steady-state terms and measurements
• they are minimum phase phenomena and can be EQed correcting the time domain in the process
• multiple-subwoofer schemes work and that includes DBA

But would you agree/disagree that this bass array idea is different from most other „mode manipulation“ concepts as it tries to „erase“ the effect of low frequency modes altogether?
If it works (and that is a big IF of course) the sound would not propagate by means of low frequency modes but would transform the rectangular room into a duct with more or less fully absorptive termination.
One difference would be that the sound field is position independent to a high degree and mode decay would be nearly absent.

The propagating "wave front" attempts to eliminate length modes only, but the initiating subwoofers do not launch a plane wave. The width and height modes are separate issues, and in their installations they are attenuated, not eliminated, by locating the subwoofers at locations 25% in from the floor, ceiling and side walls. Oblique modes might sneak in, but if all others are attenuated, they are likely to be negligible. Todd Welti's Sound Field Management calculation would recognize it as a desirable option. Having lived with and experienced several versions of SFM, they all can provide superb bass entertainment for multiple listeners. There are multiple desirable physical configurations if signal processing of individual subs is involved.