The perfect speaker is room dependent - wide vs. narrow directivity and more

[Duke]

This precedence effect gonna make your brain zoom in to the first arriving sound ( the direct sound from the speakers ) and your brain gonna attenuate the later reflected sound ( 5 ms or more ) coming from the room as much as -10 dB !

The Precedence Effect has to do with the localization of sound sources. It is the localization cues of the reflections which are suppressed by about 10 dB; the ear is still picking up loudness (and therefore timbre) cues within the time interval the Precedence Effect is in, um, effect.

 

[tmuikku]

... This effect gonna dominate if you listen to speakers at a distance longer than 5 ms soundtravel , meaning distances longer than 1,7 meters ...

I think here is a misconception, at least this easily creates one.

Precedence effect doesn't mind distance to speaker, its about path length difference between direct sound and reflected sound, delay before reflection arrives to listener after direct sound. Speaker distance to boundaries and listener affects the reflection paths and their lengths as well as direct sound path length.

Thought experiment: one could keep listening distance constant while changing distance to all boundaries relative to the speaker and listener and change the delay(s) of first reflections, affect precedence effect and all other effects due to reflections.

 

[Tangband]

I think here is a misconception, at least this easily creates one.

Precedence effect doesn't mind distance to speaker, its about path length difference between direct sound and reflected sound, delay before reflection arrives to listener after direct sound. Speaker distance to boundaries and listener affects the reflection paths and their lengths as well as direct sound path length.

Thought experiment: one could keep listening distance constant while changing distance to all boundaries relative to the speaker and listener and change the delay(s) of first reflections, affect precedence effect and all other effects due to reflections.

Yes, maybe I was unclear , and thanks for correcting .
Its a very complicated matter in reality. In a normal room you always have delayed reflections. You listen to them, can measure them with microphones, but the brain selects the sound from the reflections very different to the microphone when the reflections arrive with some delay.

The time delay from those reflections has a huge impact on how we hear the sound.
Reflections with smaller timedelays, like 1 ms - ( 34,3 cm ) the listener gonna hear only one sound.
Reflections with more than 5 ms (1,7 meter ) the brain starts to separate the first arriving sound, the direct sound from the speaker, from the reflected sound ( the walls ) - but the brain will zoom in for the first arriving sound being the dominant, and the brain will attenuate the reflected sound with as much as -10 dB. This is very different from how a measurement microphone works, - it takes up all the sound, while the brain selects sound.

This makes measurements in a room from listening position very complicated If one wants measurement results that correlate well with what the brain hears. You cant measure anything of value from the listening position in a normal room above 300 Hz. Measuring a loudspeaker from 500 -20000 Hz in a normal room needs gating technique and measurements on/off axis near the loudspeaker.

As John Atkinsson used to say at stereophile : All measurements tell lies.

The picture shows the direct sound and the reflected sound.
In my opinion, for optimal 2-channel playback in a room, the reflected sound from the walls will be delayed to the listener about 20-25 ms, because of the distance from the sidewalls. This will fill up the stereo system faults and the sound will appear to be better in most cases. This demands a rather big listening room, bigger than the picture. It takes about 5 meters to the sidewalls at the sides of every speaker to get the needed 20-25 ms delay, compared to the direct sound from the speaker .
Narrow directivity speakers or wide doesnt matter much , but good directivity is needed.
Try it.

 

[tmuikku]

20ms is roughly 6 meters, lets call 7 meters, this needs to be on top of direct sound which could be 3 meters for easy math. So, 10m from speaker through wall to listening position quite large room.

10 meter reflection path happens on smaller rooms with third order reflections, perhaps some second order ones but no way first order reflections. Speaker and listener would have to be roughly 5 meters away from any boundary. I assume most hobbyist rooms arent big enough to satisfy this requirement.

One can attenuate the first order reflections with narrow directivity system, down to some frequency. Also the reverberation field would attenuate (in relation to direct sound) with narrower directivity speaker. Thought experiment: how to get nice spaciousness and envelopment with some clarity? Use wide enough (buy not too wide) coverage angle speaker to get the reverberation just loud enough there is enough spaciousness to be enjoyable and then treat the first reflections in room to get some clarity, or imaging or whats the right term, if there is need to.

Without acoustic treatment but having narrow directivity speakers toe-in the speakers so they cross in front of listening position, this would increase sound toward opposing sidewall reflection (usually the longest first reflection path length) and reduce direct sound some as well as all other first reflections with shorter path length, decreasing direct to reflected sound ration to get some spaciousness. 99% of hifi setup photos in the internets dont do this. Perhaps because they are not constant directivity, or at least smooth directivity, and are not designed to be listened off axis. Design for this and find out which coverage angle works for those ~20 square meter rooms many have and there you have it, speaker that works with that kind of room. Don't forget to take account room height, speaker and listening height.

After all we listen pressure variations in a room and the speaker needs to radiate sound nicely, at optimal position, to all directions so that the pressure variation is fine at the observation point, fine enough so the hearing system finds it pleasing.

 

[Tangband]

Can you show me such a speaker in a picture ?

Agree to some points . Its difficult to make the smaller room fixed for really good 2 channel sound . Fixing good directivity in loudspeakers also in the lower midrange gonna be big speakers , not suitable for high WAF . Or am I missing something ?

One other approach is to just give up and build the smaller room like a studio - a lot of damping and no help to fill up the stereo system faults at all. But this might be better than a small room without acoustical treatment ?

 

[tmuikku]

I don't know what kind of speaker that is controlled directivity, many ways to do one. What sounds best, no opinion yet, its a hobby and no hurry to goal as the journey is very much fun. Currently exploring effects of early vertical reflections, and still in mono after several years slow progress prototyping DIY speaker system For family livingroom situation, near wall and all, no acoustic treatment other than furnishing.

 

[tmuikku]

Fixing good directivity in loudspeakers also in the lower midrange gonna be big speakers , not suitable for high WAF . Or am I missing something ?

Naked open baffle system, hard to beat if you can position the speakers far enough from walls. Smallest and easiest to manufacture structure, as there is hardly one, good smooth directivity given enough ways. If need to position the system near wall then cardioidish woofer and waveguided tweeter (and some bass box) would work better, not very high DI but cheats in physical size vs pattern control like open baffle, trading of bass. 8" woofer in minimal enclosure has controlled directivity down to 200Hz or so but DI is not very high, perhaps enough.

 

[test1223]

In a smaller room, you dont have that 20-25 ms reflections needed for a good illusion. The first reflections in the small room from 5 - 10 ms gonna make the sound less clear. Then you can experiment with narrow beaming directivity loudspeakers. But this is not an optimal solution for two channel playback.

Yes, the delay from the side wall reflection of about 20 to 25ms increases envelopment of the sound without much negative effect.

10 meter reflection path happens on smaller rooms with third order reflections, perhaps some second order ones but no way first order reflections. Speaker and listener would have to be roughly 5 meters away from any boundary. I assume most hobbyist rooms arent big enough to satisfy this requirement.

You also need the direction of the reflected sound. The delay on its own isn't enough, as far as I know.

Without acoustic treatment but having narrow directivity speakers toe-in the speakers so they cross in front of listening position, this would increase sound toward opposing sidewall reflection (usually the longest first reflection path length) and reduce direct sound some as well as all other first reflections with shorter path length, decreasing direct to reflected sound ration to get some spaciousness.

It is interesting if this actually works this way since the reflection has its origin form the other speaker and is therefore mirrored?

--

To get an estimate of the room size I calculated the room width. I assume a standard stereo triangle (equal triangle, listening distance = width between L and R speaker).

Listening distance [m]	Room width (20ms delay) [m]	Room width (22.5ms delay) [m]	Room width (25ms delay) [m]
1	8.14	9.00	9.87
2	9.35	10.22	11.09
3	10.53	11.40	12.38
4	11.68	12.57	13.45

You can see that the room has to be bigger for this.

If you want to calculate it by yourself you can solve the following formula:
w=2*b+a
b=-a/(2*2^0.5)+(a^2/8+0.5*c*d*a+0.25*c^2*d^2)^0.5
Where w is the room width [m], b is the distance from the speaker to the nearest side wall [m], a is the listening distance aka length of the stereo triangle [m], c the speed of sound [m/s] and d the delay (s).

Since the best angle to create envelopment is about 60° and you can't get 25ms of delay with a standard side wall, you have to apply a trick by placing absorbers or diffusors at the side wall and use a "acoustic mirror" or a blank wall at about 60° from the listening position to get the desired 60° angle reflection.

You also can mimic these reflection with additional speaker with attenuation and delay. But you have to be caution since these speakers create new early reflections.

 

[Byrdsmaniac]

After octave to octave balance and timbral accuracy, wide dispersion is an absolute requirement for me. I have never yet encountered a situation where I have preferred a narrow dispersion speaker over a wide dispersion one. Not once. And I have heard quite a few speakers in a lot of different surroundings.

 

[test1223]

After octave to octave balance and timbral accuracy, wide dispersion is an absolute requirement for me. I have never yet encountered a situation where I have preferred a narrow dispersion speaker over a wide dispersion one. Not once. And I have heard quite a few speakers in a lot of different surroundings.

Which wide di speaker do you have and at which listening distance do you think your speaker sounds the best?

I guess it is at about 1m listening distance if it is a typical wide di speaker with a sound power di around 5dB from 300Hz to 4kHz. If this is the case it is most likely that you actually prefer a more narrow but constant di speaker at higher listening distance. As @Bjorn already mentioned these speakers are rare...

 

[Byrdsmaniac]

Which wide di speaker do you have and at which listening distance do you think your speaker sounds the best?

I guess it is at about 1m listening distance if it is a typical wide di speaker with a sound power di around 5dB from 300Hz to 4kHz. If this is the case it is most likely that you actually prefer a more narrow but constant di speaker at higher listening distance. As @Bjorn already mentioned these speakers are rare...

Ascend Acoustics HTM 200 SE at 8', Monitor Audio Silver 8i at 8.5' and Ascend Acoustics Sierra 2EX AT 9.5'. All in very different locations. The 200s are close to the front wall, the Silver 8is are away from any walls, and the Sierras are in a corner. I've heard speakers from most major brands in all kinds of rooms from tiny apartments to quite big rooms, (I once had a studio that was 20x30' with 18' ceilings), to recording studios. Whenever I've auditioned speakers and then checked measurements afterwards, the ones I preferred have been the ones with flat smooth frequency response and wide even dispersion. Violins have never sounded right to me on narrow dispersion speakers.

 

[Duke]

Tmuikku wrote:

Without acoustic treatment but having narrow directivity speakers toe-in the speakers so they cross in front of listening position, this would increase sound toward opposing sidewall reflection (usually the longest first reflection path length) and reduce direct sound some as well as all other first reflections with shorter path length, decreasing direct to reflected sound ratio to get some spaciousness. [emphasis Duke's]

And Tangband asked:

Can you show me such a speaker in a picture ?

Duke chimes in with a photo of speaker whose radiation pattern from the crossover region on up is 90 degrees (-6 dB @ 45 degrees off-axis), toed in by 45 degrees, axes criss-crossing in front of the sweet spot. The guy whose knee you see is sitting to the right of the center sweet-spot chair.

The first significant lateral reflection of the right-side (nearest) speaker is the long across-the-room bounce off the left-hand wall, and vice-versa. While the time delay is less than the ideal 20-25 seconds ballpark, it is still long enough to be beneficial. Because of time/intensity trading, you still get a decent soundstage from where the photo was taken.

 

[chips666]

Tmuikku wrote:



And Tangband asked:



Duke chimes in with a photo of speaker whose radiation pattern from the crossover region on up is 90 degrees (-6 dB @ 45 degrees off-axis), toed in by 45 degrees, axes criss-crossing in front of the sweet spot:

View attachment 220266

The first significant lateral reflection of the left-side speaker is the long across-the-room bounce of the right-hand wall, and vice-versa. While the time delay is less than the ideal 20-25 seconds ballpark, it is still long enough to be beneficial.

extra advantage is wider soundstage...

 

[Tangband]

Yes, the delay from the side wall reflection of about 20 to 25ms increases envelopment of the sound without much negative effect.


You also need the direction of the reflected sound. The delay on its own isn't enough, as far as I know.


It is interesting if this actually works this way since the reflection has its origin form the other speaker and is therefore mirrored?

--

To get an estimate of the room size I calculated the room width. I assume a standard stereo triangle (equal triangle, listening distance = width between L and R speaker).

Listening distance [m]	Room width (20ms delay) [m]	Room width (22.5ms delay) [m]	Room width (25ms delay) [m]
1	8.14	9.00	9.87
2	9.35	10.22	11.09
3	10.53	11.40	12.38
4	11.68	12.57	13.45

You can see that the room has to be bigger for this.

If you want to calculate it by yourself you can solve the following formula:
w=2*b+a
b=-a/(2*2^0.5)+(a^2/8+0.5*c*d*a+0.25*c^2*d^2)^0.5
Where w is the room width [m], b is the distance from the speaker to the nearest side wall [m], a is the listening distance aka length of the stereo triangle [m], c the speed of sound [m/s] and d the delay (s).

Since the best angle to create envelopment is about 60° and you can't get 25ms of delay with a standard side wall, you have to apply a trick by placing absorbers or diffusors at the side wall and use a "acoustic mirror" or a blank wall at about 60° from the listening position to get the desired 60° angle reflection.

You also can mimic these reflection with additional speaker with attenuation and delay. But you have to be caution since these speakers create new early reflections.

Interesting

Here is also some good information :

Early Reflections Are Not Beneficial

ethanwiner.com

 

[tmuikku]

Yeah, Geddes school of thought. Not sure which speakers are in the picture Duke posted but yes surely those look like theyd work fine in this kind of setup.

I guess I misread Tangbands post or it was edited, I wrote in #86 I dont know which speakers (have right coverage angle among other features) for ~20 square meter room, certainly something with waveguide, in my current opinio, but there is lot more to it. Geddes speaker system why not, makes a lot of sense. Here is a good post on the topic by mr Geddes. https://www.diyaudio.com/community/...d-geddes-for-stereo-setup.368188/post-6542945

What is often forgotten from discussion (on loudspeakers) is how hearing works and how to exploit that. I'm trying to figure it out and thought to participate on this active thread for more info on the subject.

 

[Duke]

One other approach is to just give up and build the smaller room like a studio - a lot of damping and no help to fill up the stereo system faults at all.

The disadvantage to relying on aggressive damping to eliminate the first lateral reflections is, that energy is now GONE and cannot contribute to beneficial later reflections. And the damping material continues to absorb EVERY reflection that strikes it, not just those initial ones. So in my opinion IF you choose to absorb the first lateral reflections, use broad-band absorption (so that you aren't degrading the spectral balance of the reflections) and only cover as much area as you need to. In my opinion re-directing or diffusing those first lateral reflections is generally superior to outright absorbing them, at least for home audio.

Yeah, Geddes school of thought. Not sure which speakers are in the picture Duke posted but yes surely those look like they'd work fine in this kind of setup.

Earl Geddes was my mentor, I assembled first-generation Summas for him before he offered kits. The speaker shown is one of my designs, and the photo is from a 2019 audio show.

 

[test1223]

Ascend Acoustics HTM 200 SE at 8', Monitor Audio Silver 8i at 8.5' and Ascend Acoustics Sierra 2EX AT 9.5'. All in very different locations. The 200s are close to the front wall, the Silver 8is are away from any walls, and the Sierras are in a corner. I've heard speakers from most major brands in all kinds of rooms from tiny apartments to quite big rooms, (I once had a studio that was 20x30' with 18' ceilings), to recording studios. Whenever I've auditioned speakers and then checked measurements afterwards, the ones I preferred have been the ones with flat smooth frequency response and wide even dispersion. Violins have never sounded right to me on narrow dispersion speakers.

Did I get it right, you tried different listening distances from about 3' to 9' with adjusting the stereo triangle accordingly and you liked with all the three speakers a listening distance of about 8' to 9' best? This would surprise me.

If the speaker where free standing and you "exclude" the bass performance from the judgement about 3.3' should be optimal for all three speakers. Did you hear a higher exactness similar to hearing with good headphones at about 3' listening distance? What made you choose the listening distance of 8' to 9' over closer distances?

 

[Byrdsmaniac]

Did I get it right, you tried different listening distances from about 3' to 9' with adjusting the stereo triangle accordingly and you liked with all the three speakers a listening distance of about 8' to 9' best? This would surprise me.

If the speaker where free standing and you "exclude" the bass performance from the judgement about 3.3' should be optimal for all three speakers. Did you hear a higher exactness similar to hearing with good headphones at about 3' listening distance? What made you choose the listening distance of 8' to 9' over closer distances?

Good headphone sound is certainly not what I want to hear from speakers. I have headphones for that. I want good concert hall sound. Pinpoint imaging is not something I've ever heard in a concert hall. As an example of a good sounding hall, I'd point to the Troy Savings Bank Music Hall in Troy NY. If you've never been, you should go. http://mediad.publicbroadcasting.ne...public/201604/tsbmhwithpiano-center-stage.jpg I saw Hilary Hahn in recital there a few years back It was heavenly. There are lot's of good halls for acoustic music all over the country. That's my reference for what I want from my speakers.
Also, I generally prefer rows 5 to 10 for chamber music and rows 10 to 20 for a full orchestra.

 

[tmuikku]

Yes, the delay from the side wall reflection of about 20 to 25ms increases envelopment of the sound without much negative effect.


You also need the direction of the reflected sound. The delay on its own isn't enough, as far as I know.


It is interesting if this actually works this way since the reflection has its origin form the other speaker and is therefore mirrored?

--

To get an estimate of the room size I calculated the room width. I assume a standard stereo triangle (equal triangle, listening distance = width between L and R speaker).

Listening distance [m]	Room width (20ms delay) [m]	Room width (22.5ms delay) [m]	Room width (25ms delay) [m]
1	8.14	9.00	9.87
2	9.35	10.22	11.09
3	10.53	11.40	12.38
4	11.68	12.57	13.45

You can see that the room has to be bigger for this.

If you want to calculate it by yourself you can solve the following formula:
w=2*b+a
b=-a/(2*2^0.5)+(a^2/8+0.5*c*d*a+0.25*c^2*d^2)^0.5
Where w is the room width [m], b is the distance from the speaker to the nearest side wall [m], a is the listening distance aka length of the stereo triangle [m], c the speed of sound [m/s] and d the delay (s).

Since the best angle to create envelopment is about 60° and you can't get 25ms of delay with a standard side wall, you have to apply a trick by placing absorbers or diffusors at the side wall and use a "acoustic mirror" or a blank wall at about 60° from the listening position to get the desired 60° angle reflection.

You also can mimic these reflection with additional speaker with attenuation and delay. But you have to be caution since these speakers create new early reflections.

Yeah its just math to figure out size of the space and position of things to get the delays sorted for nice results. Unfortunately the optimal stuff seems to be concert hall category, similar in size to our three bedroom detached house in total Such big listening space would be great, and accommodate few hundred other listeners as well. While its very interesting this is not the reality for most livingrooms, certainly not mine, so I dont spend time figuring it out other than with quick simple math and little bit of imagination Thanks for digging into it though!

One could check out actual living /listening room reflection paths / times with some raytracing tools or reflection calculators like this one https://amcoustics.com/tools/amray and try figure out how to make best of it. Its logical to me to try attenuate the first arriving ones with narrow directivity speaker toe-in, if they are considered audibly bad enough. They are there no matter what due to small room so better make best of the situation. Then, there are the other four early reflections from rest of cubicle room boundaries plus some second order reflections and perhaps few third order ones that arrive before 25ms in small rooms. It would be interesting to figure out whats there, what would be reasonable compromises, or set of outcomes to pick one suitable for a situation of small room listening setup. I mean, how to make best use of directivity.

Often discussion, for the sake of simplicity and keeping on topic, forgets the fact that loudspeaker systems and especially with combination of room and hearing system are full of compromises and things that matter and that don't matter as much. I mean In the end it is not possible to isolate imaging or spaciousness from other qualities of perceived sound in a loudspeaker system and for example wide coverage angle speaker doesnt play as loud as narrow one for one simplified example of compromises. Of course, its very interesting topic and good to have discussion in isolation so everyone please carry on.

 

[Duke]

Yeah its just math to figure out size of the space and position of things to get the delays sorted for nice results. Unfortunately the optimal stuff seems to be concert hall category, similar in size to our three bedroom detached house in total Such big listening space would be great...

Sub-optimal lateral reflection path lengths can still be beneficial. Ten milliseconds delay relative to the direct sound is the target Earl Geddes mentions, and my understanding is that Earl's source for the psychoacoustics behind that figure is researcher David Griesinger: "Transients are not corrupted by reflections if the room is large enough - and 10ms of reflections free time is enough." (This is the "short version" of the thinking behind that 10 milliseconds target.)

The cross-firing configuration you mentioned earlier will delay the first significant lateral reflections by about ten milliseconds (relative to the direct sound) in a 12-foot-wide room.