Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers

[Langston]

On 7/15/2021 at 11:45 AM, VDS said:

What is going on at 4-6khz?

 

Maybe nothing, depending on how REW is presenting the phase curve. I'm not a REW user*, but if it's set to "wrap" phase at ±180˚ points, then the 4-6kHz wrapping may simply be due to the curve kissing the -180˚ point. If REW is presenting the phase curve "unwrapped", then ignore me. : )

 

On 7/15/2021 at 5:30 PM, Wirrunna said:

What does the minimum phase curve tell us?

 

This is the kind of thing @Chris A has probably defined already, but it's the minimum amount of frequency delay (the initial zero-cross of the sinusoidal) possible for a perfect device with a finite passband. Minimum phase is not measured, it is a mathematical transform of the magnitude (SPL) response. Loudspeaker drivers of good quality without external filters (passive, active, or acoustic) usually behave like minimum phase devices quite well. When plotted against the measured** phase curve, the difference reveals non-minimum phase behavior. This difference is called excess phase, thus the excess phase curve is the measured phase curve with the minimum phase component subtracted out of it. This same concept extends to the measured, minimum and excess group delay curves. There is much that can be learned by looking at each of these three curves and the REW manual has some excellent instruction on this.

 

God bless you and your precious family - Langston

 

* I've read the manual and in my opinion it would be a good deal if it sold for $1,000. I sent the author $50 a few years ago to thank him for his contribution to our sport.

** Assuming the measured phase curve is measured correctly, i.e., the microphone and electronics have their inaccuracies compensated for, the physical measurement setup is correct, and the user sets measurement system to remove the correct TOF (time of flight between microphone and driver). Correct TOF removal is no piece of cake and it's a fascinating subject. (PM me if you want this paper)

 

[VDS]

On 6/5/2019 at 11:30 AM, Chris A said:

IIR filters only--no "named" crossover filters, i.e., Danley-style. 

 

Look at the raw response of your drivers/horns.  Without FIR filtering, you can't improve on the response that you see there after they've been EQed flat (individually).  So you need to select drivers and horns that have good phase response characteristics.  It just so happens that the Jub bass bins and TAD 4002/K-402 horns have really good phase response taken individually--after they've been EQed flat by themselves.

 

So all you have to do is put the HF and LF drivers together without phase shifts:

1. Don't use the "crossover filters" that come with DSP crossovers--clear any crossover filters if they're set.
2. Set the HF or LF channel delay to get perfect impulse response in the time domain--as seen in the spectrogram view.
3. Flatten each driver's SPL response within their pass bands.
4. Match the channel gains between flattened phase drivers.
5. Use output channel PEQs to trim off response on each end of the bass and high frequency drivers until you've got overall flat SPL across the crossover interference band and smooth handover of SPL vs. frequency.  The drivers themselves will tell you where that transition/crossover should occur.  [If you're using MEHs, you'll have to use multiple PEQs to attenuate the bass bin peaks in response above the first notch frequency.]
6. Use the input channel PEQs to further flatten the overall response within the interference band to correct any dips or peaks in response within that band.

Voila!  Flat phase.  It's really that easy.

 

Chris

@Chris A,I’ve been doing this method, with steep drops and rises where drivers meet. Rereading this thread you are advocating shallow drop offs.  When you “trim the ends” do you use PEQ to get both drivers to rise and fall slowly, similar to a 6db/oct ?  Mine have been very steep, following the natural driver drop off/rise.

Ted

 

 

[Chris A]

1 hour ago, VDS said:

When you “trim the ends” do you use PEQ to get both drivers to rise and fall slowly, similar to a 6db/oct ?

I use the natural acoustic roll-off of the acoustic drivers when I can. 

 

Sometimes a large overlap of driver frequency/phase responses between "ways" makes the job more difficult with a  DSP crossover since neither driver on each side of the selected crossover frequency is really attenuating by itself at the chosen crossover point. So in those particular cases, more PEQs usually are needed to attenuate the response of both drivers at the ends of their frequency response. 

 

But generally, one or the other driver is close to its own steep roll-off in response for many/most applications., and usually combine with the applied DSP crossover attenuation PEQs to give you 10 dB/octave or more slopes on either side of the center crossover frequency that's selected.

 

In the case of the prototype K-402-MEH, the natural roll-off of both the woofer (low pass) and BMS 4592ND (high pass) is ~10 dB/octave at the nominally 550 Hz crossover point: a so-called "fractional order" acoustic crossover, neither first order nor second order.  The same is true for the Jubilee bass bins and the TAD TD-4002 drivers: I get about 10 dB/octave effective slopes

 

Chris

[Chris A]

By the way, I just thought of this, and I think it should be laid out here:

 

If using multiway loudspeakers having separate drivers/horns, typically the mouth of the horn or the front face of the driver (for direct radiating loudspeakers) is clamped to the front baffle of the loudspeaker, thus inducing a physical misalignment of the drivers' acoustic centers, with the higher frequency drivers typically leading the lower frequency drivers in terms of their time alignment.  These loudspeakers need digital delay added to their DSP channels to time align them to the lower frequency drivers.

 

 

In the case of the multiple entry horn (MEH), the physical alignment of the drivers has the lower frequency drivers mounted in front (closer to the listener's ear) than the higher frequency drivers:

 

 

So, if a zeroth order crossover filter is used, it will allow the lower frequency drivers to lead the higher frequency drivers by 1/4 wavelength at crossover (90 degrees).  While this is acceptable for an MEH time alignment, it will actually be a little closer if a first order filter is used to delay the lower frequency drivers by 90 degrees, thereby achieving time alignment without having to "cut and try" the channel delay  corrections. 

 

So bottom line:  You actually need something in the lower frequency driver channel(s) to delay them in an MEH, albeit very slightly.  Just using a first order set of crossover filters (low pass on the lower frequency drivers and high pass on the higher frequency driver[s]) is actually desirable.  I've used just straight delay on the lower frequency driver of the K-402-MEH (equivalent to 90 degrees at the crossover frequency) and alternatively, just simple straight first order filters without digital channel delay, and could see no difference in the output of the MEH.

 

JMTC.

 

Chris

[Supersteff]

Hi Chris

 

Hmm, I need to ask to be sure to understand what you are saying here.

 

Are you saying here, that it is OK to use a first order high-pass on the AXI2050, without messing with/degrading the phase-response? I presume in combination with PEQ´s.

 

In an other thread of yours, you have said that the AXI2050 does NOT need to have any high-pass filtering (only PEQ´s) to protect it from low frequencies (when not playing very loud) only utilizing the natural roll-off of the CD for cross-over.

 

I hope I made my point clear!?

 

Steffen

[gnarly]

13 minutes ago, Chris A said:

 

 

In the case of the multiple entry horn (MEH), the physical alignment of the drivers has the lower frequency drivers mounted in front (closer to the listener's ear) than the higher frequency drivers:

 

 

So, if a zeroth order crossover filter is used, it will allow the lower frequency drivers to lead the higher frequency drivers by 1/4 wavelength at crossover (90 degrees).  While this is acceptable for an MEH time alignment, it will actually be a little closer if a first order filter is used to delay the lower frequency drivers by 90 degrees, thereby achieving time alignment without having to "cut and try" the channel delay  corrections. 

 

So bottom line:  You actually need something in the lower frequency driver channel(s) to delay them in an MEH, albeit very slightly.  Just using a first order set of crossover filters (low pass on the lower frequency drivers and high pass on the higher frequency driver[s]) is actually desirable.  I've used just straight delay on the lower frequency driver of the K-402-MEH (equivalent to 90 degrees at the crossover frequency) and simple first order filters, and could see no difference in the output of the MEH.

 

JMTC.

 

Chris

 

Hi Chris, i should first ask what do you mean by a "zeroth order crossover" ? Not familiar with that idea...

 

Here's my two cents...

 

So far, on every MEH i've built (maybe a dozen counting fully developed prototypes), the woofers have always measured further from the mic, than from the CD.  This has occurred despite the fact the woofers are closer to the mic than the CD,....... as in the drawing you posted.

 

And it occurs whether taking time-of-flight measurements with no filters in place, or with linear-phase low passes of any order (which don't have any group delay).

 

I've come to the conclusion, perhaps wrongly, that the natural acoustic low pass of the woofer is the source of the delay.  Well that, and also i believe that our FFT measurement programs use the Hilbert-Transform to make TOF determinations under the assumption minimum phase represents correct timing.

 

At any rate, i know all my builds have had to add delay to the CD.  

For the first time recently, i tried adding small 4" mids between the Cd and woofer (even though not needed).

The mids also need delay relative to the woofer. 

With the woofer at 0.00ms, mids delay is 0.32ms, and CD is 0.83ms & 0.90ms (hf & vhf dcx464 coax sections)

 

You can see from the spectro, those delays put timing right on target.  (other than i got a small bobble at 250Hz, xover between woofer and mid)

 

 

 

[Chris A]

2 minutes ago, Supersteff said:

Are you saying here, that it is OK to use a first order high-pass on the AXI2050, without messing with/degrading the phase-response? I presume in combination with PEQ´s.

 

1) The Axi2050 needs PEQs to flatten its response on almost any horn that is straight-sided, etc.  All drivers on straight-sided horns having controlled directivity need EQ--and this is just physics.

 

2) All "ways" in an MEH can actually use first order filters to time align the lower frequency drivers to the higher frequency drivers that are physically farther away from the listener.  This is very different than typical multiple aperture (conventional) horn loudspeakers having separate horns and drivers, and direct radiating drivers in other non-horn-loaded loudspeakers, which benefit greatly from zeroth order filters in terms achieving better overall acoustic phase response flatness.

 

5 minutes ago, Supersteff said:

In an other thread of yours, you have said that the AXI2050 does NOT need to have any high-pass filtering (only PEQ´s) to protect it from low frequencies (when not playing very loud) only utilizing the natural roll-off of the CD for cross-over.

 

3) The Axi2050 driver is so robust in its design that unless you're using it at PA or higher SPLs, it doesn't need blocking capacitors, etc. to protect it.  In fact, essentially all 2" compression drivers can take the full output of an amplifier (less than 50 watt/channel) without issues. 

 

[It's the little K-77 phenolic dome driver in the Klipsch Heritage series that has trouble taking more than 4 watts, that you really need to do a fast limiter on peak and integrated voltage levels using a DSP crossover.]

 

Chris

[Chris A]

27 minutes ago, gnarly said:

So far, on every MEH i've built (maybe a dozen counting fully developed prototypes), the woofers have always measured further from the mic, than from the CD.  This has occurred despite the fact the woofers are closer to the mic than the CD,....... as in the drawing you posted.

 

It's 90 degrees of net phase lag on the lower frequency drivers that's induced for every order of the paired (high pass and low pass) crossover filters.  If you remove the crossover filters and measure again, the lower frequency drivers will lead the higher frequency drivers in an MEH.  You can see it in the spectrogram, step, group delay and phase plots for the loudspeaker. 

 

I think you're having trouble using these extremely high order filters having sharp cutoffs in MEHs, that absolutely don't need those kind of filters--because the reason for using them is simply not there as it is in multiple-aperture driver/horn loudspeakers.

 

I'm not sure what you're doing using "time of flight", but I don't listen to my MEHs (or any loudspeaker for that matter) using single ways, turning off the other drivers.

 

Chris

[Langston]

7 hours ago, Chris A said:

.., thus inducing a physical misalignment of the drivers' acoustic centers, with the higher frequency drivers typically leading the lower frequency drivers in terms of their time alignment.

 

This post, if understood, will save you many thousands of dollars and untold numbers of hours chasing audio nirvana with loudspeaker formats that are physically unable to achieve it. Spoiler: that includes every loudspeaker format other than MEH. Simply put, it is impossible to fix physical driver displacements with DSP. DSP can help immensely, but there is a domain that cannot be entered apart from a single acoustic wavefront containing the entire frequency spectrum of interest.

 

Just for fun:

 

All the best designers know this, but Tom Danley threw his whole audio career at the mercy of it, and for a long while it looked like the market was going to reject it. My audio background until recently was concert production and you still can't get a promoter at a major event to use the superior sounding Danley Jericho series in place of line arrays. But Tom finally found a great friend that threw in with him (Mike Hedden) and they still keep expanding the Jericho line (concert level MEH loudspeakers). They are still waiting for the market to catch up, but in the meantime installations (churches, performing arts facilities, etc.) and home audio have caught up and Danley Labs is doing just fine, thank you.

 

Here's a picture from an event I did with my line array system where Danley offered me their Jericho line FOR FREE if the promoter would accept it. Danley was going to deliver them to the site, help with setup, and take 'em back after the week long festival. I offered the promoter a significant discount. No, no, no. I pleaded with the guy and he knew that I knew more than any other company involved with the event. For fun, here's a couple of pictures:

 

 

[Chris A]

I didn't know about the customer resistance issue with the Jericho series (which really is a big step forward in commercial sound).  But I'd wondered if that was the reason why that series hadn't taken over that market sector. 

 

Tom is an experimentalist (thank goodness) that first builds and measures the phenomenon before he lets the theory inform him of the "why".  I think a lot of design nowadays is frozen in terms of innovation because we don't have more engineers like Tom.

_____________________________________________________

 

If you take an SH-50 apart and measure each "way", with and without crossover networks, you get to see some things that are not commonly known.  After doing that exercise with the SH-50 that I still own, I found some things that will never allow me to look at loudspeaker design the same way again.  I'll not steal Tom's thunder, but I have to say that it isn't what many people think it is (i.e., the "Synergy" crossover filter design).  It changed my outlook completely.

 

Chris

[gnarly]

47 minutes ago, Chris A said:

 

It's 90 degrees of net phase lag on the lower frequency drivers that's induced for every order of the paired (high pass and low pass) crossover filters.  If you remove the crossover filters and measure again, the lower frequency drivers will lead the higher frequency drivers in an MEH.  You can see it in the spectrogram, step, group delay and phase plots for the loudspeaker. 

 No, i don't think that's  it.  Phase lag induced for every order of the crossover filters applies to IIR crossovers.  I'm using complementary linear phase xovers which have the same phase lag (none), no matter what order used.

 

47 minutes ago, Chris A said:

 

I think you're having trouble using these extremely high order filters having sharp cutoffs in MEHs, that absolutely don't need those kind of filters--because the reason for using them is simply not there as it is in multiple-aperture driver/horn loudspeakers.

 

I'm not sure what you're doing using "time of flight", but I don't listen to my MEHs (or any loudspeaker for that matter) using single ways, turning off the other drivers.

 

Chris

 

Spectrograms, step, group delay are all just derivatives of mag and phase response....they all key off the same TOF measurements that place the woofer further away that the CD.  (REW:  TOF = Delay relative to loopback)

 

 

26 minutes ago, Chris A said: (This seems to have been edited away or something...???) My reply got hung up Saving...so this is a put back together attempt)

 

I think you're having trouble using these extremely high order filters having sharp cutoffs in MEHs, that absolutely don't need those kind of filters--because the reason for using them is simply not there as it is in multiple-aperture driver/horn loudspeakers.

 

 

The multiple aperture MEH's help mitigate the need for steeper xovers for sure.  But MEH's can improve further still with their use.  I have tuned them both ways and the steep acoustic rolloff of a large woofer crossing to a CD has been easier to get right using steep, for example.

 

I use outdoor polars to determine whether to use steep or shallow.  Working on such a project right now...

 

 

 

 

 

[Chris A]

20 minutes ago, gnarly said:

No, i don't think that's  it.  Phase lag induced for every order of the crossover filters applies to IIR crossovers.  I'm using complementary linear phase xovers which have the same phase lag (none), no matter what order used.

 

I do think you've got some issues with interpreting what you've got Mark, and your rationale for going the ways that you do with your MEH designs.  This is not new.  You may not know this, but perhaps half of the reason why I don't generally post at diyAudio anymore is due to history there with your approaches and methods (among a couple of others there). 

 

Without enumerating those things outright, I can say that they are fundamental differences in physics and psychophysics (i.e., how we hear).  I don't believe that your fundamental viewpoints on this subject (if there are, in fact, any) come very close to those that I've developed over time measuring, experimenting with, and developing MEHs.  I thought it easier to state this instead of trying to field your "questions that aren't questions". 

 

Thanks for being civil in your answers, but I don't believe that we're going to see eye-to-eye on this subject, i.e, MEH design and measurements, and I believe that you'll end up not thanking me if I start to enumerate the "why"--like you have beforehand on this forum, so I'll simply agree to disagree here.

 

I recommend that you continue your efforts at diyAudio for those that have been convinced of your efforts there.  I don't share your views on MEHs, and in fact have totally different viewpoints for your apparent issues that seem to drive your designs.  I think it's best if we leave it at that. 

 

Thanks again.

 

Chris

[gnarly]

On 8/24/2021 at 3:19 PM, Chris A said:

 

I do think you've got some issues with interpreting what you've got Mark, and your rationale for going the ways that you do with your MEH designs.  This is not new.  You may not know this, but perhaps half of the reason why I don't generally post at diyAudio anymore is due to history there with your approaches and methods (among a couple of others there). 

I'm sorry to hear i was part of the reason you quit posting at diyaudio. 

 

I don't understand why you would have a problem over approaches and methods that are different from yours.

Nobody's approach is the end all, be all. 

Tis healthy to discuss  various approaches and methods as long as respect and courtesy is provided to all parties presenting theirs.

On 8/24/2021 at 3:19 PM, Chris A said:

 

 I don't believe that your fundamental viewpoints on this subject (if there are, in fact, any) come very close to those that I've developed over time measuring, experimenting with, and developing MEHs.  I thought it easier to state this instead of trying to field your "questions that aren't questions". 

I hope you can realize how condescending that sounds..."(if there are, in fact, any)"....

 

You're correct about my "questions that aren't questions". 

I didn't ask any questions in this thread, other than what you mean by zeroth order? (which I'd still like to understand) 

 

I didn't disagree or present any alternatives to your first order tuning methods.

 

I  presented observations of measurements that don't reconcile with timings you discussed, and my viewpoint (theory) as to why TOF's measure as they do. 

 

 

On 8/24/2021 at 3:19 PM, Chris A said:

 

Thanks for being civil in your answers, but I don't believe that we're going to see eye-to-eye on this subject, i.e, MEH design and measurements, and I believe that you'll end up not thanking me if I start to enumerate the "why"--like you have beforehand on this forum, so I'll simply agree to disagree here.

 

Thank you for appreciating my civility, and for when you provide civility in return.

 

I need to ask :  Why is it important we see to eye-to-eye ? 

I don't care if you advocate a particular tuning strategy....why do you care if i advocate a particular different strategy?  What is there to agree or disagree about?  Both strategies work just fine.

 

(If i ever hear proof why my strategy doesn't work, i'll happily and very thankfully accept the instruction, knowing better audible results are forthcoming !! )

 

 

On 8/24/2021 at 3:19 PM, Chris A said:

 

I recommend that you continue your efforts at diyAudio for those that have been convinced of your efforts there.  I don't share your views on MEHs, and in fact have totally different viewpoints for your apparent issues that seem to drive your designs.  I think it's best if we leave it at that. 

 

Thanks again.

 

Chris

 

 

I'm sorry it appears you don't like it when folks do not adopt your viewpoints, approaches, and methods.

I don't know if i'll keep posting here or not, as you seem to be disinviting, directing me back to diyAudio.

Seems like a pity to me...I know you have a lot to offer, that there is a lot i could learn from you. 

And believe it or not, i have a lot to offer you too.

 

Anyway, later.. for now.

Sincerely,  Mark

 

 Peace...

[billybob]

"Convinced myself, I seek not to convince..."

[Chris A]

Mark, I think you missed two key words: "...fundamental differences...". I do acknowledge your apparent effervescent personality and your characteristic lack of being dissuaded. 

 

The problem is when you show (to me, at least) that you really don't understand the basic physics.  And in apparently not wishing to deal with that, choose instead to forge ahead rather forcefully voicing your own opinions--which I've found to be poorly calibrated.  You then typically seem to exercise what I consider to be questionable observational capabilities (i.e., lack of care in understanding the root causes of phenomena measured and sensed), and compound that with what I see to be poor judgment...both in informational threads that I have set up over time.  The problem is that you don't seem to relent until someone has to tell you this so explicitly, and for that I really do apologize that I have to be so explicit, because I find it distasteful that I have to say it here.  You've apparently followed me to this forum from diyAudio, itself looking like a predatory gesture.

 

It's a simple request: could you please start your own threads that you've set up for yourself and please not leapfrog into the threads where I've put in a fair amount of effort over time.  And that distinction applies just for you, Mark.  For all others here, I've apparently been able to address all manner of questions and comments. The alternative is that I will have to start ignoring all of your comments, Mark (which is a simple operation in the preferences dialogue) which I don't think is a good approach for either of us. 

 

I try hard to answer each person's questions on the forum, but when so many wide-of-the-mark questions and assertions keep occurring with regularity,  I have to say that I personally tire of trying to address them in the threads that I've spent a fair amount of time compiling--for others.

 

Chris

[gnarly]

6 hours ago, Chris A said:

You've apparently followed me to this forum from diyAudio, itself looking like a predatory gesture.

 

Yikes !  That's quite the imagination !

 

 

6 hours ago, Chris A said:

 

It's a simple request: could you please start your own threads that you've set up for yourself and please not leapfrog into the threads where I've put in a fair amount of effort over time. 

 

No problem at all, your request will be honored.

 

 

 

[billybob]

Just one of those things.

[prs]

more on this from 2007.

i'm nobody special, just a diy'r who plays with heil drivers at times and likes r and d.

https://www.researchgate.net/publication/289049366_Correction_of_crossover_phase_distortion_using_reversed_time_all-pass_IIR_filter

[Born2RockU]

On 5/3/2019 at 8:55 AM, Chris A said:

 

 

If you build a horn to have flat on-axis response from 400 Hz to even 5 kHz, you wind up with a non-controlled directivity horn ... ;  that's an approach that existed before DSP crossovers--i.e., the world that PWK had to live with until the very last few years of his life (John Dunlavy also planned to switch to DSP crossovers in his "Magnus" loudspeaker about the same time,...

 

 

 

Chris

"It may come as a surprise—this is giving away a trade secret—but when I design a loudspeaker, I first design it by looking at the step response. I find that by playing around with the crossover network while observing the step response in real time, any change I make is immediately available. " - DUNLAVY