Phase unwrap questions

[user44455555]

please see video.it is avi extension so i need zip it because forum do not accept it. the phase jumps depend on cursor area when i call unwrap phase but it did not put the curve at 0 at cursor position. I look for a way to set 40 hz as phase reference 0 so i can compare which speaker is the best. same microphone position same speaker position, same buffer size, but phases are very diffrent shift. there is tell that phase did not show correct because of room. but look at the JBL 104 which have the best bass clarity. the phase look also at 40 hz also best in compare to other. so maybe most speakers are worse, because in speaker tests, nobody show the phase in unwraped. with wraped can see nothing. or wy can the phase of the JBL 104 in bass so much better in compare to other speakers when room is most problem ? https://www.avnirvana.com/threads/i...her-resolution-as-the-t60m.16397/#post-122162

teh mdat is here

 

[John Mulcahy]

Unwrapping phase does exactly what it says. The region where the cursor is stays within -180..180 and phase wraps before and after the cursor position are unwrapped.

The phase trace depends on the position defined as t=0, if that varies so will the phase trace.

 

[user44455555]

Unwrapping phase does exactly what it says. The region where the cursor is stays within -180..180 and phase wraps before and after the cursor position are unwrapped.

The phase trace depends on the position defined as t=0, if that varies so will the phase trace.

I have add mdat with 2 measure same place of microphone at 1,5 cm distance so no room influence possible. I do estimate IR delay and set t0 to IR peak and set cursor to 20 hz(begin of measure) and call on overlays unwrap phase(this i notice unwrap alle phases of all speakers i enable to show) . I get between the speakers when add the diffrence 170 degree at 20 hz ( . this is huge diffrence. why does it at 20 hz not show near 0 ?. 20 hz have wavelength 17,5 meter. so 170 degree is then ~9 meter diffrence between speaker. when understand how phase display work correct

 

[John Mulcahy]

no room influence possible

Both measurements have a reflection at around 13 ms, for the Kali it is only 9.5 dB below the peak.

I do estimate IR delay

Changes t = 0

and set t0 to IR peak

Changes t = 0

I get between the speakers when add the diffrence 170 degree at 20 hz

Why should different speakers, with different low frequency responses, have the same phase? That would not make any sense at all. Where responses are minimum phase the phase response is directly related to the magnitude response. Try setting up some simulated speakers consisting of just a high pass filter for the low frequency roll-off and a low pass filter for the high frequency roll-off (use Generate measurement from filters). Do not change t=0. Examine what happens to the phase responses of those filters as you change the filter frequencies and roll-off slope, that may help you understand the factors affecting the phase response.

 

[user44455555]

Why should different speakers, with different low frequency responses, have the same phase? That would not make any sense at all. Where responses are minimum phase the phase response is directly related to the magnitude response. Try setting up some simulated speakers consisting of just a high pass filter for the low frequency roll-off and a low pass filter for the high frequency roll-off (use Generate measurement from filters). Do not change t=0. Examine what happens to the phase responses of those filters as you change the filter frequencies and roll-off slope, that may help you understand the factors affecting the phase response.

I mean because the measure begin with frequency 20 hz and i choose 20 hz for unwrap wy did not both speakers at 20 hz begin with 0 degree in phase because thats start point of measure ?. if REW choose a frequency maybe 1 khz to set phase reference to 0, then the phase of both speakers at 1 khz should same ?. but the phase of both speakers is ~170 degree different from 20 hz to 300 hz offset of phase shift stay near same. so my question is, how this constant offset of 170 degrees happens ?. microphone position and speaker position was same. the measures can see are only 7 min later done. The kali have 6.5 inch bass the jbl 104 4.5 bass and much less low bass. but the phaseoffset between speakers have near same constant offset from 20 to 300. and 300 hz

 

[John Mulcahy]

That is not how phase works.

 

[user44455555]

That is not how phase works.

also strange in the measures, the kali have at 20 hz a phase of 131 degree. this mean the recordet wave come earlier than the measure tone. at 48 hz it cross zero. the jbl have at 20 hz -46 degree. the JBL result look realistic

I think this is correct explain how phase offsets work ? https://www.electronics-tutorials.ws/accircuits/phase-difference.html is this correct ?
with T0 they mean zero crossing, but i try values when REW use peak of sine .

So i try change T0 in REW to let start at 20 hz with 0. the start point shift, but then it give very unrealistic strange curves. this happen always when try shift t0 in REW

 

[John Mulcahy]

this mean the recordet wave come earlier than the measure tone

No, it doesn't.

I think this is correct explain how phase offsets work ?

No. Shifting t=0 applies a phase shift that varies linearly with frequency.

Please consider my suggestion about examining the behaviour of a simple high pass, low pass filter arrangement, or find some other resource to explain the phase of a transfer function. Your questions are nonsensical and will continue to be that way until you start to understand the phase component of a transfer function, I will not respond further.

 

[user44455555]

No, it doesn't.



No. Shifting t=0 applies a phase shift that varies linearly with frequency.

Please consider my suggestion about examining the behaviour of a simple high pass, low pass filter arrangement, or find some other resource to explain the phase of a transfer function. Your questions are nonsensical and will continue to be that way until you start to understand the phase component of a transfer function, I will not respond further.

you mean i should simulate with EQ the bass range of kali and compare phase then ?. this i have done with REW EQ. The Jbl 104 get FR in the Bass part as the Kali. you can see the jbl SPL change alot but the phase only slightly to the original JBL 104. the huge diffrence in phase to the kali stay, it is reduce now but still over 130 degree. phase of the EQ version begin near 0 at 20 hz I use for left right filter in prefs the default tucey 0.25

I understand a high pass min phase filter do phase shift, but a speaker do more phase shift as a min phase filter. maybe call it as worse factor or so.

 

[user44455555]

thanks to your help and hint with try out filters now i understand what excess phase and predicted phase is, because when i EQ the JBL it have same excess phase as the non EQ version . only kali have very diffrent excess phase
So i assume min phase is the result the frequency response of the speaker have if it is min phase. and when show excess phase REW subtract the min phase from the measures phase, and excess phase is the additional phase of the worseness of speakers ?

REW have no overlay for excess phase button, but predictated phase seem do the same. The version i EQ have the same predicted phase result as the non EQ version. so to compare speakers in overlay, the predictated phase is the important thing and the phase only should ignore.

both the JBL EQ version and JBL non EQ version have at 20 hz 10 degree phase. the kali have 131 degree and this mean the kali is worse and this also explain wy the clarity at 50 hz is better for the JBL 104

sense of all this is have tests that show how good a speaker is so can choose without pay snakeoil addition price. because i notice by hearing that big expensive speakers sound much worser in bass clarity also when eq them same and compare

now the question is only how the kali come with excess phase to 131 degree which mean it is faster than the measure tone ?

 

[user44455555]

the group delay is also frequency dependent and there is no excess group delay or predicted group delay. you can see the groupdelay slightly change when i FR the JBL to the kali, but still much better group delay on JBL. the mdat i have upload too