Jump to content

Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers


Chris A

Recommended Posts

20 minutes ago, Marvel said:

It also makes it closer in alignment to the woofers in the Forte and Cornwall. That's a win/win...

Yes...the closer to being within 1/4 wavelength of time/phase alignment and vertical offset spacing at the center crossover frequency, the better the loudspeaker sounds. 

 

I assume this is a good portion of the "secret sauce" that Roy used with the Cornwall and Forte upgrades (and using better performing drivers) over previous models, and why they suddenly became so popular for buyers.

 

Chris

Link to comment
Share on other sites

  • Klipsch Employees
10 hours ago, Chris A said:

I should add that the advice I gave above on adding a "half first order" crossover filter (i.e., high pass only) is the same type of filter configuration in DSP as the passive crossover filters used in analog, and was used while PWK was still at the helm of the company. I think I remember someone saying that PWK mentioned they  "sounded better".  An example of a type "B" crossover for the Cornwall:

 

CornwallTypeB.jpg

 

Note the absence of a low pass inductor on the midrange, and the single capacitor on the tweeter with a "T2A" autoformer to pad down the tweeter's  and midrange's output.  What PWK didn't have was the ability to directly add delay to the tweeter and woofer via DSP (in the Cornwall's case) to align them in time with the midrange, which had a long full exponential horn that created the time delay.

 

More recent updates of the Heritage line of loudspeakers have seen the replacement of the long exponential midrange horns, K-400, K500, K-600, etc., with a K-510 like midrange (straight-sided/tractrix mouth flare) having a much shallower horn that reduces the time delay of the midrange to the tweeter.  This horn is constructed very much like the following figure from Dr. Bruce Edgar's original 1981 Speaker Builder article on the tractrix horn (approximately the red line):

 

1510934234_TractrixhorncrosssectionconstructionfromBruceEdgar.thumb.GIF.da2e1714c738a4928550b80b18d61127.GIF

 

As you can see, this horn is much shorter--and it's actually about 1/3 the length of the exponential horn it replaced.  That means that instead of the midrange being ~3 wavelengths behind the tweeter at the 4.5-->7 kHz crossover point using the full exponential midrange horn, it's been reduced to about one or two full wavelengths behind the tweeter in the newest versions of the Cornwall and Forte loudspeakers. Vertical separation of drivers from each other also add to the diffraction of the radiated sound field in front of the tweeter+midrange horns/drivers.  I believe that the crossovers used currently are probably using second order filters (Roy hasn't explained that yet), but the second order filters (adding an additional 180 degrees of crossover filter induced phase lag on the lower frequency drivers) only add ~0.68 inches to the physical time misalignment over using first order filters...an added 90 degrees of phase shift over first order that could be avoided.  Given the rave reviews of the Forte and Cornwall updates (and to a lesser degree the Heresy), it seems likely that at least a portion of that increased satisfaction by listeners is due to the reduced time misalignments.

 

Chris

What you are missing is the fact that Mr K moved on to steeper slopes (whatever was required) because of the coverage interfere band and the distortion the drivers would create because they did not have the proper filtering. Just because they do not create acoustic pressure does not mean that the diap is not moving. 

Link to comment
Share on other sites

Most of the people discussing this were apparently worried about doing damage to drivers.  I believe that the weight of evidence of PWK's crossovers over decades of use show that (except for the very low power K77s at 4-5 watt input limitation which can be addressed via fast limiters in the DSP crossovers), that damage isn't really a big concern unless someone is using a lot higher power amplifiers than Klipsch recommended (which I've posted elsewhere: a ca. 1979 Dope from Hope article).  That was the real question being discussed, above. 

 

Now as far as the distortion that you refer to:  if that is an issue for you, then use the higher slope filters. 

 

For me and my setup (TAD TD-4002s, BMS 4592ND dual diaphragm, AMT-1s, and horn-loaded bass bins), my ears tell me that the reduction in phase growth due to the crossover networks themselves is more than a good trade.  Of course, I'm not using my loudspeakers for in-home concerts at 110 dB+ (1 m).  The subject of this thread--started about 2 years ago now (March 2019), is the effects of lowering the crossover-induced phase growth, and the techniques for DIYers to try it for their own ears. These effects that I found were so surprising that I chose to share it with others--that's all.  

 

Chris

  • Like 1
Link to comment
Share on other sites

One other point that I feel needs to be raised, and that's the default idea that "one size fits all" (...which it almost never does):

 

In the past, the idea of having different crossover filters for different usage (in-home hi-fi, outside PA, rock concerts, etc.) was not even discussed.  Nowadays, if you want to have different settings for outside temperatures (a minor factor in time alignment of drivers), celebrating with a block party at outside PA levels, or even taking your loudspeakers on the road to support a rock concert, these conditions are about as complicated to address via front panel or even smartphone control with a flick of a button. 

 

So if you're using your loudspeakers inside at sub-80 dB or sub-105 dB (1m) levels 99.9% of the time, you can set up for that condition and optimize the settings for your listening pleasure, but in the event that you're using your loudspeakers in a dorm window blasting at the full power of the amplifier(s), then different settings can be used--including those that include the fast limiters.  No one is going to be close enough to the loudspeakers to hear the effect of phase flattening under the "dorm window" conditions anyway. 

 

This is the same issue with stereo music tracks--they're typically sold in a one-size-fits-all manner, when the incremental costs to provide more than one mastering condition to the tracks (for different customers having different quality playback equipment) is effectively zero if considering music downloads.  One-size-fits-all thinking has apparently attached itself to the music and audio industry, and the effects of this are significantly decreased enjoyment, in my experience.  Even Coca Cola figured it out about 30 years ago to provide different products for different tastes of customers.

 

The tendency of one-size-fits-all reminds me of the time when British troops were obliged to wear boots that were not differentiated left from right: only one boot stock number was issued (uni-foot)--in order to "simplify the logistics of supplying boots".  The British troops celebrated for a long time after that practice was finally rescinded by the military leadership in the early 19th century (IIRC). 

 

I'm still waiting for the music industry to figure it out.

 

Chris

  • Like 1
Link to comment
Share on other sites

Hi Chris 

                                I have the same 2 way Jubilee /Tad 4002 beryllium  setup as yourself , and would like to try the K402 horn repositioning that you mentioned at the beginning of this post that gave you such good results . Is there an image of your Jubilee `s  with the K402 horn position change you made  you can post so I can be sure that I have it right .  They say a picture is worth a thousand words .

                                                     

Link to comment
Share on other sites

  • Klipsch Employees
5 hours ago, Chris A said:

Most of the people discussing this were apparently worried about doing damage to drivers.  I believe that the weight of evidence of PWK's crossovers over decades of use show that (except for the very low power K77s at 4-5 watt input limitation which can be addressed via fast limiters in the DSP crossovers), that damage isn't really a big concern unless someone is using a lot higher power amplifiers than Klipsch recommended (which I've posted elsewhere: a ca. 1979 Dope from Hope article).  That was the real question being discussed, above. 

 

Now as far as the distortion that you refer to:  if that is an issue for you, then use the higher slope filters. 

 

For me and my setup (TAD TD-4002s, BMS 4592ND dual diaphragm, AMT-1s, and horn-loaded bass bins--my ears tell me that the reduction in phase growth due to the crossover networks themselves is more than a good trade.  Of course, I'm not using my loudspeakers for in-home concerts at 110 dB+ (1 m).  The subject of this thread--started about 2 years ago now (March 2019), is the effects of lowering the crossover-induced phase growth, and the techniques for DIYers to try it for their own ears. These effects that I found were so surprising that I chose to share it with others--that's all.  

 

Chris

So “flat” phase at the expensive of distortion and coverage?  Ok. Low order filters cause so much lobing and nulling that the ear tends to “average” it. Kinda of like line arrays. 

Link to comment
Share on other sites

8 hours ago, Chief bonehead said:

So “flat” phase at the expense of distortion and coverage?  Ok. Low order filters cause so much lobing and nulling that the ear tends to “average” it. Kinda of like line arrays. 

So it's the idea of flat phase crossovers that you don't like...and not so much what someone else's ears are actually responding to?  If this is the problem, I would recommend staying with the higher order crossover filter settings. You're probably not listening to the type of music that I listen to regularly (briefly described below) and you more highly value something other than what I hear.  We've discussed this before.

 

I've personally found that each loudspeaker in my 5.2 channel array subjectively benefits from the flat phase crossover settings, especially when I listen to the best recordings (notably acoustic-only instrumentation, with few or no electrically amplified instruments/voices).  The results are actually breathtaking at times.  I have to say that I do respond to that and enjoy listening in a way that I've never experienced before, even to the point of collecting music of certain genres now that I haven't in the past--something extremely difficult to reproduce: full orchestra.

 

I would liken the experience sort of like that of little loudspeakers on sticks (like Mike B. uses) but without the audible disadvantages that those little loudspeakers bring in terms of absent bass response, strong floor and other room boundary bounce (due to their lack of good directivity control vs. frequency), and the inevitably much higher modulation distortion of the direct radiating loudspeakers that severely limit their dynamic range.  I get flat phase without the drawbacks just mentioned.

 

I mentioned that when I moved the K-402s slightly closer to the centerline of the KHJ bass bins, the subjective listening experience got even better, not only moving the high axis of the K-402s down to the recommended 1.2 m ITU home theater standard height, but also in-room measured phase response, posted here.  The only real performance issue of the Jubilee configuration for home hi-fi duty is its directivity issue around the crossover frequency.  I find that moving the K-402s down a little audibly addresses that issue much more effectively than trying to use higher order crossover networks to hide the problem--that induce all the phase shifts.

 

Chris

Link to comment
Share on other sites

On 2/16/2021 at 8:30 PM, Ivanhoe56 said:

Hi Chris 

I have the same 2 way Jubilee /Tad 4002 beryllium  setup as yourself , and would like to try the K402 horn repositioning that you mentioned at the beginning of this post that gave you such good results . Is there an image of your Jubilee `s  with the K402 horn position change you made  you can post so I can be sure that I have it right .  They say a picture is worth a thousand words .

                                                     

I don't have a problem posting my Xilica settings, but note that my delays will probably differ from yours (i.e., I shorten-down the K-402s on top of the bass bins to reduce the vertical distance between the center of the K-402s and the geometric center of the bass bins--like the below picture courtesy of @MisterVego.

 

Jubes.jpg

 

When you get to the smaller offset distances between "ways" of a loudspeaker, I find that the effects of driver separation actually increase until you get within 1/4 wavelength of each other at crossover, at which point the drivers/horns "couple" to each other.  This means that small offsets just outside of 1/4 wavelength become more critical than the further effects of increasing the vertical or horizontal offset distances to even larger values, since they continue to lose the effects of coupling until perhaps one or two wavelengths of sound exist between them at the crossover frequency.

 

My Jubilee settings from XConsole:

 

1988961002_AskewDeviceWindow(two-channelJubs).GIF.0bbb7d73197b1664be85fabd99780b55.GIF

 

The left bass bin channel (output channel #1):

255707971_Askewleftbassbinwindow(two-channelJubs).thumb.GIF.f0f08ccbec73060c82c3fcd17721ce6d.GIF

 

My left output channel TAD TD-4002 settings:

 

2124078374_AskewleftHFwindow(two-channelJubs).thumb.GIF.e888eba13af19771d3f16fc5cca918c9.GIF

 

And I have a couple of PEQs in the input channels:

 

419716710_Askewleftinputchannelwindow(two-channelJubs).thumb.GIF.e10e1ac9b3e4fc0f22729aafcce1f0e4.GIF

 

If you're worried about the issues that Roy raises above, simply add more attenuating PEQs on the output HF (TAD) channels below 300 Hz to further curtail any issues with out-of-band low frequencies exercising the TAD diaphragms.  You can move one or more of the higher frequency output PEQs (well above the crossover point) that are taking up all 8 slots presently in the output channel, above, to the input available input channels, then add the attenuating PEQs below the crossover point with the newly unused output PEQ slot(s) that you opened up.

 

Here is the stereo-only .xdat Xilica file that reflects my settings, only for the Jubs only:  Askew's Jubilee settings (left right only).xdat

 

 

 

Chris

 

 

Link to comment
Share on other sites

Link to comment
Share on other sites

Chris,

Again, thanks for the effort and write-up which was no mean feet.

 

With some frequency, I am at a loss for the correct descriptive audio terminology, so thanks for going easy on me in advance.

 

Years ago (20+) an article estimate the DSP performance of the Human Brain for graphics processing as equal to 26 Cray Supercomputers. Might have been self driving robots.

 

For simplicity, lets say the Human audio processing is similar.

 

A motorcycle article Super Bike Shoot Out

had ten professional drivers ride all of the top Super Bikes, and rate them on scales and various criteria. Acceleration, top speed, braking etc. Suzuki (if I recall correctly IIRC) was the winner. Last question was subjective, "Which bike would you like take home", all of them chose the Ducati F1 (IIRC), which finished like 4th place. Conclusion, either they were measuring the wrong thing-s, or there is a quality I will call summation. The whole is better than the parts.

The motorcycle Shoot Out was measuring the wrong thing.......


Je ne sais quoi

 

I had such a girlfriend, from Montreal,  if you put her in a room with other women, the guys would gravitate towards her. The other girls noticed it too. Why ?

How do you measure Wiles ? but it exists.....

 

Quote

Camille Chevalier-Karfis

Updated October 06, 2019

"Je ne sais quoi" is a French idiomatic expression used so much in English that it has made it into the leading English dictionaries. In other words, it has been assimilated into the English language.

Merriam-Webster describes je ne sais quoi as "something (such as an appealing quality) that cannot be adequately described or expressed," as in "This woman has a certain je ne sais quoi that I really like." In French, Larousse calls je ne sais quoi " a thing one wouldn't know how to define but whose existence is understood intuitively."

 

How the French Use It

 

But the French also use it as we do in English: a quality you cannot describe. We connect je ne sais quoi to the adjective describing it with de, like this:

Cette fille a je ne sais quoi de fascinant.

"There is something fascinating about that girl."

 

 

Another example

I was speaking with an x-mil friend with lots of all kinds of experience. I told him when I am traveling,  lots outside CONUS, that sometimes I would intuitively not want to enter a street, and for no apparent reason go around, backtrack or hop a cab even for a few blocks. I asked my friends

"Should I pay attention to these feelings or an I being Paranoid" ?

 

Response,

"if you don't listen to your instincts, in war zones you die quickly".

I asked for clarification,

"your brain recorded something dangerous or out of place and signaled you, don't go".

"there was some kind of danger on the street"

 

I think that we need to accept that the brain is tremendously powerful, especially with the DSP aspects and background processing.

 

An example of perception, I think it was TIM that took some time and effort to identify the source, characterize and solve.

 

Another example from JBL 4340

IIRC the designer of the 4340 Studio monitors,

loved by musicians for their studio work

The designer was asked "How did you do it "

Answer

Did all of the design and component homework

and chose an assistant with great ears,

voiced and equalized the design using the assistant's ears as my guide.

When they sounded great, the design was released.

People gravitate towards these speakers........

 

The article, if you can find it,  the design, components and complex equalizer combination are worth a look.

 

Trial and error with alignment and positioning and equalization, should not be dismissed just because it is difficult to characterize before the experimentation.

 

Life Like Sound Reproduction like Wiles is difficult to define

but you know it when you see it

or in the case of audio hear it.


 

Quote

 

wiles

verb

present tense third-person singular of wile

Synonyms & Antonyms of wiles (Entry 2 of 2)

to attract or delight as if by magic

her stories of the Old South could wile anyone

 

Synonyms for wiles

 

allures,

beguiles,

bewitches,

captivates,

charms,

enchants,

fascinates,

kills,

magnetizes,

witches

[archaic]

 

Words Related to wiles

 

disarms,

draws,

entices,

lures,

pulls,

seduces,

tempts

delights,

gratifies,

pleases

arrests,

enraptures,

enthralls

(or enthrals),

entrances

appeals (to),

interests,

intrigues

beckons,

courts,

invites,

solicits,

woos

 

https://www.merriam-webster.com/thesaurus/wiles

 

  • Like 1
Link to comment
Share on other sites

My thoughts : what problem or application was the Jub designed to solve ?

It was designed as a Med-Large space speaker.

 

If the source of the sound is not localized compared to experience, it's off.

The further from the source the distance between drivers may become less perceptible, the argument for small room, single full spectrum driver.

 

We want the human voice to be at ear level, this is our experience, that the brain compares to. It is the reason I chose the LaScalas over the K-horn, and placement flexibility.

 

Drivers closer together passes common sense, as does driver alignment  by electronic means since physical is not possible.

 

This is the reason my eye goes to the 4340 and Altec 820c, I'm sure both have their own issues, but driver alignment and driver to ear alignment aren't two of them....

 

EQ what sounds good in a small space, almost any house of any size.

 

 

Link to comment
Share on other sites

2 hours ago, Chief bonehead said:

Ummmm.....no. You are assuming something I didn’t say. 

Then I suppose that its really disagreement at its core as the central issue.  That's okay--we can agree to disagree.  I'll stand by what I've said, and you can certainly do the same.  I'm okay with that. 

 

3 hours ago, Bubo said:

Years ago (20+) an article estimate the DSP performance of the Human Brain for graphics processing as equal to 26 Cray Supercomputers. Might have been self driving robots.

I do have a fair amount of experience in this domain from a system design standpoint (ground autonomous vehicles and things that fly autonomously through the air--with the flying problem being MUCH less technically demanding than the former).  But I digress.

 

3 hours ago, Bubo said:

...Another example from JBL 4340. IIRC the designer of the 4340 Studio monitors, loved by musicians for their studio work.  The designer was asked "How did you do it?"

Answer:

Did all of the design and component homework and chose an assistant with great ears, voiced and equalized the design using the assistant's ears as my guide.  When they sounded great, the design was released.  People gravitate towards these speakers...

I found this thread but haven't delved into it any further: http://www.audioheritage.org/vbulletin/showthread.php?16493-4340-amp-4341-references-thread

 

But I believe I get your point and the dialing-in process. I've found that over time, my ears don't really accommodate as much as people would generally have our ears accommodate. I still yearn for the actual live sound of the performances that I spent much of my youth participating in.  And that's what fuels my efforts in this particular subject area. It takes many, many hours of listening and experimentation with a focus on live acoustic presentation as the standard, I believe.  The more I listen, the more sensitized to sound reproduction issues that I find myself focusing on--something that would be difficult to hear in just an afternoon of listening or an afternoon DBT (blind test), etc. 

 

As an engineer, I have to be careful that this thread doesn't devolve into yet another "opinions only--subjectivism vs. objectivism" thread.  I'm not really into those, and I try to stay as close as I can to the engineering domain--as it intersects the psychoacoustics domain.  It's only when we get to the dilemmas of the type highlighted just above that sometimes I have to pull out the subjectivism card--and I don't do it cavalierly. In fact, I hate to do it.  The 1990 JAES Greenfield and Hawksford article that I posted in the third post of this thread is one such reference that explores the intersection of acoustic engineering (where I think Roy is primarily focused) and the field of psychoacoustics (i.e., repeatable evidence about how we hear as humans), and is one that I believe anchors the discussion to objectivist engineering-only discussions. 

 

What this thread was meant to do is to convey what I found using crossover schemas that don't inject all-pass delays in the loudspeakers and listening rooms--over and above what we have to live with, and how I achieved it.  Nothing more.  If it becomes only pure subjectivism arguments, or purely engineering--measurement arguments, devoid of the psychoacoustic aspects of how we hear, then I think we need to conclude the discussion and let what I wrote above stand on its own.

 

3 hours ago, Bubo said:

My thoughts : what problem or application was the Jub designed to solve?  It was designed as a Med-Large space speaker.  If the source of the sound is not localized compared to experience, it's off.

Let's be clear: I really don't agree with the implications of this statement.  What I've found is that the Jubilees actually do better in a home-sized listening room than one might suppose by just looking at them or considering their marketing provenance.  And they responded extremely well to the tweaks that I talk about in this thread.  I believe that their full-range directivity was a good portion of that, as well as their ability to hold down modulation distortion and compression distortion.  Once the phase response of the Jubs was flattened by removing most of the crossover filter-induced phase growth at lower frequencies, it was like flipping on the switch to the house Christmas lights display:  It was really something to experience--the lights went on, at least for my listening pleasure and others that have tried it and listened to it in A-B fashion.

 

What I've found is that using a single horn aperture (i.e., using a multiple entry horn to eliminate the issues that were being discussed, just above) is actually better in terms of midbass and lower midrange acoustic performance, without apparently trading away its low frequency performance--which was a really big surprise for me, and one that I think can be explained by its use of a larger mouth horn.  The good thing is, the K-402 is made by Klipsch and can easily be converted for MEH use. But the Jubilees hold their own quite well, better than I thought from purely top-level (visual and anticipated measurement--i.e., engineering) reasons only.

 

Chris

  • Like 2
Link to comment
Share on other sites

This thread on DIYaudio discusses another way to achieve low phase growth in a crossover.  I don't speak French, so I can't make much of the original source material, but the OP seems to get very good phase behavior out of his speakers.

 

https://www.diyaudio.com/forums/multi-way/277691-harsch-xo.html

 

I may have to try this technique out.  His phase measurements are impressive.

Link to comment
Share on other sites

I tried that crossover in 2018.  I found that a straight first order crossover worked better and had lower phase growth.  Bessel second order worked not as well as first order, but was still much better than the Harsch crossover in terms of avoiding phase growth. 

 

Then on April 1, 2019, I found this:

 

https://community.klipsch.com/index.php?/topic/182419-subconscious-auditory-effects-of-quasi-linear-phase-loudspeakers/page/3/&tab=comments#comment-2379562

 

and the method I used is described here:

 

https://community.klipsch.com/index.php?/topic/182419-subconscious-auditory-effects-of-quasi-linear-phase-loudspeakers/page/4/&tab=comments#comment-2388972

 

These basically eliminated the phase growth altogether.

 

Chris

  • Like 1
Link to comment
Share on other sites

Chris,

   I'd love to get your thoughts on some more questions I have regarding phase growth.  First, it looks like you achieve less than 90 degrees of phase growth from 100hz up with your Jubilee/TAD setup.  So far, I have managed to reduce the phase growth in my setup (LaScala bass bins, B&C 2" CD on ZXPC horn) from ~700 degrees of phase growth above 100hz, down to about 240 degrees, by using your "fractional order" method.  The improvement is great, but naturally, I'd like to go farther.  To what do you attribute your lower phase growth?  Is it a characteristic of the TAD drivers or the K-402?  Or is there something in the crossover setup that is driving this?  Is there something else I could be doing to farther reduce phase growth?

  Second, It seems that in all of these measurements, phase grows considerably below 100hz, regardless of crossover settings.  Why is this?  Is there something about bass frequencies or drivers that fundamentally creates phase growth?  Is there anything that can be done to reduce phase growth at lower frequencies?

 

Lee

 

Link to comment
Share on other sites

1 hour ago, Tarheel TJ said:

...So far, I have managed to reduce the phase growth in my setup (LaScala bass bins, B&C 2" CD on ZXPC horn) from ~700 degrees of phase growth above 100hz, down to about 240 degrees, by using your "fractional order" method.  The improvement is great, but naturally, I'd like to go farther.  To what do you attribute your lower phase growth?  Is it a characteristic of the TAD drivers or the K-402?  Or is there something in the crossover setup that is driving this?  Is there something else I could be doing to farther reduce phase growth?...

If you look at a case study, say a La Scala II tri-amp with time alignment.  First let's look at the Spectrogram of the completed dial-in and time alignment to verify we've done our job as best we can:

 

927179822_LaScalaIITri-AmpTimeAlignmentSpectrogram.jpg.cd0c47eee6b39e64fb05935d6e0ca23f.jpg

 

Note the hook to the right of the "Peak energy time" trace (the dotted black line) starting just above 1 kHz, and then returning to zero time datum around 200 Hz.  This is characteristic of the midrange phase growth of the K-401 horn and driver (I believe it's a K-55, but I could be mistaken).  This is the best we can do, and the total misalignment of the peak energy time curve is below about 0.6 ms, which is extremely good for the frequencies involved.  More important is the tweeter/midrange time alignment at ~5-6 kHz, which looks to be perfect, i.e., you can't see much happening there.

 

Now if we turn our attention to the SPL and phase plot of the same measurement, we get a completely different picture:

 

1812189749_LaScalaIITri-AmpTimeAlignmentSPLPhaseExcessPhase.jpg.b358746539335419dc47629aa81502f4.jpg

 

This shows the SPL response to be almost perfect (on-axis)--within ±1.5 dB using psychoacoustic smoothing.  But the phase plot is all over the place.  Some of this is due to in-room reflections that disrupt the calculation of phase within REW (it's an inherently noisy numerical process, apparently, but floor bounce is a really bad issue, as well as nearby wall reflections due to the loudspeakers being close to the walls in a smaller listening room). 

 

So the question is: what's real and what's not?  We can also look at the group delay plot to answer that question, and in particular, the excess group delay curve to see if we've gotten all the "all pass" delays out of the loudspeaker crossover regions:

 

593468584_LaScalaIITri-AmpTimeAlignmentGroupDelayandExcessGroupDelay.jpg.0a404cfbccdb8c4830d2d67705a62bff.jpg

 

The excess group delay curve shows that midrange horn-driver movement of its acoustic center backward (away from the listener) as we go below ~1.5 kHz, then it returns to zero excess group delay at ~300 Hz.  This says that we've taken out all the excess group delay (and therefore all the phase growth due to the crossover).  It also shows a fairly wide total group delay response (oscillating, that is) that we can't do anything about using the type of digital filters that we have on the DSP crossovers we use. 

 

We also see a blip on excess group delay at 6.3 kHz (apparently the crossover point between the tweeter and midrange), but again, the excess group delay curve returns to zero below that frequency.  This is due to the tweeter being physically more than 1/4 wavelength away from the centerline of the midrange horn at the crossover frequency.  At 6.3 kHz, that 1/4 wavelength distance is 0.53 inches, which is a smaller distance than the vertical separation of the tweeter and midrange horns centerlines.  So you've got that excess group delay bump that's going to be there due to the vertical separation of the drivers/horns at the very small wavelengths present at that high crossover frequency.  (This is the reason why supertweeters aren't a very good idea.  It's better to use one driver or a coaxial driver with two diaphragms above about 1-2 kHz.)

 

So the bottom line is: although we're reading a large excursion in the phase curve, the reality is that this is the best we can do without resorting to using the much more computationally intensive FIR (finite impulse response) filtering.  And it's a function of the drivers/horns themselves and their vertical separation distance.

 

The reason why the Jubilee phase response looks so flat...

 

TAD TD-4002 Jubilee vs. Danley SH-50 phase response.jpg

 

...is due to the fact that there are no crossovers occurring between separate horns above the only crossover frequency (about 500 Hz in this case), and the horn/driver (a TAD TD-4002 beryllium dome diaphragm) doesn't exhibit that big shift in acoustic center backwards as we decrease the frequency--all the way down to its crossover point with the bass bin. 

 

We could improve the La Scala II's measured results in these areas by using FIR filters, but note that this won't correct the lobing due to the tweeter and midrange centerline being more than 1/4 wavelength apart vertically.  FIR filtering will correct the shift in the acoustic center of the K-55/K-401 horn backward vs. frequency (the latter effect most likely due to the pattern flip midrange horn used on the La Scala). 

 

Questions?  (I'm sure there are many...)

 

Chris

  • Like 2
Link to comment
Share on other sites

1 hour ago, Tarheel TJ said:

Second, It seems that in all of these measurements, phase grows considerably below 100hz, regardless of crossover settings.  Why is this?  Is there something about bass frequencies or drivers that fundamentally creates phase growth?  Is there anything that can be done to reduce phase growth at lower frequencies?

If you look at the total phase and total group delay curves below 100 Hz, you will see that the phase and group delay growth is really dominated by the "minimum phase" (the other part of phase and group delay not related to crossover filters, etc.).  This minimum phase group delay can be corrected, but the solution is to use FIR filters having an extremely high number of "taps" (the quantity of individual numbers in the column of row of numbers which defines each FIR filter).  The time delays associated with correcting these can be bigger than the available digital synchronization correction television available from AVRs and AVPs and flat screen displays. 

 

You can talk to @Edgar (Greg Berchin) to see what can be done to flatten the phase down there (and Dirac, etc. "room correction software" tries to do this within its own framework).  But in general, the computational requirements become enormous to do it. 

 

Would it be audible to correct the phase growth below 100 Hz.  YES. Is it feasible?  Only in extreme DIY implementations using PCs with lots of computing horsepower have I seen it corrected, and generally, there is a frequency below which the correction has to stop due to computing limitations and synchronization issues with flat screen displays.

 

Chris

  • Like 2
Link to comment
Share on other sites

   Thank you for the explanation.  

   So it sounds like the phase growth in my setup (above 100hz) is mostly related to the inherent characteristics of the horn/driver combo.  So potentially going to a different horn/driver could lead to improvements in that area?

   Below 100hz, it seems there isn’t much that can be done besides using FIR filters.  
  So it sounds like I have made about as much of an improvement as I can with my current hardware.  Would you agree?
    I have one more question.  How do you get REW to generate the spectrogram image?  How about group delay and impulse response?  I can do SPL and phase response easily enough, but the others just show a blank screen when I click on them.  Thanks again.

 

Lee 

Link to comment
Share on other sites

29 minutes ago, Tarheel TJ said:

So potentially going to a different horn/driver could lead to improvements in that area?

Yes.  This is, in fact, a "measurement-based" way to assess drivers/horns, and is probably just as important as SPL response and coverage vs. frequency.  There is also the subjective/listening portion that is required, too.  Some drivers measure well but don't sound very good.  I believe that diffraction and other higher-order modes are the root of most of these issues (like "constant directivity" horns having long slots in their mouths).

 

29 minutes ago, Tarheel TJ said:

 So it sounds like I have made about as much of an improvement as I can with my current hardware.  Would you agree?

Likely.

 

29 minutes ago, Tarheel TJ said:

How do you get REW to generate the spectrogram image?  How about group delay and impulse response?  I can do SPL and phase response easily enough, but the others just show a blank screen when I click on them.

There is a "generate" button in the bottom left corner of the spectrogram plot when you move the cursor into the plot area.  Press that.

 

The excess phase and excess group delay plot is available after you go to the "Controls" menu on the right hand side and press "Generate minimum phase" button, then go to the bottom of the plot and check the box next to "excess phase" or "excess group delay" to turn on the plot.

 

The impulse response plot should be there. Try selecting a "dBFS" or "%" from the control at the top of the vertical axis when the cursor is in the plot field.

 

Chris

  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...