Getting the acoustic phase

Lex_audio

New Member
Thread Starter
Joined
Jan 17, 2019
Messages
10
Hello!
Please, say Me.
I need to isolate the acoustic phase to calculate the intersection of the loudspeakers. How can I do it?
1. I assembled the box and inserted the drivers into it.
2. I installed the microphone in front of the tweeter and took measurements from one microphone position. First tweeter, then midbass.
3. I see frequency response, which is good.
4. Analyze the impulse response and set the same temporary gate for both drivers.
5. I see how the frequency response changes and this is understandable.
But I don’t understand how to get the acoustic phase, taking into account that the distance from the microphone to the drivers is different? This is important for the design of crossovers.
Thank!
Oleg
 

EarlK

Member
Joined
May 22, 2017
Messages
221
You can obtain the "time-of-flight" phase info needed for crossover design ( easiest ) by using a 2-channel sound-card with one of the channels acting as the "0" time reference.

REW will display the time differential of the captured impulse ( when compared to the "0" reference ).

:)
 

Lex_audio

New Member
Thread Starter
Joined
Jan 17, 2019
Messages
10
Thanks for the answer. All the necessary equipment I have. And there is a bit of experience. I do not understand what actions in the REV need to take to get the phase of the flight time. Now I get this kind of phase schedule. This is unacceptable.
 

Attachments

  • 111.jpg
    111.jpg
    58.1 KB · Views: 25

EarlK

Member
Joined
May 22, 2017
Messages
221
Here's a nice pdf put together by the author of VituixCAD software ( for crossover design ).

Making Measurements for use in VituixCad pdf

Simply substitute REW processes for those of ARTA ( when mentioned within the pdf notes ).

If you don't fully know your way around REW then you have more homework to accomplish ( ie; you need to walk before running, etc. since crossover design = "running" ).

BTW, the measurement of that 10" woofer in my pic had the microphone capsule @ 1/2" from the paper cone ( iow; the measurement was made nearfield ).
I was looking to see if I could establish a usable near-field Acoustic Center ( of the JBL le10H when measured from the front of the basket rim ).
- I still don't know if that measurement is useful.

:)

2-Chnl_SoundCard_info.PNG

2-Chnl_SoundCard_LoopBack_info.PNG

MeasurementPanel_.PNG

Time-of-Flight_info.PNG
 

Attachments

  • VituixCAD Measurement Preparations.pdf
    1.2 MB · Views: 49
Last edited:

jtalden

Senior Member
Joined
May 22, 2017
Messages
887
Location
Arizona, USA
More  
Preamp, Processor or Receiver
Marantz AV7705 Pre/Pro
Main Amp
VTV 6 chnl NC252MP P-amp x 2
Additional Amp
Behringer DCX2496 x 2
Universal / Blu-ray / CD Player
OPPO BDP-103 Universal Player
Front Speakers
DIY SEAS H1456/H1212 Spkr x 5
Subwoofers
DIY JBL 2235H 15" SW x 2
Video Display Device
JVC DLA-X790R
Screen
Da-Lite Da-Snap 39105V - 92"
Oleg,
It sounds like you have the correct measurements; loopback timing activated, mic at 0.5-1 m and on the line of sight of the listening position (the TW level is a good generic position), full range sweeps.
The next steps to see phase correctly:
  • View the impulse of the TW and use the cursor to help identify the timing location of the peak (xx.xxx ms).
  • In the 'controls' panel enter that value into the 'T=0 offset (ms)' and then 'apply'
  • Offset the MW the same XX.XXX value used for the TW. [Both impulses are now near 0 ms and have the same relative timing relationship to each other.]
  • Make a 'Response Copy' of the MW using the 'All SPL > Controls' panel.
  • Open 'IR Windows' and set the window 'window Ref Time' to 0 ms and select 'Apply Windows to all'. [Left and right windows are noncritical for relatively close measurements, but can be very helpful in some cases.]
  • In overlay/phase window deselect the original phase measurement so only the MW-copy is visible. [we can now see the relative phase of the 2 drivers.]
  • Adjust the MW-copy impulse position as needed until the phase through the XO range closely tracks each other. [We have now identified the best impulse delay relationship for a stable frontal lobe of the direct sound arriving at the LP.]
  • Now measure the time offset between the original MW impulse and the MW-copy Impulse that was shifted. That is the time adjustment needed to achieve the phase alignment.
  • If you have a DSP you can make the timing adjustment and remeasure to confirm the result.
This should be helpful/enough for your situation.
There may be elements missed here and other specific cases may require additional steps/information. If you want to provide the an example mdat file containing you original TW and MW measurements I can provide my analysis result so you can confirm you have understood the process correctly. It also will help me see if I missed any steps critical to the process.

Below is an example result for a 2k XO.
0-Impulse.jpg
1-Phase.jpg
2-SPL.jpg
 

EarlK

Member
Joined
May 22, 2017
Messages
221
Hi John,

Nice of you to weigh in with your wisdom on matching phase traces.

Your method is a form of "Curve-Fitting" practiced in a couple of other ways ( by those seeking to make passive crossovers ).

At some point I'll need to check out your method to see how it compares to the others ( one of which is mentioned in my Acoustic Centers link ).

Cheers!
 
Last edited:

jtalden

Senior Member
Joined
May 22, 2017
Messages
887
Location
Arizona, USA
More  
Preamp, Processor or Receiver
Marantz AV7705 Pre/Pro
Main Amp
VTV 6 chnl NC252MP P-amp x 2
Additional Amp
Behringer DCX2496 x 2
Universal / Blu-ray / CD Player
OPPO BDP-103 Universal Player
Front Speakers
DIY SEAS H1456/H1212 Spkr x 5
Subwoofers
DIY JBL 2235H 15" SW x 2
Video Display Device
JVC DLA-X790R
Screen
Da-Lite Da-Snap 39105V - 92"
Hi Earl,
Thanks.

There are lots of methods that lead to good results. The curve-fitting target is indeed practiced in all 3-4 of the passive XO modeling processes I have seen. This is just one process to get there based on actual measurements in place of the modeling. It's helpful for those of us that put a couple of drivers in a box and then want to determine the ideal
DSP
delay to apply for our setup.

This overall process method is just one that I developed years ago when I started using REW. There may well be easier ways to get to the same answer. I have never even tried most of the alternative processes that I have seen referenced including the process you pointed at. I have started to read a couple of them, but usually get lost along the way. I would probably need to see and understand some specific step by step examples.

My process has a difficult learning curve, but works reliability for me. I have learned to recognize and deal with most all of the potential issues so I just continue using it.

Any process that leads to close phase tracking or provides a crossing point at the XO and does not diverge too quickly through the XO range is a good one. The SPL will be good and we are not very sensitive to anything else.
 

EarlK

Member
Joined
May 22, 2017
Messages
221
Hi Earl,
Thanks.

There are lots of methods that lead to good results. The curve-fitting target is indeed practiced in all 3-4 of the passive XO modeling processes I have seen. This is just one process to get there based on actual measurements in place of the modeling. It's helpful for those of us that put a couple of drivers in a box and then want to determine the ideal
DSP
delay to apply for our setup.

This overall process method is just one that I developed years ago when I started using REW. There may well be easier ways to get to the same answer. I have never even tried most of the alternative processes that I have seen referenced including the process you pointed at. I have started to read a couple of them, but usually get lost along the way. I would probably need to see and understand some specific step by step examples.

My process has a difficult learning curve, but works reliability for me. I have learned to recognize and deal with most all of the potential issues so I just continue using it.

Any process that leads to close phase tracking or provides a crossing point at the XO and does not diverge too quickly through the XO range is a good one. The SPL will be good and we are not very sensitive to anything else.


Thanks for your thoughts John.

Yes I have watched you fine tune this process over the last couple of years ( starting with syncing subs to the Main LR's etc. ).

Anyways, I see your process as returning an Offset ( Z ) number that is as accurate as is possible ( when using Minimum Phase info ).

I should be stated ( for the OPs sake ) that one can get a similar offset ( Z ) value simply by solving the differential existing between the two different time-of-flights ( for two drivers ).
- Your number should be slightly different due to the low-pass applied within the Hilbert Transform ( used to extract minimum phase ).

Hopefully this isn't confusing Oleg.

:)
 

jtalden

Senior Member
Joined
May 22, 2017
Messages
887
Location
Arizona, USA
More  
Preamp, Processor or Receiver
Marantz AV7705 Pre/Pro
Main Amp
VTV 6 chnl NC252MP P-amp x 2
Additional Amp
Behringer DCX2496 x 2
Universal / Blu-ray / CD Player
OPPO BDP-103 Universal Player
Front Speakers
DIY SEAS H1456/H1212 Spkr x 5
Subwoofers
DIY JBL 2235H 15" SW x 2
Video Display Device
JVC DLA-X790R
Screen
Da-Lite Da-Snap 39105V - 92"
Oh, I finally understand the OP was looking for info for input into a model for a passive XO. My mistake. My method is indeed only for those who have a DSP XO in place and active and just need to determine the required delay adjustment.
Sorry for the confusion.
 

Lex_audio

New Member
Thread Starter
Joined
Jan 17, 2019
Messages
10
Hello colleagues. :) I received a lot of information from you and now I need a little time to think about all the options and try. My friend has SoundEasy, and I made several measurements with him. Everything is simple there. We set up the microphone and measure tweeter and modbas from a common position. Immediately get the phase in which the excess phase and the minimum phase are included. It is convenient and can be immediately used in the calculation of the crossover. It is even possible to carry out the Gilbert-Bode transform to determine displacements of drivers relative to each other. This is easy and understandable. However, a friend lives in another city and it is not convenient for me.
Thanks for the advice. I will definitely try and if I have any questions, I will ask.
Thank! :)
 
Top Bottom