• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Dr. Toole's comments in anticipation to his new book on Acoustics and Psychoacoustics

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,056
Here is a better explanation about why room curves are not using flat as a target. From here:


Both Dr. Olive's initial speaker measurements, the JBL Synthesis measurements, and the B&K curve agree that high-quality speakers measuring flat in an anechoic chamber tend to have a ~1dB/octave "room gain" curve when placed in a good-sounding room with no equalization.
Read more at https://www.innerfidelity.com/conte...man-listener-target-curve#lXl6fEI9EQkrbvgp.99

Notice measuring flat in an anechoic chamber results in the room measurement having a 1 db/octave gain in response as frequency ascends. So speakers should have flat and smooth response. If you have to measure them indoors your measurements is effected.
 

Jakob1863

Addicted to Fun and Learning
Joined
Jul 21, 2016
Messages
573
Likes
155
Location
Germany
Could someone please define "linear"?
The term "linearity" in systemtheory means that a system responds to a change of the input signal with a change of the output signal with a fixed ratio, delta y / delta x = const. ( y is the output signal, x is the input signal).
A LTI system is a L inear T ime I nvariant system, when the linearity is given and the superposition principle holds and the system reacts in the same manner regardless of the time (input change today or tomorrow does not matter).

Two different types of distortion are possible
-) linear distortions, which means that the above mentioned delta y/delta x = const. condition is violated
-) nonlinear distortions, which means the output signal contains frequencies that weren´t present in the input signal

Therefore a perfectly linear system can´t have harmonic distortions, therefore no intermodulation distortions either and of course no linear distortion.

So i can´t agree with amirm´s post above; a room introduces usually _linear_ distortions.
 
Last edited:

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
Flat response is not preferred.
I think there is some confusion about this. The way I have come to think about it is this:

Your speaker is not ideal, in that it probably doesn't have uniform dispersion at all frequencies. In an anechoic chamber this wouldn't matter because you would only hear the direct sound. So as long as recordings were created in order to be heard flat*, a flat on-axis response would sound correct.

In a real room, you hear a mixture of direct and reflected sound. The reflected sound has a different frequency response than it would if your speaker was ideal. It is quite likely that this means that it will sound 'wrong' in terms of the overall frequency response. The effect will vary depending on the furnishings in the room, speaker placement, etc. You therefore need to fudge some compensation to the direct frequency response in order to make the overall sound subjectively 'about right'. Most people here think that this is via a simple measured in-room target curve (I bet the old speaker manufacturers feel a bit stupid for going to the expense of building anechoic chambers! Just slap some drivers in a box and EQ it until it measures flat in the listening room...), but they find, mysteriously, that it doesn't sound right when it is flat. So they 'turn it down' a bit and find that it is less offensive to the ears.

I would say that this simple in-room measurement technique contravenes the implications of Dr. Toole's observation that we largely "hear through the room". It seems to me that the most likely optimal result is going to come from minimal, smooth EQ, that does just enough to compensate by about the right amount - it is a fudge, after all. Rather like the baffle step compensation curve that might be present in many speakers, anyway, and which was known long before DSP was invented...

* Circle of confusion I hear you cry. Well, not if the recording person is using near field monitoring with flat speakers, and not if the recording is a purist recording where the mic effectively feeds straight onto the 'tape' without any fiddling about.
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,202
Likes
16,982
Location
Riverview FL
In a real room, you hear a mixture of direct and reflected sound.

Looking at my most recent measurement set impulse response:

My highest reflection SPL in the first 25ms from the room is 29dB below the impulse with the panel speakers, but only -16dB with the little JBLs.

If considered "distortion", those numbers would be something like 3.5% and 15.8%.

---

Crude Experiment:

I have the TV playing through its speakers about 29dB below the music level right now. The TV (talk) is all but inaudible behind the music, though just intelligible with the music stopped.

Raising it to 16dB below the music, it becomes an annoying but unintelligible noise (though clearly intelligible when the music is stopped).

---

Disclaimer: I have not festooned my room with sidewall nor ceiling absorbers.
 

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
Looking at my most recent measurement set impulse response:

My highest reflection SPL in the first 25ms from the room is 29dB below the impulse with the panel speakers, but only -16dB with the little JBLs.

If considered "distortion", those numbers would be something like 3.5% and 15.8%.

---

Crude Experiment:

I have the TV playing through its speakers about 29dB below the music level right now. The TV (talk) is all but inaudible behind the music, though just intelligible with the music stopped.

Raising it to 16dB below the music, it becomes an annoying but unintelligible noise (though clearly intelligible when the music is stopped).

---

Disclaimer: I have not festooned my room with sidewall nor ceiling absorbers.
Am I interpreting this page correctly?
http://www.linkwitzlab.com/rooms.htm

Mr. Linkwitz seems to be saying that the reverberated sound in total is between 2dB and 8dB louder than the direct when you're 2.4m away from a (dipole) speaker?
 

The Smokester

Active Member
Joined
May 11, 2017
Messages
136
Likes
39
Location
SF Bay

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
I can't see exactly what you are referring to...I like this paper which is a little more self-contained:

http://www.linkwitzlab.com/The_Magic/The_Magic.htm

Yes, in this article he says that direct and reverberant levels are equal when you sit at 0.8m (monopole speaker) and 1.39m (dipole), so I guess it looks consistent with what I said above (which I had gleaned from the table in the middle of the other article).
 

The Smokester

Active Member
Joined
May 11, 2017
Messages
136
Likes
39
Location
SF Bay
Linear means total absence of distortion.

Of note, room impact on the sound of the speakers is 100% linear (sans rattles and such). In other words, the room does not add any distortion products.

In the context here, we are talking about the shape of the frequency response. That can be flat, smooth or both. Listening tests show that what is important is smoothness (i.e. lack of variations). Flat response is not preferred. And at any rate, being flat has nothing to do with a system being linear.

Okay...I will pick on you since your answer was so clear and short.

For clarity, let's say the output, A, of an audio system is characterized by A(f) = k a(f), where a is the input amplitude and f is frequency. Seems like everyone would agree that that is a linear system. No distortion.

Lets expand this to A(f) = k(f) a(f). The k now changes with frequency. This is just a sneaky way of introducing a non-linearity. So the shape of A, the sum of complex Fourier components, will be changed relative to a. So, isn't it distorted? Are we still calling this a linear system?

The reason I'm asking all these dumb questions is because I'm continually confused about what people are saying and am trying to understand.
 
Last edited:

The Smokester

Active Member
Joined
May 11, 2017
Messages
136
Likes
39
Location
SF Bay
Yes, in this article he says that direct and reverberant levels are equal when you sit at 0.8m (monopole speaker) and 1.39m (dipole), so I guess it looks consistent with what I said above (which I had gleaned from the table in the middle of the other article).

Have you heard Linkwitz's Orion speakers? I find them pretty impressive.
 

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
Have you heard Linkwitz's Orion speakers? I find them pretty impressive.
No I haven't. I have to admit to a bias against open baffle speakers. How do they sound to you? Any evidence of 'phasiness'?
 

The Smokester

Active Member
Joined
May 11, 2017
Messages
136
Likes
39
Location
SF Bay
No I haven't. I have to admit to a bias against open baffle speakers. How do they sound to you? Any evidence of 'phasiness'?

What do you mean by "phasey"? I must admit they are very "suspicious" looking. But, they rival the best I've heard in the under $50k category. As I resurrect my system they will be contenders. Depends on what room I end up with to put them in.

There is listening program where interested parties can go to a volunteer host's house and audition them. Linkwitz and his wife are gracious hosts.
 

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
What do you mean by "phasey"? I must admit they are very "suspicious" looking. But, they rival the best I've heard in the under $50k category. As I resurrect my system they will be contenders. Depends on what room I end up with to put them in.
Well, I'll try and audition them some time. But it just doesn't 'sit right' with me... If you'd never heard of these and were asked what the ideal speaker should be, I think you would probably come up with something like a Kii Three: small, controlled, uniform dispersion or maybe omnidirectional in some applications..? But would you think to say "Actually, I think it will sound better if we spray the inverse of what's coming out at the front out of the back"?

Why have the major speaker manufacturers not cottoned onto the fact that they could make brilliant speakers so much more easily and cheaply if they ditch the box? I have yet to be convinced by them...
 

The Smokester

Active Member
Joined
May 11, 2017
Messages
136
Likes
39
Location
SF Bay
...But would you think to say "Actually, I think it will sound better if we spray the inverse of what's coming out at the front out of the back"?

Why have the major speaker manufacturers not cottoned onto the fact that they could make brilliant speakers so much more easily and cheaply if they ditch the box? I have yet to be convinced by them...

Yes. As I say, they are suspicious looking because they break stereotypes. We are accustomed to using a box to enrich a speakers capabilities, and adapting speaker position to the room. Some of Linkwitz's reasoning is in the above-referenced paper. Also, I recall him saying that bass woofers with sufficiently long movement didn't (reasonably) exist until the early 21st century. And note there is a DSP for each individual speaker.

Also, note that they need to be well out into the room. Linkwitz's livingroom has a wall of glass on one side (left) which helps dissipate the low frequency reflections. But, I have a friend who drank the coolaid years ago and his Orion system also sounds terrific.
 

Jakob1863

Addicted to Fun and Learning
Joined
Jul 21, 2016
Messages
573
Likes
155
Location
Germany
Yes, in this article he says that direct and reverberant levels are equal when you sit at 0.8m (monopole speaker) and 1.39m (dipole), so I guess it looks consistent with what I said above (which I had gleaned from the table in the middle of the other article).

That is the "reverberation radius" or "critical distance" defined as the distance from the sound source (loudspeaker) where direct sound and reverb field are of equal loudness/level. It depends on the fact that the direct sound drops at a constant rate with distance while in theory the diffuse sound field remains constant.

In real usually quite small listening rooms the diffuse field isn´t really diffuse and so the premise of constant level doesn´t hold and the calculation of the critical distance gets a bit more complicated.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,389
Location
Seattle Area
Okay...I will pick on you since your answer was so clear and short.

For clarity, let's say the output, A, of an audio system is characterized by A(f) = k a(f), where a is the input amplitude and f is frequency. Seems like everyone would agree that that is a linear system. No distortion.

Lets expand this to A(f) = k(f) a(f). The k now changes with frequency. This is just a sneaky way of introducing a non-linearity. So the shape of A, the sum of complex Fourier components, will be changed relative to a. So, isn't it distorted? Are we still calling this a linear system?

The reason I'm asking all these dumb questions is because I'm continually confused about what people are saying and am trying to understand.
Ah, you went from asking me for a simple explanation to a much more complex one. :)

The answer is yes, your system is still linear. Its transfer function of input values to output is a straight line with the slope of k(f)*a. Linear system only need to be linear within a single frequency. It is OK if the gain (i.e. slope of input to output) changes if the frequency changes.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
I always hate these linearity discussions and how people fold them into distortion. It is something I do not claim to have reconciled myself.

If the response is linear, then it's a linear system, but speakers get tricky. If it is linear, there should be no distortion, but in fact if the output varies with frequency due to the room or whatever then the overall output no longer matches the input and thus we can say it is distorted. What is particularly vexing is that the lowest-distortion speaker in the world in the wrong place in a room can present a very distorted/modified frequency response to the listener. So did the room distort the output? In one sense no; it did not add (significantly) spurs or signals other than what came into the system. However, if you look across the audio band, the frequency response is no longer flat, so it has distorted the sound we hear. That would be Smokester's "linear distortion". Blah. Glad I work with simple GHz stuff... :)
 

The Smokester

Active Member
Joined
May 11, 2017
Messages
136
Likes
39
Location
SF Bay
Ah, you went from asking me for a simple explanation to a much more complex one. :)

The answer is yes, your system is still linear. Its transfer function of input values to output is a straight line with the slope of k(f)*a. Linear system only need to be linear within a single frequency. It is OK if the gain (i.e. slope of input to output) changes if the frequency changes.

Didn't mean to sandbag you. Guess that's just for bonus points. :D

I always hate these linearity discussions and how people fold them into distortion. It is something I do not claim to have reconciled myself.

If the response is linear, then it's a linear system, but speakers get tricky. If it is linear, there should be no distortion, but in fact if the output varies with frequency due to the room or whatever then the overall output no longer matches the input and thus we can say it is distorted. What is particularly vexing is that the lowest-distortion speaker in the world in the wrong place in a room can present a very distorted/modified frequency response to the listener. So did the room distort the output? In one sense no; it did not add (significantly) spurs or signals other than what came into the system. However, if you look across the audio band, the frequency response is no longer flat, so it has distorted the sound we hear. That would be Smokester's "linear distortion". Blah. Glad I work with simple GHz stuff... :)

Yes. I guess it is indeed called linear distortion. But it makes it confusing and probably originally coined by someone confused. Like I said, it's a sneaky way to add nonlinearity. So it's not really a linear relation anymore which was the source of my confusion and probably the source of your dislike.
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,202
Likes
16,982
Location
Riverview FL
Room Reverberation Experiment:

Mono signal applied to both (stereo) speakers, recorded waveform.

4 ms of 1000Hz and then 4 ms of 3000Hz, then about 80 ms of room contribution only.

Plots: Signal, JBL LSR 308, MartinLogan reQuest (dipole)

upload_2017-6-19_17-58-54.png


Attached: signal, JBL, reQuest. Drag into Audacity and play, try slower playback rates, and repeat play.
 

Attachments

  • Reverb Experiment.zip
    55.8 KB · Views: 115
Last edited:

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
Room Reverberation Experiment:

Mono signal applied to both (stereo) speakers, recorded waveform.

4 ms of 1000Hz and then 4 ms of 3000Hz, then about 80 ms of room contribution only.

Plots: Signal, JBL LSR 308, MartinLogan reQuest (dipole)

View attachment 7578

Attached: signal, JBL, reQuest. Drag into Audacity and play, try slower playback rates, and repeat play.
And if you were to keep a signal going for a second or two in order to let the reverb 'build up', and then stopped the signal..?
 

Brad

Active Member
Joined
Nov 8, 2016
Messages
114
Likes
35
I think the difficulty is that the numbers in your A(f)=k(f) a(f) are complex.
As shown above, the direct sound can still obey A(f)=k a(f), but integrated over some longer time interval you measure k(f)
 
Top Bottom