Jump to content

Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers


Chris A

Recommended Posts

1 hour ago, glens said:

That's (frequency amplitude) not fair to the driver itself...  It don't look bad, but is largely horn.

 

You're apparently thinking about trying to use passive crossovers.  The idea that HF drivers don't require significant EQ is a notion that isn't true.  Perhaps you aren't used to controlled directivity horns. They all require significant EQ (IIRC, it was Don Keele that proved that in a white paper many decades ago). 

 

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 (like the K-400 series of midrange horns) that spill their vertical acoustic energy below 2.2 kHz, putting that energy on your floor and ceiling.  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, but the disorder that took his life overtook him before he could complete it.)

 

With a driver the caliber of a TAD TD-4002, I certainly wouldn't try to use passive crossovers.  It's much better to direct-couple the driver to the amplifier output using a DSP crossover and correct its minimum phase behavior upstream of the amplifier.  It takes merely 3 PEQs on the controlled directivity K-402 horn to bring the overall SPL response to within ±1.5 dB from 400 Hz to 20 kHz.  That's pretty spectacular.  I know of no other 2" driver that can do that plus also have the flat phase, extended HF response, outstanding impulse response, and extremely low modulation distortion of the TAD TD-4002.

 

Chris

  • Like 2
Link to comment
Share on other sites

Increased Perception of Bass:

 

One of the interesting changes that occurred when reworking the crossover filters on the Jubs to achieve the flat phase response posted above (and reproduced again here, below) is that my wife started to comment on how much more bass response there was.  While this may seem far fetched, Toole talks about this subjective effect in his book:

 

Quote

Craven and Gerzon (1992) stated that the phase distortion caused by the high-pass response is audible, even if the cutoff frequency is reduced to 5 Hz. They say it causes the bass to lack “tightness” and become “woolly.” Phase equalization of the bass...subjectively extends the effective bass response by the order of half an octave.

This effect did seem to occur in my listening room, even though the measured on-axis SPL response was flat or slightly rising to 17 Hz both beforehand and afterward.  Some of this effect might be explainable in part by the phase match-up with the two TH subs in the corners behind the Jubs...but note that the crossover frequency between the subs and the Jub bass bins is 30 Hz.  

 

139925984_TADTD-4002Jubilee(RedTrace)vs.DanleySH-50(GreenTrace)Phase.thumb.jpg.71246150ee40a01836a4ddd6e9bbd4eb.jpg

TAD TD-4002 Jubilee Impulse Spectrogram (Fractional Order Crossover Filters).jpg

 

Another assessment is that these "quasi-linear-phase responses" have further improved upon the overall subjective sound quality as I reported at the top of this thread.  It seems that phase response flatness really does have a large part to play in the hi-fi reproduction of music.  I believe that a large part of the "little monitors on stands" or full-range drivers (a misnomer) is wrapped up with this effect.  When you flatten the phase with, for instance, Jubilees, a center K-402-MEH and surround AMT-1/Belle bass bins (in a 5.1 array), the effect is pronounced when playing the best acoustic recordings that I own (whose EQ hasn't been substantially altered from its as-recorded state). 

 

All this brings up a point about loudspeakers and crossovers: without measuring and correcting the in-room response of the loudspeakers, this effect cannot be experienced, i.e., passive crossovers are pretty much ruled out. Those that intend to achieve this level of performance via mixing and matching existing horns, drivers, and passive crossovers are pretty much locked into a resulting sound that one can identify with the sound of loudspeakers instead of "the real thing". 

 

Chris

  • Like 2
Link to comment
Share on other sites

  • 3 weeks later...

Elimination of Harshness:

 

Last night, my wife mentioned that she could listen to just music instead of watching video (Netflix, Prime, and BDs), and that she is no longer afraid of harsh sounds from the loudspeakers.  Recall that I recently further flattened not only the frequency response of the surround 5.1 setup, but also flattened the phase response of all five surround loudspeakers and two TH subwoofers. 

 

Additionally, any time that we play a recording that hasn't been multitrack recorded in recording booths, i.e., symphonic, solo concertos and sonatas, choral pieces, etc., the sound that we are treated to is, well...like we're sitting in the audience of a large music hall.  It is actually difficult to stop listening to these recordings once they're turned on.  This is a completely new experience...and is the central subject of this thread.

 

Recently I've seen people asking how to reduce the "fatigue factor" and the harshness of sound in various home theater threads.  I believe that I now know exactly how to achieve that--but the price of admission is bi/tri-amping and DSP crossovers to flatten not only the frequency response, but also the phase response of all of the loudspeakers in the array.  In stereo mode, the effect is actually heightened.

 

Chris

  • Like 2
Link to comment
Share on other sites

Very interesting thread. Ive got some thoughts but no time to get them all out currently. I do have a couple questions: you have FDW enabled on these graphs, how many cycles? And how different do your phase graphs look unsmoothed? Just pondering about the effect of room interactions.

Link to comment
Share on other sites

3 hours ago, Dkfan9 said:

...you have FDW enabled on these graphs...

Only the first two measurements, after which I turned off FDW within REW.

 

3 hours ago, Dkfan9 said:

...how many cycles?

The REW default value for those first two measurements (only). FDW was inadvertently on due to a re-install of REW last fall after having to use a much earlier version of REW to open a measurement sent to me from another user not using the latest version of REW. 

 

All FDW does is smooth the raw measurements to ~1/48th octave.  I didn't need FDW to be on, so I turned it off (there is a really interesting discussion in the Home Theater Shack forum between Bob Katz--the noted mastering engineer who requested the capability so REW could mimic Acourate...and John Mulcahy on why FDW was implemented within REW.  It's a good discussion that I recommend highly.

 

The much more interesting measurements however  are found on page 3 of this thread.  Those measurements didn't have FDW turned on. 

 

3 hours ago, Dkfan9 said:

And how different do your phase graphs look unsmoothed?

Pretty much like the smoothed measurements.  I'm not sure why you're worried about smoothing.  If it were me, I think that I'd be asking questions about where the microphone was (1 m in front of the loudspeaker) and how much absorption was used to capture such good phase information in-room (several inches thickness of fuzzy blankets and comforters the long axis is the width from side-to-side, and 1 m depth to the loudspeaker), not really the smoothing used on the data after acquisition.

 

 

3 hours ago, Dkfan9 said:

...Just pondering about the effect of room interactions.

My listening room photo and reverberation time plots can be found in my profile under the "About Me" tab.

 

Chris

  • Like 1
Link to comment
Share on other sites

  • 2 weeks later...

I'd have to use something like JRiver in order to get enough taps to do anything further to flatten phase.  The problem is that I'm not currently running a PC-based stereo setup, rather a pre/pro-based 5.1 using dedicated DSP crossovers.  At some time in the future, I'll see if I can spring for another workstation with sufficient compute horsepower to run JRiver in 5.1 mode (as I read it supports, but I'm not sure how well it works, how much compute capability it requires, or how many users actually use that capability), and output through HDMI to the pre/pro. 

 

Right now, both Jubilees are measuring within 45 degrees of flat phase from 100 Hz to 20 kHz, and group delay is similarly improved as seen above in the plots above.  Considering that I listened to the same Jubilees with the same TAD drivers for over a decade with over a thousand of degrees of phase growth from HF to LF, I suppose that ±45 degrees is "good enough for the time being".  I wish you could hear them.  It's been a real pleasure listening to them every day since I made the crossover change around the first of May.

 

1314114214_TAD-JubileePhaseFlattening(Dec22014BlueMay12019Orange).thumb.jpg.1e88f7b0828d29e8046b5e372ef755e3.jpg

 

Chris

  • Like 3
Link to comment
Share on other sites

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

 

1 hour ago, Arcgotic said:

Is there a 'secret' to get that nice phase?

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

  • Like 5
  • Thanks 1
Link to comment
Share on other sites

3 minutes ago, 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 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, you make it sound easy  ;)

 

Link to comment
Share on other sites

Once you see how it works...and do it a couple of times...it's like riding a bicycle. 🚲  You'll never go back to using canned crossover filters either once you hear the difference.  The sound quality difference isn't close.

 

This little process works fairly fast if you're in-room and can make observations on measurements and make immediate adjustments, then rerun the sweep.  You'll have to check that you've got the driver polarity right ( :laugh: ).  Make sure that there's plenty of absorption on the floor and nearby walls before you start the process--at least a couple inches thick and several feet wide, all the way to the microphone from the loudspeaker under test/adjustment.

 

Chris

  • Like 3
Link to comment
Share on other sites

5 minutes ago, Chris A said:

Once you see how it works...and do it a couple of times...it's like riding a bicycle. 🚲  You'll never will go back to using canned crossover filters either once you hear the difference.  The sound quality difference isn't close.

 

This little process works fairly fast if you're in-room and can make observations on measurements and make immediate adjustments, then rerun the sweep.  You'll have to check that you've got the driver polarity right ( :laugh: ).  Make sure that there's plenty of absorption on the floor and nearby walls before you start the process--at least a couple inches thick and several feet wide, all the way to the microphone from the loudspeaker under test/adjustment.

 

Chris

I'm looking forward to loading the latest revision then walking though the steps again to see the effect on each change.

Once i get this other issue straightened out, I'll be all over it...   

-e

Link to comment
Share on other sites

Yes...don't use filters that have men's last names attached to them, or the word "order" in their names--those all introduce phase shifts.  PEQs don't introduce systemic phase shifts, only local shifts (i.e., minimum phase).

 

Chris

 

 

  • Like 1
Link to comment
Share on other sites

Here is the above Jubilee SPL response with the overall, bass bin, and TAD channels overlaid to demonstrate the ~10 dB/octave "fractional order" crossover steepness due to the drivers' responses themselves--no "named" crossover filters are used:

 

328842806_JubileewithDanley-StyleCrossoversSPLResponse.thumb.jpg.d571dcdab35092123229dd2f9b7f3c8f.jpg

  • Like 3
Link to comment
Share on other sites

On 3/11/2019 at 7:55 PM, Don Richard said:

I hear (no pun intended) the pro sound guys refer to "coherence". This refers to how the output of a loudspeaker compares with the input signal. When the phase and the amplitude are flat, and impulse response is clean, the speaker sounds more lifelike and measures better. This seems to describe what you report hearing and observing. A coherent speaker is better able to deal with noise and acoustic interferences, and will not suffer from polar pattern degradation as badly when the wind is blowing.

 

https://www.prosoundweb.com/channels/live-sound/tech-topic-coherence-reverberation/

As a test procedure, Tom Danley has often referred to recording what comes out of a loudspeaker when playing a test recording, then seeing how many generations it takes of doing this over and over to deteriorate,  via listening tests. In other words, it's like making a Photocopy of a Photocopy of a Photocopy and so on until the TEXT is no longer legible. Pretty good process, if you ask me.

Link to comment
Share on other sites

On 6/5/2019 at 12:30 PM, 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

FYI, I did a brief recorded interview with Tom Danley when I met him a few years ago. He told me he uses LSP CAD to design his SH-50 (and all other speakers) PASSIVE Networks and rePhase (Free Software) for the final touch in a Mini DSP box that can use it. In my Living Room corners, these truly SMALL horns are flat to 40 Hz. where the two subs (soon to be 3) take over. These plots are from Tom Danley's own SH-50's, not mine, with no subs.

sh50_in_livingroom_with_and_without_fir.png

SH-50guts.jpg

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...