Importance of impulse response

[thewas]

that why there is jumper option for both brand for other problem occurs. I don't own both speaker, I can't tell how they perform in real life.

Both mentioned brands don't have jumpers for phase variations:

https://www.focal.com/sites/www.focal.fr/files/shared/catalog/document/notice_grandeutopiaemevo_web_0.pdf

https://www.wilsonaudio.com/media/917/SabrinaX-Owners-Manual-Web.pdf

 

[MAB]

that why there is jumper option for both brand for other problem occurs. I don't own both speaker, I can't tell how they perform in real life.

Yeah, but the tweeter arrangement and movements were explained to you, with examples. It can't be moved in a meaningful way to adjust for delay. Delay adjustments are not possible here.

 

[fineMen]

Hahaa, then I'm very happy to quote Audibility of Loudspeaker Group-Delay Characteristics by Liski, Mäkivirta and Välimäki once again. I can continue as long as you.

View attachment 247891



Your turn to repeat your dogma...

You live or die with the dogma, that jumping out of a window will hurt you. Look, I take it with humour.

And here it comes again, that "investigation" that was founded on a misunderstanding of some real literature, given in the article as a reference. Nobody could tell why these guys used a 'time inversion' as to 'strengthen' the effect. Except that people said: "Can't you read? They say so! Stupid."

In contrast to many others I read the original article and additionally that reference. And I'm so arrogant to say that I've got an understanding of what the authors wanted to convey, and I even understand the reference perfectly.

So, keep on continuing to present an open question: "why these guys used a 'time inversion' as to 'strengthen' the effect" For info, what's your schedule with reiterating the asking?

 

[Thomas_A]

Not sure why this is discussed. The audibility of phase shift in traditional loudspeaker crossovers is zero if properly executed above 1 kHz or below 100 Hz. Which should cover many speakers. Messing with phase in the ”voice” region may be audible depending on the symmetry of the signal.

 

[fineMen]

As it is shown in that link just the angle can be changed, not the distance.

Referring to this: https://www.focal.com/en/focal-teach/focus-time

Angle changes distance; it does. Of course the technical design is hilarious in my eyes. That little crank, you can't make that up, really, really?
But anyway, angle changes distance. Won't teach you, just motivate to reconsider the case (after the LoL at the "concept").

 

[fineMen]

Anyway for really perfect phase alignment an active FIR crossover is needed, also unless the design is concentric you will only have perfect alignment just at one point but the off-axis reflected sound won't be.

So many misunderstandings. That "phase alignment" may be understood as zero group delay. But of course it is not the same! Phase alignment as such is straight perfect with e/g an infinite impulse response (IIR, means: it "rings") Linkwitz/Riley cross-over.
"Concentric" is not key, but that the spacing of the radiating sources is very small compared to the wavelength at x/o frequency. It has nothing to do with "reflected sound". The effect is known as 'lobing'.

 

[Plcamp]

Not sure why this is discussed. The audibility of phase shift in traditional loudspeaker crossovers is zero if properly executed above 1 kHz or below 100 Hz. Which should cover many speakers. Messing with phase in the ”voice” region may be audible depending on the symmetry of the signal.

I need to understand this…it implies a need to avoid crossovers in the range 100-1khz?

Not sure that’s what you meant …would appreciate if you could elaborate a bit?

 

[KSTR]

That "phase alignment" may be understood as zero group delay. But of course it is not the same! Phase alignment as such is straight perfect with e/g an infinite impulse response (IIR, means: it "rings") Linkwitz/Riley cross-over.

Good point! I prefer to use the term "constant phase offset" between ways for this as I found that to be of great importance for coherence of the sound, even when the total (summed) phase response is not linear phase. That is, constantness of whatever phase offset is between drivers (unwrapped phase, of course) makes a speaker sound coherent. That is to say, a minimum phase XO with precisely tracking phases tends to sound better than a linear phase XO with bad phase tracking. Once you have phase tracking, you can apply FIR-based phase unwrapping on the result, being the icing on the cake (preferably using an analytically derived correction kernel to avoid artifacts).

Ideally, I found zero is the best best constant phase offset (as achieved by proper LR-type crossovers, acoustical responses of course, not electrical network alone). This creates the desired characteristic of blips (short shaped envelope sine bursts) looking exactly the same in the waveform for both ways, only amplitude differs, depending on distance to XO point. When we have several XO points spaced closer than a few octaves apart phase contributions from one XO point affect the other and vice versa which means a LR amplitude target will not lead to constant zero phase offset. Same goes for the final roll-off (system's lower corner frequency). Therefore I usually design from the phase target backwards which yields a sum that may not be flat but can be fixed with overall global EQ.

But others like 90deg (classic Butterworth) also work best with precise constantness down to -40dB if possible. I often like to use 60deg constant offset (a hybrid between LR and Butterworth) as it has good overall properties like final 3rd(or 5th) order slopes but a 2nd(4th) order'ish knee and a 2nd(4th) order total phase response and little lobing compromise assumed spacing can be kept low (or when its a coaxial or D'Appolito-style speaker).

 

[thewas]

Referring to this: https://www.focal.com/en/focal-teach/focus-time

Angle changes distance; it does. Of course the technical design is hilarious in my eyes. That little crank, you can't make that up, really, really?
But anyway, angle changes distance. Won't teach you, just motivate to reconsider the case (after the LoL at the "concept").

Yes, see what I wrote above:

Yes, but as said this distance change is very small and cannot really make passive multiway loudspeakers perfectly time aligned.

 

[fineMen]

I need to understand this…it implies a need to avoid crossovers in the range 100-1khz?

Not sure that’s what you meant …would appreciate if you could elaborate a bit?

I can't imagine that you would take advise from my side. Settled science ( for hearing aids and for the luxury product of dead 'perfect' gear to replay not-so perfect, the least not ideal recordings for bored people ... ), yep, these guys and gals tell, that in the region you mention the => group delay (*) shall not exceed 2ms .. 1ms. Then, as the saying goes, the human hearing apparatus simply cannot tell. It is as with vision. Some supermen see, or at least hear ultra-violet or infra-red light, but most of us lucky normal individuals can't. Simple. Pure guys who can see sound ...

If that news wasn't good enough, we already know, that it doesn't need much to achieve such a goal. Even careless speaker designers (as me) hardly make it only close to those limits. I mean, from below. My current project realizes some 1ms @ 350Hz, < 0,2ms @2kHz with tons of steep and vicious cross-over plus mad equalization on the way.

(*) as the only reasonable measure for frequency dependent phase

 

[thewas]

So many misunderstandings. That "phase alignment" may be understood as zero group delay. But of course it is not the same! Phase alignment as such is straight perfect with e/g an infinite impulse response (IIR, means: it "rings") Linkwitz/Riley cross-over.

Correct, my quoted sentence was too short and could lead to misunderstandings.

"Concentric" is not key, but that the spacing of the radiating sources is very small compared to the wavelength at x/o frequency.

This is with what is meant with "concentric" and the reason why I didn't write "coaxial".

It has nothing to do with "reflected sound".

It has, with a non-"concentric" design you can only align one direction and thus reflections from other directions won't be as aligned.

 

[Thomas_A]

I need to understand this…it implies a need to avoid crossovers in the range 100-1khz?

Not sure that’s what you meant …would appreciate if you could elaborate a bit?

Some asymmetric signals can be audible when you flip polarity or shift phase in bands that cause a symmetry change. Not everone appears to hear it though.

 

[Holmz]

So many misunderstandings. That "phase alignment" may be understood as zero group delay. But of course it is not the same! Phase alignment as such is straight perfect with e/g an infinite impulse response (IIR, means: it "rings") Linkwitz/Riley cross-over.
"Concentric" is not key, but that the spacing of the radiating sources is very small compared to the wavelength at x/o frequency. It has nothing to do with "reflected sound". The effect is known as 'lobing'.

Which means that two drivers physically close together, and the time offset corrected digitally (and maybe phase), will have less of lobing issue and they appear more concentric?
(Lobing is primarily a physical separation distance deal.)

 

[fineMen]

Which means that two drivers physically close together, and the time offset corrected digitally (and maybe phase), will have less of lobing issue and they appear more concentric?
(Lobing is primarily a physical separation distance deal.)

As the me I am wrote afterwards, please do not fall for people who infect you with an everlasting doubt regarding the capabilities of your stereo. Group delay, or whatever they name it, is actually of NO concern. (Except the infamous real big Klipsch horn.)

It is a marketing hoax and nothing else. I bet my life for it. (I trust technology when I step into a commercial jet. The typical audiophile is the infamous 'nervous flyer' who takes your hand on a cross-wind landing. I won't marry her ... . These anxious, even leaking contemporaries just don't understand anything.)

I'm writing this whilst listening to F/ Zappa's 'We're Only In It For The Money'. As time coherence is requested. "Psychedelic Dungeons", you name it ...

 

[kimmosto]

And here it comes again, that "investigation" that was founded on a misunderstanding of some real literature, given in the article as a reference. Nobody could tell why these guys used a 'time inversion' as to 'strengthen' the effect.

In contrast to many others I read the original article and additionally that reference. And I'm so arrogant to say that I've got an understanding of what the authors wanted to convey, and I even understand the reference perfectly.

So, keep on continuing to present an open question: "why these guys used a 'time inversion' as to 'strengthen' the effect"

The reason why these guys compared impulse response to it's time-reversed version is obvious at least in my opinion, though difference in GD between compared impulses could be bigger than expected from the names of test signals. Why you stuck on this detail? Are you trying to deny whole study and every "not same" result no matter what GD difference between original and time-reversed actually was?

 

[kimmosto]

It is a marketing hoax and nothing else. I bet my life for it. (I trust technology when I step into a commercial jet. The typical audiophile is the infamous 'nervous flyer' who takes your hand on a cross-wind landing. I won't marry her ... . These anxious, even leaking contemporaries just don't understand anything.)

Are you okay? Looks that you need a break and rest.

 

[fineMen]

Are you okay? Looks that you need a break and rest.

Seems I eventually made my point.

 

[fineMen]

The reason why these guys compared impulse response to it's time-reversed version is obvious at least in my opinion,...

No, reiterated no. To trust a blatant claim is not scientific.

The reason why these guys compared impulse response to it's time-reversed version is obvious at least in my opinion, though difference in GD between compared impulses could be bigger than expected from the names of test signals. Why you stuck on this detail?

The detail makes it.

 

[gnarly]

I'm happy to just dismiss the idea of that violent imperfection in 'time response'. I'm glad with the the music given to me.

Real instruments in a stereo recording? I know a few musicians who are in the recording business. They actually do not like the sound of recorded instruments to put it mildly. A recording is something else in its own right, not a replica of the real sound. To talk about "the real" is a typical, if not the original audiophile's fairy tale; only the naive, not familiar with the real thing can believe it.

Yeah? What happens when real instruments, their recordings playing through a speaker, start sounding more real?
I'm hearing that all the time...and I abhor audiophoolery (and home audio in general as overpriced prey on the inexperienced/uneducated)

What happens when as a DIY speaker designer/ speaker tuner, you find that realism clearly increases the more you move towards a "perfect impulse"....
....which for heavens-frigging-stupid-simple sake, is no more than flat mag and phase.

Are you gonna listen to naysayers like you who are apparently happy with the idea mediocre sound suffices, and we can't differentiate further?
Are you going to say why bother, I'm kidding myself..... ignorant bliss has to be right.....?

Or, having heard more realism from working with basic audio engineering, trying to take it to excellence, are you going to enjoy ever increasing realism by staying with it?
I know what I'm gonna do....

 

[kongwee]

Both mentioned brands don't have jumpers for phase variations:

https://www.focal.com/sites/www.focal.fr/files/shared/catalog/document/notice_grandeutopiaemevo_web_0.pdf

https://www.wilsonaudio.com/media/917/SabrinaX-Owners-Manual-Web.pdf

https://www.focal.com/sites/www.focal.fr/files/shared/catalog/document/notice_scalautopiaevo_web.pdf

https://wilsonaudio.com/media/858/Sasha-DAW-Manual.pdf

Some has. I never take notice for all the model.