Step Response: Does It Really Matter?

[ctrl]

I have done much tests with ribbon tweeter at addon on my speakers and use DSP crossover. i notice step response show best if the tweeter and the mid/bass have fewest delay together. when in stereo change the delay of both tweeter is set to 0.1 ms it have lots diffrent sound. with mono sound i hear not so much diffrence. magnitude and phase are in correlation, and influence Magnitude at a specific frequency too. both have together influence of a sound. there can not say when the result of a step response look faster it is only because FR diffrence.

It is precisely because of these "conclusions" that it is important to deal with the basics.

Let's just look at an "ideal" loudspeaker with a fourth order crossover at 2.5kHz. So something like a standard 2-way loudspeaker.
So we get a linear frequency response over the entire bandwidth of the "ideal" drivers used.


Your procedure is now to look at the step response and determine the time peak-to-peak distance of the drivers:

Your conclusion: Wow, 0.4ms difference between tweeter and woofer peaks, the woofer is extremely slow, I can clearly hear a delay of 0.1ms** between tweeter and woofer.

** We will look at this in more detail below

However, the group delay is frequency-dependent and cannot be determined in this way.
If we consider the group delay of our standard 2-way loudspeaker we get the following:

In the above-mentioned study, it was again clearly shown that, according to the current status, it is not possible to perceive the increase in group delay caused by the allpass filter.

With for example 0.22ms group delay around 1.5kHz of our 2-way, according to the study, the perceptibility threshold in the frequency range 500-4000Hz is not exceeded by far.

Now I can already hear you saying, "That's crap, my test with the tweeter delayed by only 0.1ms clearly shows that even 0.1ms are audible without any problems. You're full of sh..." - or something like that

What does your experiment?
The 0.1ms delay means that you move the tweeter 3.5cm from the woofer (on listening axis). Of course this also leads to a phase shift.

For our 2-way speaker with the [email protected] this means a phase shift of 90°:
2500Hz --> 1/2500 s = 4.10^-4s per oscillation period (360°) --> 0.1ms delay = 1*10-4ms is 1/4 oscillation period --> 90°

With this you change the sound addition of the high and low-pass filter slopes.
For our 2-way loudspeaker this results in the following frequency response:

It would be astonishing if this frequency response dip (with a small Q, i.e. wide frequency range) would not be perceptible.

If our loudspeaker were driven with a third-order crossover and as usual one driver with reversed polarity, the effect would be even more spectacular and would lead to a perfect zero in the FR.

You see, your attempt has nothing at all to do with the audibility of group delay or "fast and slow bass".

It only shows that when a driver is delayed on one side, the frequency response changes significantly and the crossover becomes unusable... well, again that's why it's important to familiarize yourself with the basics.

 

[bennybbbx]

I have done RTA with sinus measure tone on the Kali LP6(which have large changes in distortion depend on level). a big change happen on level 61 db 3 rd 0.15% distortion 114 deg and on 63 db 0.45% 131 deg. I measure more often and long time degree change 1 or 2 only when volume stay same. the distortion can hear on the measuretone too.
this is really not loud measure. you can see on the noise floor on Bass too that it is not loud. when go to 70 db distortion go to 0.9%

131 deg - 114 deg is 17 degree. when calculate the delay that happen at 17 degree i am not 100% sure. I think can calculate as this. 1/300 = 0.0033 sec = 3.3 ms or 3300 microseconds 1 degree is 3300 /360 = 9.1 microseconds. now take the 17 degree and calc delay diffrence. because it can measure inaccuratc we calc only 10 degree. this are 91 microseconds delay. and because left and right speaker do on stereo diffrent movements so distortions get diffrent. and 90 microseconds delay are much hearable when can hear ITD

the question is if the harmonic is delayed more than 90 microseconds how much is the main tone(300 hz) delayed. here are screenshots. i can also upload a video that show that measure is repeatable and give same results

 

[bennybbbx]

It is precisely because of these "conclusions" that it is important to deal with the basics.

Let's just look at an "ideal" loudspeaker with a fourth order crossover at 2.5kHz. So something like a standard 2-way loudspeaker.
So we get a linear frequency response over the entire bandwidth of the "ideal" drivers used.
View attachment 145956

Your procedure is now to look at the step response and determine the time peak-to-peak distance of the drivers:
View attachment 145957

Your conclusion: Wow, 0.4ms difference between tweeter and woofer peaks, the woofer is extremely slow, I can clearly hear a delay of 0.1ms** between tweeter and woofer.

** We will look at this in more detail below

However, the group delay is frequency-dependent and cannot be determined in this way.
If we consider the group delay of our standard 2-way loudspeaker we get the following:
View attachment 145958

In the above-mentioned study, it was again clearly shown that, according to the current status, it is not possible to perceive the increase in group delay caused by the allpass filter.

With for example 0.22ms group delay around 1.5kHz of our 2-way, according to the study, the perceptibility threshold in the frequency range 500-4000Hz is not exceeded by far.

Now I can already hear you saying, "That's crap, my test with the tweeter delayed by only 0.1ms clearly shows that even 0.1ms are audible without any problems. You're full of sh..." - or something like that

What does your experiment?
The 0.1ms delay means that you move the tweeter 3.5cm from the woofer (on listening axis). Of course this also leads to a phase shift.

For our 2-way speaker with the [email protected] this means a phase shift of 90°:
2500Hz --> 1/2500 s = 4.10^-4s per oscillation period (360°) --> 0.1ms delay = 1*10-4ms is 1/4 oscillation period --> 90°

With this you change the sound addition of the high and low-pass filter slopes.
For our 2-way loudspeaker this results in the following frequency response:
View attachment 145961

It would be astonishing if this frequency response dip (with a small Q, i.e. wide frequency range) would not be perceptible.

If our loudspeaker were driven with a third-order crossover with, as usual, a reversed driver, the effect would be even more spectacular and would lead to a perfect zero in the FR.

You see, your attempt has nothing at all to do with the audibility of group delay or "fast and slow bass".

It only shows that when a driver is delayed on one side, the frequency response changes significantly and the crossover becomes unusable... well, again that's why it's important to familiarize yourself with the basics.

No, i mean the beginning point of the woofer at 40% and the speed it raise to 100 and fall to 40 %. better is measure the impulses 1 cm away from woofer. this i have done. so tweeter have no influence. see here between the arrows. and the slower speaker i think have a slower woofer . the top show a fast speaker and button a slow speaker. the example is not measure very near at 1 cm you can see on FR

and are you sure the curce for a ideal speaker is really right ?. wy headphones have then faster step response ?

and here how a headphone step response look. this look different in compare to your ideal speaker step response. when play the impulse then a ideal speaker is more simular to headphone. think about car move. the step response say move from position 1 to position 2 and back. now when plot the time need from position 1 to position 2 a slow car need more time. so the rise and fall time are slower. compare with headpone.

edit: please look at the diagram about step response in wikipedia https://en.wikipedia.org/wiki/Step_response this picture i mean. the time to reach position 2 is called settling time from them https://en.wikipedia.org/wiki/Step_...uracy_settling_time_measurements_figure_1.png

Edit 2: here i have the function explain from mathlab what the step response function return.

https://de.mathworks.com/help/control/ref/lti.stepinfo.html

S

= stepinfo(sys)computes the step-response characteristics for a dynamic system model sys. The function returns the characteristics in a structure containing the fields:

• RiseTime — Time it takes for the response to rise from 10% to 90% of the steady-state response.
• SettlingTime — Time it takes for the error e(t) = |y(t) – yfinal| between the response y(t) and the steady-state response yfinal to fall below 2% of the peak value of e(t).
• SettlingMin — Minimum value of y(t) once the response has risen.
• SettlingMax — Maximum value of y(t) once the response has risen.
• Overshoot — Percentage overshoot, relative to yfinal.
• Undershoot — Percentage undershoot.
• Peak — Peak absolute value of y(t)
• PeakTime — Time at which the peak value occurs.

 

[ctrl]

No, i mean the beginning point of the woofer at 40% and the speed it raise to 100 and fall to 40 %. better is measure the impulses 1 cm away from woofer. this i have done. so tweeter have no influence. see here between the arrows. and the slower speaker i think have a slower woofer . the top show a fast speaker and button a slow speaker. the example is not measure very near at 1 cm you can see on FR

You are going around in circles with your argument. As far as I know, we had already discussed this once at the beginning of the thread.

The bandwidth of the driver's frequency range significantly determines the appearance of the step response. This has been told to you several times now... so it is important to know the basics to not draw wrong conclusions.

Our ideal driver is always equally fast, let's see how your "40% theory" estimates the "speed" of the driver at different frequency bandwidths.

1. Bandwidth: 25-2500Hz


2. Bandwidth: 250-2500Hz


3. Bandwidth: 250-5000Hz


The first driver is total crap, very slow. The second one is okay-isch, but the third one is lightning fast... Wait! Oh no it's the same ideal fast driver with different frequency bandwidths.

and are you sure the curce for a ideal speaker is really right ?. wy headphones have then faster step response ?

Yes I'm sure and because you measure your headphones in the worst way possible. You completely "lose" the low frequency content of the headphones this way (the low frequency range is 10-20dB quieter than the rest in this way) and thus it looks extremely fast to you... so the basics are important to not draw such wrong conclusions.

You have to create a situation for a measurement that is as close as possible to wearing the headphones.
For simple measurements, it doesn't have to be a Gras 45CA like Amir uses. A small wooden plate covered with felt and a small hole for inserting the microphone is sufficient. This covers the headphone cup during the measurement.

Then the step response of a headphone will again look like what is usual for a full range loudspeaker.
DT990 (bass strong) DT660 (bass weak)


Please, please read up on the basics and experiment a bit with the many tools that are available.
It's no fun to keep putting the same things right over and over again so that beginners don't get misguided.

 

[bennybbbx]

You are going around in circles with your argument. As far as I know, we had already discussed this once at the beginning of the thread.

The bandwidth of the driver's frequency range significantly determines the appearance of the step response. This has been told to you several times now... so it is important to know the basics to not draw wrong conclusions.

Our ideal driver is always equally fast, let's see how your "40% theory" estimates the "speed" of the driver at different frequency bandwidths.

1. Bandwidth: 25-2500Hz
View attachment 146038

2. Bandwidth: 250-2500Hz
View attachment 146039

3. Bandwidth: 250-5000Hz
View attachment 146041

The first driver is total crap, very slow. The second one is okay-isch, but the third one is lightning fast... Wait! Oh no it's the same ideal fast driver with different frequency bandwidths.



Yes I'm sure and because you measure your headphones in the worst way possible. You completely "lose" the low frequency content of the headphones this way (the low frequency range is 10-20dB quieter than the rest in this way) and thus it looks extremely fast to you... so the basics are important to not draw such wrong conclusions.

You have to create a situation for a measurement that is as close as possible to wearing the headphones.
For simple measurements, it doesn't have to be a Gras 45CA like Amir uses. A small wooden plate covered with felt and a small hole for inserting the microphone is sufficient. This covers the headphone cup during the measurement.

Then the step response of a headphone will again look like what is usual for a full range loudspeaker.
DT990 (bass strong) View attachment 146034 DT660 (bass weak) View attachment 146035


Please, please read up on the basics and experiment a bit with the many tools that are available.
It's no fun to keep putting the same things right over and over again so that beginners don't get misguided.

of course EQ do phase shifting and change the step response. but all in all the step response should show the rise time
maybe it is more clear when you look at the step response of electrical hydraulic systems. see page 9

https://www.sunhydraulics.com/sites...ech_resources/rel-Prop_terms-definitionsN.pdf

Step Response
Defined as, the time it takes to reach the expected output pressure or flow from when the input
command is applied; commonly evaluated at 100%.

maybe it is more clear what step response should show. on hydraulic there is no EQ use. that it look diffrent on spekaers depend also on the crossover frequency because if you have non linear filters it shift the time on speakers too

in this link and page 9 is a diagram with 1 hz and 10 hz. you can see the hydraulic reach 10 hz, but it delay alot in compare to 1 hz

a speaker is of course faster as hydraulic but problems are the same. of course the best step response of a speaker can measure when use no crossover and no EQ or it is phaselinear

EDIT; here is a video that explain step response.

this video show what the step response is in a example with a drone controller. it is at sec 58. the link should jump to it

there can see how filters influnece step response. its of course the task of speaker developer to choose a good crossover for the speaker so it do not delay much

 

[Frgirard]

Like the k+h o500 https://gearspace.com/board/showpost.php?p=9633758&postcount=9

 

[ctrl]

of course EQ do phase shifting and change the step response. but all in all the step response should show the rise time

... and the rise time is bandwidth dependent.

Two equivalent woofers, both play down to 50Hz, but one woofer is crossed over to the tweeter at 5kHz (e.g. a Dynaudio 2-way), the other woofer is crossed over to the tweeter at 2kHz (e.g. Neumann 2-way).
Then the rise time (of the woofers within the step response measurement of the speakers) says nothing about how "fast" the individual woofers are.

You will get a different group delay for the speakers with the same filter order of the two crossovers.
The speaker with the higher crossover frequency has a lower group delay.

If the frequency responses of the two woofers would be absolutely congruent in a measurement and the step responses are not congruent after correction of a possible delay, then you can start to interpret the step responses regarding transient and decay behavior.

 

[bennybbbx]

you

... and the rise time is bandwidth dependent.

Two equivalent woofers, both play down to 50Hz, but one woofer is crossed over to the tweeter at 5kHz (e.g. a Dynaudio 2-way), the other woofer is crossed over to the tweeter at 2kHz (e.g. Neumann 2-way).
Then the rise time (of the woofers within the step response measurement of the speakers) says nothing about how "fast" the individual woofers are.

You will get a different group delay for the speakers with the same filter order of the two crossovers.
The speaker with the higher crossover frequency has a lower group delay.

If the frequency responses of the two woofers would be absolutely congruent in a measurement and the step responses are not congruent after correction of a possible delay, then you can start to interpret the step responses regarding transient and decay behavior.

in the compare with Kali LP the crossover is very simular. 1.5 khz on kali and JBL 104 BT is 1.7 khz https://jblpro.com/en/site_elements/jbl-professional-104-spec-sheet

so very simular. but the diffrence in step response is much as you can see and the 104 look much faster. and dont forget my step responses from the iloud MTM. It have crossover of 3.3 khz. so with your theory, wy the iloud mtm is not faster as the JBL in step response ?. also the iloud mtm have much more bass as the lp6. iloud step response look much faster as the LP6 and it sound much wider. so how is this possible

also a speaker is not bandlimit at crossover frequency. it is a 2 way full range speaker. step response also show overshot undershoot

Edit: here is step response of the 104 without and with EQ (bass boost 50 hz 12 db) so it have more and lower bass as the lp6. you see the step response is near same and also with bass boost much faster as with the LP6
the measures are in this post
https://www.audiosciencereview.com/...oes-it-really-matter.1999/page-15#post-842795

 

[ctrl]

this video show what the step response is in a example with a drone controller. it is at sec 58. the link should jump to it

...and now think about what the big difference is to a woofer or even more especially a woofer with crossover and why under these circumstances the rise time is not a reliable statement about the "speed" of a woofer, compared to a woofer with different bandwidth (there you have it, i said it).

in the compare with Kali LP the crossover is very simular. 1.5 khz on kali and JBL 104 BT is 1.7 khz https://jblpro.com/en/site_elements/jbl-professional-104-spec-sheet
so very simular. but the diffrence in step response is much as you can see and the 104 look much faster.

I hope that the step response is very different, because the speakers are also very different (FR, f3, f6,...).

Source: https://pierreaubert.github.io/spinorama/index.html

 

[j_j]

Ok, time to make a couple of artificial signals with substantial variation in group delay, I can tell.

Then you can listen for yourselves.

Will add link in a few.

https://we.tl/t-iS5Ghmmdyk

There are two files, am.wav and fm.wav Keep the level down until you're sure of your listening level set, these are about 6dB down from peak sine wave level.

You tell me these two do not sound different. What is the difference in these two monophonic signals?

Phase shift. NOTHING else. Phase shift. The two power spectra are identical to 16 bit accuracy.

The center frequency is 700 Hz.

If you claim phase shift doesn't matter there at any level, well, have a nice day, um hm.

 

[j_j]

of course EQ do phase shifting and change the step response. but all in all the step response should show the rise time
maybe it is more clear when you look at the step response of electrical hydraulic systems. see page 9

Not all EQ's change phase response. Some add pure delay.

 

[bennybbbx]

...and now think about what the big difference is to a woofer or even more especially a woofer with crossover and why under these circumstances the rise time is not a reliable statement about the "speed" of a woofer, compared to a woofer with different bandwidth (there you have it, i said it).



I hope that the step response is very different, because the speakers are also very different (FR, f3, f6,...).

View attachment 146079

View attachment 146080
Source: https://pierreaubert.github.io/spinorama/index.html

have you look the video step response control design ?. there can see on sec 321 the rise time depend not on better bass . a bass filter(highpass) have only influence of decay time. I have boost with EQ the bass of the 104. see my EQ version that it go lower as the LP6. but no change in rise time. only decay time is a little longer
The influence of rise time happen because of the crossover lowpass.

Of course i agree that filters change the rise time and decay time. My opinion is that the rise time depend on filters AND speaker speed. both together give the results. But the slowness in rise time from filter is not much that the rise time slowdown of the filter have not important influence. here is the step response of a 1.5 khz 12 db Lowpass filter (LP setting in REW then choose export impiulse response of Filters). you can see 90% of level is reach at 0.3 ms. so the 1.5 khz 12 db filter delay the rise time about 0.3 ms or 300 microseconds on on the 6 db crossover the rise time is delay 0.18 ms or 184 microsecconds.

If you want more precise step responses you can subtract the rise time. calculate it depend on crossover freq and order. on speakers that have crossover higher as 1.5 khz (most speakers have around 2.5 khz) the rise time slowdown of filter is lower

this is the step response of a 100 HZ 6 db LP filter. you see the rise time stay fast. a speaker that have much or less bass can have the same rise time. the rise time and decay time depend not bass lowness. in real world a bass/mid that go lower is often slower, because it need move more air and membrane is heavier. their need also do hear tests which speed is need for people that hear ITD that stereo width is correct.

you can easy compare this with the drone example in the video that explain step response. the step input tell the drone to move fastest way to 100% . it doesnt matter if the drone is able to climb not slow. same is with speakers. and of course when not need any filter and measure speaker is best. but for most speakertesting it is not possible to remove the filter.

Or speaker developer should add their bass/mid step response without any Filter. did you then confirm that without any filter the step response shows which speaker is fastest ?

 

[j_j]

Benny, you need to study how bandwidth and rise time are related. Please?

Also, understand that a sum of transfer functions added together can become a very good impulse or step response, even if part of one seems somehow slow.

 

[bennybbbx]

Ok, time to make a couple of artificial signals with substantial variation in group delay, I can tell.

Then you can listen for yourselves.

Will add link in a few.

https://we.tl/t-iS5Ghmmdyk

There are two files, am.wav and fm.wav Keep the level down until you're sure of your listening level set, these are about 6dB down from peak sine wave level.

You tell me these two do not sound different. What is the difference in these two monophonic signals?

Phase shift. NOTHING else. Phase shift. The two power spectra are identical to 16 bit accuracy.

The center frequency is 700 Hz.

If you claim phase shift doesn't matter there at any level, well, have a nice day, um hm.

on monophonic signals the changes can not good hear. try this phase shifts in stereo signal. everybody can use my stereo example here. https://www.audiosciencereview.com/...tep-response-and-audio-record-examples.19812/
i choose guitar because on guitar their can very easy phase or crossover problems hear. their happen nasal sound on some frequency or strange stereo width when speaker is not good

 

[ctrl]

have you look the video step response control design ?. there can see on sec 321 the rise time depend not on better bass . a bass filter(highpass) have only influence of decay time.

Really, I wonder what Post#384 or Post#131!! was supposed to point out?

You said "but the diffrence in step response is much as you can see and the 104 look much faster. "
... and my answer was that I expect a big difference in step response, since the speakers differ significantly in their bandwidth.

Therefore, your 40% rule given in Post#383 as a measure of the "speed of the bass"...

No, i mean the beginning point of the woofer at 40% and the speed it raise to 100 and fall to 40 %. better is measure the impulses 1 cm away from woofer. this i have done. so tweeter have no influence. see here between the arrows. and the slower speaker i think have a slower woofer . the top show a fast speaker and button a slow speaker.

...is completely meaningless - except under very specific circumstances, which I gave in Post#387 or Post#131.

That's why your video example that you keep referencing is not applicable. The drone in the video reaches a "steady state".

If you apply a DC voltage to a driver and then measure the membrane displacement, the driver will reach a constant displacement after some time (but this brings new problems regarding the comparability of drivers).
This would then be comparable to your drone example video.

My opinion is that the rise time depend on filters AND speaker speed.

That took you a long time.

I mean to accept that the frequency bandwidth significantly determines the step response.

As far as the term "speaker speed" is concerned, the matter is a bit more complex.

You have compared loudspeakers with 4'' and 8'' woofer.
What happens when the sound pressure is increased in both speakers and the 4'' driver clearly leaves the linear excursion range of the voice coil at 60Hz, for example, and the 8'' driver hardly has to excursion at all? Which is then the "fast" driver?

 

[bennybbbx]

Really, I wonder what Post#384 or Post#131!! was supposed to point out?

You said "but the diffrence in step response is much as you can see and the 104 look much faster. "
... and my answer was that I expect a big difference in step response, since the speakers differ significantly in their bandwidth.

Therefore, your 40% rule given in Post#383 as a measure of the "speed of the bass"...

...is completely meaningless - except under very specific circumstances, which I gave in Post#387 or Post#131.

That's why your video example that you keep referencing is not applicable. The drone in the video reaches a "steady state".

i am over 50 years old and when i go to school we have very few english learn. I use english not often. so it is not easy to explain in english. in short the diffrence between step response and impulse response is that the step response(on hydraluic or drone example too) show rise time and ignore the fall time. impulse response measure give fall time. but it is not so important on speaker. because i see only when rise time is fast then fall time is fast too. in my example i have not draw an arrow to the peak. i thougt it is clear the time should look from the left arrow to peak when i write rise time. here i have with an arrow on the peak too.

I have measurements done with the kali and ribbon tweeter with crossover at 1 khz. step response look much faster, when see step response of lowpass 1 khz have more rise time as 1.5 khz. but because the ribbon is faster step response look faster

the ribbon (rbt 35 s monacor) have no problems with the low frequency, it have only Problem with a distortion at around 1.5 khz. with and without foam under it, ot can hear much on some sounds. So it is only to show that the kali step response get much faster and sound better in stereo when reduce the crossover from 1500 hz to 1000 hz that the kali woofer need not do so high frequency

I have the ribbon in this picture more inside because then angle to ear measure microphone is better.
Edit: use same time scale as other step response

Edit 2: and here is the rise time of the Lowpass 12 db 1 khz. you see is slower as with 1.5 khz. and because the ribbon is much faster the slower rise time(phase shift) of the filter have not much influence and with ribbon it look faster rise time

edit 3: here is a headphone measure of a sony headphone with much bass. 1 measure is from 20 hz. other measure from 80 hz. step response is near same. this example show that it doesnt matter for step response(i look for speed) if speaker can play low frequency or not

 

[j_j]

on monophonic signals the changes can not good hear. try this phase shifts in stereo signal. everybody can use my stereo example here. https://www.audiosciencereview.com/...tep-response-and-audio-record-examples.19812/
i choose guitar because on guitar their can very easy phase or crossover problems hear. their happen nasal sound on some frequency or strange stereo width when speaker is not good

Since this is a blatantly obvious example, what do you mean by "not good hear"? Blatant, obvious differences are just that, obvious differences. In a monophonic signal.

Did you even listen, or are you too busy arguing?

 

[tuga]

I found these transient response measurements of the B&W 801F and the Ref 105 (both 3-way sealed) in old RDS magazines.
Is the information they provide meaningful in any way? I wonder how the measurement correlates with listening at 60Hz and 100Hz.

 

[bennybbbx]

I found these transient response measurements of the B&W 801F and the Ref 105 (both 3-way sealed) in old RDS magazines.
Is the information they provide meaningful in any way? I wonder how the measurement correlates with listening at 60Hz and 100Hz.

yes, such tests are the best to show undershoot and overshoot and of course precision of a system and the truth. the only need show more the wave cycles at higher frequency.the 60 hz and 100 hz in your post look as they use a burst. i post such wave screenshots. https://www.audiosciencereview.com/...ct-canton-look-much-better.17854/#post-581235 there can too see the LP6 look very worse. this tests are very old. some here told me to use rew to test the speed and presion of a speaker. so i do this. but there is no measure with rew possible that the experts here accept that it show a speaker is fast and precise or not for ITD.

a burst input

look as this

 

[bennybbbx]

Since this is a blatantly obvious example, what do you mean by "not good hear"? Blatant, obvious differences are just that, obvious differences. In a monophonic signal.

Did you even listen, or are you too busy arguing?

yes i have listen. i want wait what others hear. others do not report what they hear. now i write. I hear in the AM tone huge wobbling(i think it wobbles around 5 hz, but i am not sure. can also be 10 hz) with headphone and my jbl same strong. the fm.wav is more clean but it wobbles too clear noticable. maybe you reduce the phase shifts to much lower. in last post i want say. when this happen with stereo signals the diffrences are much much more hear because the wobble of left and right speaker independent give the room feeling of a big reflective room