Phase Distortion ABX testing

[thewas]

Would be interesting to see how the response is without filters, as for example filling a dip with a minimum phase filter adds ringing, on the other hand a room mode (peak) which mainly shows minimum phase behaviour is preferably corrected with a minimum phase filter as they cancel each other out.

 

[dasdoing]

as for example filling a dip with a minimum phase filter adds ringing, on the other hand a room mode (peak) which mainly shows minimum phase behaviour is preferably corrected with a minimum phase filter as they cancel each other out.

the software used uses both types of filters for the correction

Would be interesting to see how the response is without filters

 

[andreasmaaan]

the software used uses both types of filters for the correction



View attachment 109868View attachment 109869

@dasdoing can I suggest you use a standard scale when posting these responses? A vertical axis spanning 50dB is the convention. It's very difficult to make sense of the graphs when they span such a wide range.

 

[dasdoing]

@dasdoing can I suggest you use a standard scale when posting these responses? A vertical axis spanning 50dB is the convention. It's very difficult to make sense of the graphs when they span such a wide range.

ok, here are the 3, var smoothing, aligned at 1000Hz

and with 5 cycle FDW

5 cycle FDW of the the two corected only

OBS: no SPL calibration
for the info: the goal of this filters was flat direct sound

 

[thewas]

the software used uses both types of filters for the correction

But we don't know how it generates the filter coefficients, if you would want to do a minimum phase and linear phase comparison you should manually create the same amplitude filters for example in Rephase.
But even then in your above plots you can see that the problems with the minimum phase filters mainly occur like I had predicted above when they are used to fill dips.

for the info: the goal of this filters was flat direct sound

In my past experience FDW used is not enough for flat direct sound in usual room sizes, thats why I personally rather use anechoic measurements above transition frequency and only below (where we don't just want flat direct sound) listening position measurements for room correction.

 

[dasdoing]

I imported the minimum phase corrected one (in 5 cycles FDW) to Rephase and tried to make the phase flat. I din't complete it because this is so hard to do in Rephase. but I did enough to show that only by flatting the phase you fill those FR gaps in the direct sound; just ingnore below 30Hz:

now the problem is that this creates pre-ringing, so Rephase is not realy a solution

and here again the MP corrected one with the phase, it is visable those 2 dips are caused by phase shift

 

[dasdoing]

In my past experience FDW used is not enough for flat direct sound in usual room sizes, thats why I personally rather use anechoic measurements above transition frequency and only below (where we don't just want flat direct sound) listening position measurements for room correction.

but how translatable is the anoechoic meassurement to the direct sound at you LP? you don't have the exact same speaker, and you are not at the same distance

 

[BenB]

now look what happens if we put those two into a 5 cycle freq. dependent window

It's not at all clear to me what processing you're performing when you run, nor when you analyze your results. Can you describe the process more comprehensively?

 

[dasdoing]

when you run

I don't understand

nor when you analyze your results

you get this windowing by clicking "IR windows" in REW and "add frequency dependent window" with "widht in cycles" of 5

 

[thewas]

but how translatable is the anoechoic meassurement to the direct sound at you LP? you don't have the exact same speaker, and you are not at the same distance

I compare to windowed measurements of my exact same loudspeakers (which are identical to anechoic above 500 Hz) and see that FDW doesn't give the same results which is to be expected as a FDW cannot fully remove early reflections which I remove by placing a loudspeaker in the middle of a room for measurements.

 

[dasdoing]

I compare to windowed measurements of my exact same loudspeakers (which are identical to anechoic above 500 Hz) and see that FDW doesn't give the same results which is to be expected as a FDW cannot fully remove early reflections which I remove by placing a loudspeaker in the middle of a room for measurements.

the question is what is: which on are we hearing at LP?
if the 5 cycle FDW includes some reflection, doesn't it mean that we percieve those as part of the direct sound?

 

[andreasmaaan]

the question is what is: which on are we hearing at LP?
if the 5 cycle FDW includes some reflection, doesn't it mean that we percieve those as part of the direct sound?

What leads you to believe that our ability to discern direct sound from reflected sound tracks a 5-cycle FDW?

 

[dasdoing]

What leads you to believe that our ability to discern direct sound from reflected sound tracks a 5-cycle FDW?

nothing,
I am actualy tinking the other way around here. if it is in this window, can we seperate it?

 

[dasdoing]

which leads to the next question: how big of a window would be the threshold?

 

[dasdoing]

If the width is in cycles a 15 cycle window (for example) would have a width of 150 ms at 100 Hz (15 times 10 ms), 15 ms at 1 kHz (15 times 1 ms) and 1.5 ms at 10 kHz (15 times 0.1 ms)

So a 5 cycle window covers 5ms at 1000Hz. if there is, say, a 1000Hz reflection at 4ms will you hear it as direct sound, or reflected sound? at 10000Hz the window is 0,5ms.

 

[thewas]

the question is what is: which on are we hearing at LP?
if the 5 cycle FDW includes some reflection, doesn't it mean that we percieve those as part of the direct sound?

That's exactly the question of the questions and my personal experience is that correcting to direct sound sounds more neutral to me than correcting to 5-6 cycles FDW above transition frequency.
I can only recommend everyone to do his own experiments and find out what works best for him with his setup, as it possibly depends also on the loudspeakers, room acoustics and listening distance.

 

[ernestcarl]

even those who claim not to hear the delays will find this intresting:

two meassurements, one minimum phase corrected, the other linear phase corrected. they are basicly similar in the 500ms window:

View attachment 109862

now look what happens if we put those two into a 5 cycle freq. dependent window

View attachment 109863

the green is the phase correted. it is not perfect because else pre-delay occures

if anybody thinks this is faked, I post the mdat

I still don't find the frequency dependent windowed SPL magnitude graphs all that informative. It's far better to look at the wavelet transform if you want to see the obvious improvement in both the time and frequency domains at the same time.

There's no direct ability to apply windowing in the wavelet view of REW (wish it had more filtering options), but you can export the 'windowed response' and reload it later if so desired:

KH120 + F12 sub

This was from a previous simple listening experiment that I did.

Didn't use a single linear phase xo filter, but several banks from rePhase's paragraphic phase EQ.

BTW, I found the delay in this case with streamed videos marginally acceptable.

Enough correction was applied to significantly reduce time spread/lag seen in the wavelet. I found the correction audibly better (less dull, a bit "brighter") than what's default. Dunno how this holds with longer listening distances and with lots of room reflections which is the reality in many home living room situations.

Further reading on the advantages of wavelet analysis in loudspeaker testing:
Time-Frequency Characterization of Loudspeaker Responses Using Wavelet Analysis
A Spectrogram Display for Loudspeaker Transient Response


*Okay, so I used more banks in rePhase a bit further and added one additional PEQ for the sub bass and this looks a little better:

**Not quite sure how visible "pre-ringing distortion" should be expected in the following views:

Some added pre-ringing in the bass I suppose...

To anyone who may have any idea, please do tell if there's anything that may indicate "bad pre-ringing" in the measurements. I've only seen/heard one or two extreme case examples. Haven't myself been able to notice after over an hour of listening to regular bass heavy music -- thus far anyway.

 

[bennybbbx]

the left measure (54) have good step response to reach 0. Is this with phase correct ?. what do you think ?. are the transients and stereo width get better ?.
maybe you can do audio records from the example guitar test i use ?. in other thread from you (with sceptre s8 kh 120 cmpare) i have explain how this can record with 1 microphone best.

I read that ears can recognize 10 microsecond delay between left and right ears.. this are 100 khz. I hear only 22 microseconds delay(less is not possible with 44.1 khz). I need test more with 96khz khz sample rate

I also notice when i hear loud longer time that the stereo width feeling get reduce suddenly. so maybe people that hear often loud damage ears. first maybe come the stereo hear loss and then come frequency loss of hear. I also notice that with the small 3.5 inch speakers the MTM and eris 3.5 have, i think the sound with frequency correct great so i need not hear loud. i hear not loud, i look on mixer levelmeter that i reach not more as 75 db now. because this is enough for mtm

the small speakers i have are better in falling slope and reach not near -100. maybe this is important for precise speaker

 

[ernestcarl]

the left measure (54) have good step response to reach 0. Is this with phase correct ?. what do you think ?. are the transients and stereo width get better ?.

My answer is pretty much the same to the concluding paragraph in the same post I linked to: simple listening experiment.

Improving the transient response via phase linearization improves 'clarity' and time-frequency coherence. Other properties of the speakers stay more or less the same e.g. directivity characteristics.

maybe you can do audio records from the example guitar test i use ?. in other thread from you (with sceptre s8 kh 120 cmpare) i have explain how this can record with 1 microphone best.

Sure, it's possible that you might be able to hear a tiny improvement with this technique. However, IMO, you're much better off trying this out with the monitors you already have (use a convolution engine in Foobar/JRiver etc.) Depending in your speakers and room acoustics (and your own hearing), you may or may not hear any significant change worth noting. Also, sorry, but I'm finding myself rather incorrigibly lazy for this task...

 

[bennybbbx]

maybe you can explain when you play the guitar test original on your speakers. so you need not recrord it. which sound in stereo width between the KH120 and sceptre s8 is larger or are they same ?. or which speaker sound more simular crisp to headphone. https://www.audiosciencereview.com/...tep-response-and-audio-record-examples.19812/ here you can download original. Recording of a speaker is usefull because you can then do blind tests with the free version
https://hofa-plugins.de/en/plugins/4u-blindtest/non headphones. hear the audio examples of kali and small woofer speakers. the diffrence is very large.better phase i hear give a little better room feeling on kali but it is not so much as need. test with distort guitar shows best slow speakers because when see wave of recordet guitar the levels change much (but not on high frequency). so can hear good when a speaker is too slow. maybe you hear this not as reduce stereo width but as muddy sound(when both speakers are frequency correct)