First and foremost, this is not a personal attack on Linkwitz. It is merely a discussion into alternate applications of phychoacoustic principals for maximum stereo fidelity. Notably with a focus of maximizing Stereo soundstage and imaging.
Linkwitz's conclusion is that stereo speakers can not accurately reproduce the soundstage of recorded material; it can only produce an "illusion" as long as reflections are copies of the direct sound (constant directivity designs). He states that speakers should optimally be placed to where first reflections are 6ms delayed or greater (based on his own experiments into the precedence effect). Below is a placement guide if you want to try this theory, with or without CD speakers, although CD speakers would be optimal.
Loudspeaker Placement in Small Rooms | Richard's Stuff
Doing some research into alternate approaches, I came across a website with completely different conclusions. These conclusions are marketed primarily for the creation of music, not reproduction, but can still be applied. A summary of conclusions as follows:
• Not all early reflections are bad -most are beneficial- but first reflections are bad, as they perceptually mask first reflections present in the recording.
• Treatment of first reflections makes it possible so you can accurately produce imaging that is the same as recorded. Treatment of the first reflection points of a room allows you to hear the first reflections in a recording.
• Untreated first reflection points cause comb filtering/ time smearing.
• To maximize precedence effect, first reflection points less than 15ms must be treated to attenuate 1khz-8khz by at least 10dB (This is incredibly easy).
• Hearing of the originally recorded acoustic space is maximized by ITDG/ISD maximization.
• This is done by having your rooms ISD gap larger than the recorded ITDG gap. World class concert halls have an ITDG gap of 12-25ms. Obviously recordings made with smaller rooms will have smaller values. To aching a ISD gap of 20ms, the direct sound must travel 22.5 feet before hitting your ears again. This means you must sit 11.25 feet from the back wall, or use absorption/diffusion.
http://arqen.com/acoustics-101/reflection-free-zone/
To DIY sound absorbers, 4" of Roxul Rockboard 40 will absorb all sound from 125hz to 4khz. What happens to sound above and below isn't known. I'd imagine all sounds above 4khz are absorbed, while attenuation below 125hz will decrease with frequency.
Linkwitz's conclusion is that stereo speakers can not accurately reproduce the soundstage of recorded material; it can only produce an "illusion" as long as reflections are copies of the direct sound (constant directivity designs). He states that speakers should optimally be placed to where first reflections are 6ms delayed or greater (based on his own experiments into the precedence effect). Below is a placement guide if you want to try this theory, with or without CD speakers, although CD speakers would be optimal.
Loudspeaker Placement in Small Rooms | Richard's Stuff
Doing some research into alternate approaches, I came across a website with completely different conclusions. These conclusions are marketed primarily for the creation of music, not reproduction, but can still be applied. A summary of conclusions as follows:
• Not all early reflections are bad -most are beneficial- but first reflections are bad, as they perceptually mask first reflections present in the recording.
• Treatment of first reflections makes it possible so you can accurately produce imaging that is the same as recorded. Treatment of the first reflection points of a room allows you to hear the first reflections in a recording.
• Untreated first reflection points cause comb filtering/ time smearing.
• To maximize precedence effect, first reflection points less than 15ms must be treated to attenuate 1khz-8khz by at least 10dB (This is incredibly easy).
• Hearing of the originally recorded acoustic space is maximized by ITDG/ISD maximization.
• This is done by having your rooms ISD gap larger than the recorded ITDG gap. World class concert halls have an ITDG gap of 12-25ms. Obviously recordings made with smaller rooms will have smaller values. To aching a ISD gap of 20ms, the direct sound must travel 22.5 feet before hitting your ears again. This means you must sit 11.25 feet from the back wall, or use absorption/diffusion.
http://arqen.com/acoustics-101/reflection-free-zone/
To DIY sound absorbers, 4" of Roxul Rockboard 40 will absorb all sound from 125hz to 4khz. What happens to sound above and below isn't known. I'd imagine all sounds above 4khz are absorbed, while attenuation below 125hz will decrease with frequency.
Last edited:
His speakers give a very diffuse sound field. By treating the first reflection, IME, the sound field becomes more focused. It all comes down to personal preference. While I find Linkwitz's speakers pleasant to listen to, between the diffuse sound field and rolled off top end, they're not my cup of tea.
This is why RFZ is now so popular in studio acoustic (Reflexion Free Zone).
I would not say that attenuation in the band you give is a panacea nor incredibly easy. If that was the case we probably all be doing it in our living room.
Yes but what do you do about side wall reflections? And floor and ceiling bounce?
Ok try to google about LEDE (live end dead end) to know how it's applyed in 'real life' and to know about Haas effect (the itd rely heavily on Haas effect.
Try to google CID (controlled image design), an other way to control IR devellopped in the 90's by BBC researcher B.Walker (there is many white paper about that one on BBC archives).
A lot of theory but no actual blind listening tests supporting it.
Ears are everything.
Are there blind listening tests on this subject?
I think what Face means is that the direct/reverberation ratio of omnipolar and dipoles aren't for him. I don't believe he was talking about room modes.
For example, a true omni design will have the lowest direct sound/ reverberation sound ratio, followed by a CD dipole, followed by a true CD horn. The horn has the tightest beam of sound, so less energy "fills" the room with reverberation.
Designs with lower direct sound / reverberated sound ratio will sound "farther", with a higher ratio sounding closer. I believe this is why Linkwitz likes to set up his monopole design more near field (to increase the sound intensity difference between direct and reverberated sound) than his dipolar designs.
The ears behind practically all modern recorded music is done in this way during the mixing and mastering stages
.To be honest though, I don't know of any blind testing done to compare with and without absorption at first reflection points- but I believe that is because the effect is so apparent, blind testing isn't needed to show it's a huge difference. Now what would be preferred under a blind test is a good question.
Absorption at first reflection points has a good consensus of "tightening" the soundstage (making it smaller) but improving pin-point imaging. While diffusion at first reflection points (notably you should only be using phase-coherent diffusers here, which are poly diffusers) creates a wider soundstage with less pin-point imaging. It really comes up to personal preferences.
Both are generally found to be improvements over no treatment at first reflection points.
Here is a quick, easy, and cheap DIY for poly diffusers for anyone who wants to try them out.
https://www.lifewire.com/make-your-own-audio-diffuser-3134903
But that has nothing to do with diffuseness. The ratio of direct to reverberant sound is also too simplistic since it does not include frequency - and also therefore the power response of the speaker. Here Linkwitz scores highly and whenever a speaker scores highly our perception is able to extract the original information regardless of the treatments that make the acoustics better or worse. But acoustic treatments are never a fix for poor speakers.
I find reduced early reflections help make the imaging more as intended, but the image can also be wide (wider than the room in some recordings). When played in mono, the image can be very compact.
A quick and easy DIY sound absorber that will fulfill this requirement can be made from "Knauf Ecose" insulation board, and you want the 2" & 6LB variant. If you google it, you can see that they have distributors across the world, so it should be easy to get anywhere. They come in 24in x 36in sheets, which is the perfect size to wrap with used burlap coffee bags for an artistic look
. To hang them, use rotofast snap-on anchors, which are less than $2/piece. Link to Knauf Ecose: Knauf ECOSE® Insulation Board, 2 Inch, 6 lb (Single Pc)
Rotofast: Rotofast Panel Anchors - Installation Instructions
The NRC for the Knauf Ecose 2" 6LB is:
• 125Hz: .32 • 250hz .81 • 500hz 1.08 • 1000hz 1.06 • 2000hz 1.03 • 4000hz 1.04
This means that they absorb 100% of the sound from 500hz and up
. They still do a pretty damn good job down to 250hz too! Especially for 2 inches of product.--------------------------
I believe the side walls and roof aren't problems because they are already treated at first reflection points. For some reason, the pro guys actually prefer a reflective floor.
From my own experiments and experience, I can only second the use of diffusors to treat first reflections.
Absorbers will take the life out of the music.
I have tried both in my room, and it is such a difference.
My room is difficult as the speakers have to be very close to the side walls which of course only makes matters worse.
Overall I think a mix of moderate absorption, especially in the corners and as many diffusors as you can stomache will deliver the best sound.
DIY Diffusor | Baldin's Blog
Building a Home Theater and Listening Room
Absorbers will take the life out of the music.
I have tried both in my room, and it is such a difference.
My room is difficult as the speakers have to be very close to the side walls which of course only makes matters worse.
Overall I think a mix of moderate absorption, especially in the corners and as many diffusors as you can stomache will deliver the best sound.
DIY Diffusor | Baldin's Blog
Building a Home Theater and Listening Room
Assuming Linkwitz assumption is true that aslong as CD speakers are used and the reflected sound are attenuated copies of the direct sound, the first reflection points would still benefit from phase-coherent diffusion, otherwise untreated first reflection points will cause comb filtering. The diffusion will not selectively attenuate frequencies, like absorbers do.
Now, 99.999% of speakers out there are not constant directivity from 20hz to 20khz. Not even Geddes'. So reflected sounds are not attenuated copies of the direct sound, if we are strict.
Additionally, no untreated room reflection point is 100% reflective, and different materials for reflection points will impart their own sonic imprint to the reflected sound. Have you ever been in a: marble, stone, wooden, normal, glass room? I'm sure you know what I mean in that they are all different sounding.
One other thing I think about is if frequency-dependent absorption makes it harder to "localize" and discern direction because reflections are not attenuated copies, why do we not notice this with common furniture and talking to each other? If Linkwitz was right, any absorptive furniture would make it noticeably more difficult to tell where a person is when they are talking. I don't know about you, but I have never observed this. I have observed it being extremely difficult to be able to understand what someone is saying in, extremely reflective, minimally dampened rooms though. Something to think about.
Last edited:
.
Not a blind test but there is a report from (a panel of manager of BBC crew at the time) the first CID room developped in the BBC paper i talked about.
The whole concept is to make a RFZ from 1khz up with ITD of 19ms and -20db first wave of reflections and to lesser the overall need for absorbsion. Walker relaxed on the attenuation after listening to the results eventually.
Yes i know.
In europe we usually use Rockwool for this purpose and as far as i remember this have approximately same characteristics: Absorb nearly everything past 500hz. For lower frequency, make it a panel and give some room from wall (1 to 3 inch) and it will work further down... Classic absorber design.
I'm not saying it won't work, what i would say is that 'good' room acoustic does not rely only on absorbtion.
It's a mix a absorbtion/diffraction/diffusion which work great in general (T. Hidley 'Non Environnment' design or G. Massenburg's 'Black Bird' multichannel room are different -and working- point of view).
I would not say that: side walls are usually angled to give an rfz (sorry to repeat but...
) and are treated you are right. Roof are nearly almost 'cathedral ceiling' (even if you don't see them to be, usually control room are much, much bigger than what you actually see once in the room) and usually the whole visible ceiling is a kind of absorber (using suspended rockwool tiles). You are right when you say it's treated to give diffusion at listening point eiter using 'clouds' or other things (diffusor tiles).
The floor is almost always reflective for practical purpose, and because of compromise to be done. In Non Environment rooms, the whole front of the room and the floor is made reflective to have some sort of liveness when you are inside the room (for users, musicians where annoyed in the first iterations of the concept and disliked the rooms at first for this reason). Otherwise it would be to oppressive for long sessions.
Last edited:
Common practice doesn't mean its actually beneficial. For that you need blind listening tests. Ears are everything, theory is just fun.The ears behind practically all modern recorded music is done in this way during the mixing and mastering stages
I've been looking. Here's a test: AES E-Library The Practical Effects of Lateral Energy in Critical Listening Environments
It turns out humans can adapt very fast to changing acoustics. If you think about it, if we couldn't we would be in serious trouble fast when hunting pray/escaping predators.
If we want accurate reproduction, then personal preference is useless. Its accurate or its not. To test it we need actual blind listening tests, not theory or anecdotal evidence from people who just spend a lot of money on something.
I wouldn't say it's that easy to get the situation as outlined in the first post.
You'd need a way to check what you've got too.
I did choose that model as a guide for my personal setup. I started with a design that has less floor and ceiling reflections and treated all first reflection points with damping panels.
It got me this result (after including FIR processing):
This picture is without the delayed reflections (my room has no means to get that in a passive fashion, it simply isn't big enough).
If we look at output at the listening position:
This will give the impression that all reflections after the main peak are down by more than 20 dB. But that isn't totally true. The Filtered IR graph is showing a rendered picture of the total output. Once you start looking at smaller slices there will be frequencies that are louder than -20 dB.
As a spectrogram shows with a 15 dB range:
If I set that to 10 dB (as outlined in the proposal):
Now that's pretty clean. See the differences between the filtered IR's -20 dB or more and the spectrogram showing a 15 dB range? Compare it to the APL_TDA plot to "see" the room reflections.
My next project was to add (diffuse lateral) late reflections. In the filtered IR that looks like this:
That part is purely synthetic, created by extra speakers behind the listening position that don't fire directly to the listening position.
This brings back the "life" into the music that was robbed with absorption. It results in a more 3D like imaging. You don't hear the extra speakers because they are within the Haas limit. But they do change perception in a very positive way.
Now why would you say:
I do not agree with that at all. Be sure a floor reflection creates combing problems just as other walls do.
You'd need a way to check what you've got too.
I did choose that model as a guide for my personal setup. I started with a design that has less floor and ceiling reflections and treated all first reflection points with damping panels.
It got me this result (after including FIR processing):
This picture is without the delayed reflections (my room has no means to get that in a passive fashion, it simply isn't big enough).
If we look at output at the listening position:
This will give the impression that all reflections after the main peak are down by more than 20 dB. But that isn't totally true. The Filtered IR graph is showing a rendered picture of the total output. Once you start looking at smaller slices there will be frequencies that are louder than -20 dB.
As a spectrogram shows with a 15 dB range:
If I set that to 10 dB (as outlined in the proposal):
Now that's pretty clean. See the differences between the filtered IR's -20 dB or more and the spectrogram showing a 15 dB range? Compare it to the APL_TDA plot to "see" the room reflections.
My next project was to add (diffuse lateral) late reflections. In the filtered IR that looks like this:
That part is purely synthetic, created by extra speakers behind the listening position that don't fire directly to the listening position.
This brings back the "life" into the music that was robbed with absorption. It results in a more 3D like imaging. You don't hear the extra speakers because they are within the Haas limit. But they do change perception in a very positive way.
Now why would you say:
I do not agree with that at all. Be sure a floor reflection creates combing problems just as other walls do.
Last edited:
I disagree. If a waveguide worked strictly to its angle and was placed in a corner, it could fill the room thoroughly (OK, floor and ceiling are more tricky). As the room is filled by this point source the reverberant field should be as diffuse as the room would allow without treatment. The difference is the energy that would be invested in early reflections stays coherent with the main wavefront instead.
The reverberant energy is therefore as spread out as it can be. The positive to that is that it will be less distinct and more delayed, but it will also be more quiet. If this energy was invested in early reflections then you'd get more pressure and for a shorter time or over a smaller area with a given amount of power. It is a choice whether to include these early reflections or to keep the reverberant field in the background.
Few reflections create comb filtering effects, it doesn't matter where they come from.
Lots of diffuse and frequency independent reflections don't.
Simple listening test: Add a delay of a few msec to a sound vs adding reverb to a sound. Huge difference. I can make some soundfiles if people are interested.
And you can add a very large amount of reverb to a sound before you can't pin point its location anymore. Edit: I can make some soundfiles if people are interested.
Lots of diffuse and frequency independent reflections don't.
Simple listening test: Add a delay of a few msec to a sound vs adding reverb to a sound. Huge difference. I can make some soundfiles if people are interested.
And you can add a very large amount of reverb to a sound before you can't pin point its location anymore. Edit: I can make some soundfiles if people are interested.
Well, i see your point but reverbs are a multitude of delay summed together.... I would have said 'discrete reflections'.
Even if i tend to agree, in acoustic usually too much diffuse field (for reproduction) is not seen as a good thing as it tends to unfocus/mess the stereo image.
But Massenburg seems to had great results in the room i talked before (i've never heard the room but really would like to!).
http://2steelgirls.com/wp-content/uploads/2014/07/10556446_730667506993862_7682785253316360452_n.jpg
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.