ProSoundWeb Community

Sound Reinforcement - Forums for Live Sound Professionals - Your Displayed Name Must Be Your Real Full Name To Post In The Live Sound Forums => LAB Lounge => Topic started by: JoeyCurran on February 28, 2014, 02:00:03 PM

Title: Phase align subs to mains
Post by: JoeyCurran on February 28, 2014, 02:00:03 PM
Hi Folks,

I'm hoping someone may help me learn to use Smaart to align  subs to mains systems.

As I understand it, we can’t time align a subwoofer to the mains, as subwoofers are stretched over time per frequency. So we cannot use the Impulse Response function in Smaart 6 to accomplish this. However, when attempting to measure with Impulse Response, the arrival times of the mains and subs at the mix position, I ran into some issues I hope someone can help me understand.

Measuring the arrival time of the mains at the mix position seemed to be very easy. I simply set a level, chose the IR Analysis page, started the Signal Generator using pink noise, and started the IR Analysis. The resulting trace was an easily discernible trace showing arrival time and positive polarity. I utilized the following settings in IR analysis: Type – Linear IR, FFT – 64k, TC (ms) – 1365, Avg – 4.

I then attempted to measure the subs arrival at the mix, with the same settings and method, but failed miserably. I tried increasing the FFT size and averages with no success. The trace that resulted was a nearly flat line, with no way to determine polarity or arrival time.

I hope someone who sees this post will know why I was unsuccessful in my first attempt. I must say I am not completely surprised at the result, as the more I attempt to come to terms with this very powerful tool the more I realize how little I know regarding its use.

Best Regards,
Joey

edited for spelling
Title: Phase align subs to mains post
Post by: JoeyCurran on February 28, 2014, 02:25:26 PM
Hi Folks,

I'm hoping someone may help me learn to use Smaart to align  subs to mains systems.

As I understand it, we can’t time align a subwoofer to the mains, as subwoofers are stretched over time per frequency. So we cannot use the Impulse Response function in Smaart 6 to accomplish this. However, when attempting to measure with Impulse Response, the arrival times of the mains and subs at the mix position, I ran into some issues I hope someone can help me understand.

Measuring the arrival time of the mains at the mix position seemed to be very easy. I simply set a level, chose the IR Analysis page, started the Signal Generator using pink noise, and started the IR Analysis. The resulting trace was an easily discernible trace showing arrival time and positive polarity. I utilized the following settings in IR analysis: Type – Linear IR, FFT – 64k, TC (ms) – 1365, Avg – 4.

I then attempted to measure the subs arrival at the mix, with the same settings and method, but failed miserably. I tried increasing the FFT size and averages with no success. The trace that resulted was a nearly flat line, with no way to determine polarity or arrival time.

I hope someone who sees this post will know why I was unsuccessful in my first attempt. I must say I am not completely surprised at the result, as the more I attempt to come to terms with this very powerful tool the more I realize how little I know regarding its use.

Best Regards,
Joey

edited for spelling
Title: Re: Phase align subs to mains post
Post by: Art Welter on February 28, 2014, 03:41:16 PM
Hi Folks,

I'm hoping someone may help me learn to use Smaart to align  subs to mains systems.

As I understand it, we can’t time align a subwoofer to the mains, as subwoofers are stretched over time per frequency.
Joey,

This thread explains the procedures:

http://www.rationalacoustics.com/forums/showthread.php?5137-How-to-measure-Sub-with-Fullrange

It also helps to know where you are "starting from", testing your system outside with the subs next to the mains with the mic on the ground equidistant between the two at about 2 meters (and away from big objects by 10 meters or so) will make it much more easy to see when the phase traces line up.
Then when you move indoors you can have a more educated guess as where to start with delay times- roughly a millisecond per foot of physical offset between subs and tops in relation to the listening/measurement position from your outdoor figures. If your mains are always stacked on your subs, the delay will not need to be changed regardless of the venue.

One "gotcha" that messes up alignment is you can over delay the tops by one wavelength and  still phase align the subs, but if your crossover is around 100 Hz the subs would then end up lagging by around 10 ms, they can measure flat, but sound "slow". That said, the output of a ported top cabinet lags the main output by 180 degrees at Fb (box tuning) so depending on the acoustical crossover point in relation to Fb, things can be interesting to figure out...

Title: Re: Phase align subs to mains post
Post by: JoeyCurran on February 28, 2014, 04:07:09 PM
Hi Art,

Thanks for your reply. Thanks for the link, I'll check that out straight away. The system is installed and is not available to take outdoors.



One "gotcha" that messes up alignment is you can over delay the tops by one wavelength and  still phase align the subs, but if your crossover is around 100 Hz the subs would then end up lagging by around 10 ms, they can measure flat, but sound "slow". That said, the output of a ported top cabinet lags the main output by 180 degrees at Fb (box tuning) so depending on the acoustical crossover point in relation to Fb, things can be interesting to figure out...

I am fairly certain I have heard this effect!!!
Title: Re: Phase align subs to mains
Post by: Martyn ferrit Rowe on February 28, 2014, 05:11:15 PM
Hey Joey,
Try this:   http://bobmccarthy.wordpress.com/2010/02/08/phase-alignment-of-subs-why-i-dont-use-the-impulse-response/

Cheers
ferrit
Title: Re: Phase align subs to mains
Post by: Doug Fowler on February 28, 2014, 06:47:13 PM
Hi Folks,

I'm hoping someone may help me learn to use Smaart to align  subs to mains systems.

As I understand it, we can’t time align a subwoofer to the mains, as subwoofers are stretched over time per frequency. So we cannot use the Impulse Response function in Smaart 6 to accomplish this. However, when attempting to measure with Impulse Response, the arrival times of the mains and subs at the mix position, I ran into some issues I hope someone can help me understand.

Measuring the arrival time of the mains at the mix position seemed to be very easy. I simply set a level, chose the IR Analysis page, started the Signal Generator using pink noise, and started the IR Analysis. The resulting trace was an easily discernible trace showing arrival time and positive polarity. I utilized the following settings in IR analysis: Type – Linear IR, FFT – 64k, TC (ms) – 1365, Avg – 4.

I then attempted to measure the subs arrival at the mix, with the same settings and method, but failed miserably. I tried increasing the FFT size and averages with no success. The trace that resulted was a nearly flat line, with no way to determine polarity or arrival time.

I hope someone who sees this post will know why I was unsuccessful in my first attempt. I must say I am not completely surprised at the result, as the more I attempt to come to terms with this very powerful tool the more I realize how little I know regarding its use.

Best Regards,
Joey

edited for spelling

The energy from the sub is so spread out in time it's very difficult to see what's actually going on in IR.   Because it's so spread out, there is much "less amplitude" than in the HF, where everything is pretty much (but not exactly) arriving at the same time.

Smaart 6 was not very good at locking onto LF arrival times IIRC.  It seems to be quite a bit better now.

One of the other things you're dealing with here is the LPF on the subs.  The crossover will cause the peak energy of the sub to be later than if it were a full range signal.  You can verify this easily with electronic measurements.  Measure through a crossover, put the delay finder on "auto", and start to lower the frequency on a LPF.  You'll see the peak really start to move as you get near the sub region.

This is precisely where many have (wrongly) rolled up the LPF over 1 KHz, gotten their "delay time", and called the job done.  Clever, but totally wrong.
Title: Re: Phase align subs to mains
Post by: JoeyCurran on March 01, 2014, 04:57:22 AM
Hey Joey,
Try this:   http://bobmccarthy.wordpress.com/2010/02/08/phase-alignment-of-subs-why-i-dont-use-the-impulse-response/

Cheers
ferrit

Hi Martyn,

Thanks for the link. Exactly what I needed!

Cheers,
Joey
Title: Re: Phase align subs to mains
Post by: JoeyCurran on March 01, 2014, 05:02:53 AM
The energy from the sub is so spread out in time it's very difficult to see what's actually going on in IR.   Because it's so spread out, there is much "less amplitude" than in the HF, where everything is pretty much (but not exactly) arriving at the same time.

Smaart 6 was not very good at locking onto LF arrival times IIRC.  It seems to be quite a bit better now.

One of the other things you're dealing with here is the LPF on the subs.  The crossover will cause the peak energy of the sub to be later than if it were a full range signal.  You can verify this easily with electronic measurements.  Measure through a crossover, put the delay finder on "auto", and start to lower the frequency on a LPF.  You'll see the peak really start to move as you get near the sub region.

This is precisely where many have (wrongly) rolled up the LPF over 1 KHz, gotten their "delay time", and called the job done.  Clever, but totally wrong.

Hi Doug,

Thanks for explaining that. So in order to get in the ballpark with setting a delay for the subs to mains, would I be better off measuring propagation to the listening area of each system separately, then adjusting the time from one or the other through electronic delay? After I'm in the ballpark, then I can fine tune Phase and slope?

Cheers,
Joey
Title: Re: Phase align subs to mains
Post by: JoeyCurran on March 01, 2014, 05:39:44 AM
Hi Folks,

Which would be more appropriate for measuring phase and slope between these pass bands? Noise/sine/program? Sine wave at the crossover freq in the processor? Or pink noise in the octave 63-125hz, covering the crossover area? Sub heavy music?

Cheers,
Joey
Title: Re: Phase align subs to mains post
Post by: JoeyCurran on March 01, 2014, 06:07:01 AM
Hi Folks,

Thanks for the great info. I was tying to find delay/propagation times for the different sub-systems I was measuring, hence the use of Impulse Response. I understand now why this is an inappropriate approach, as the energy from the sub is so spread out in time it's very difficult to see what's actually going on in IR. Thanks again for your help.

Which would be more appropriate for measuring phase and slope between these pass bands? Noise/sine/program? Sine wave at the crossover freq in the processor? Or pink noise in the octave 63-125hz, covering the crossover area? Sub heavy music?

Cheers,
Joey
Title: Re: Phase align subs to mains
Post by: Raimonds.Skuruls on March 01, 2014, 06:54:13 AM
Dear Colleagues,

May I try to help to find solution of this topic problem?
First, I would like to introduce you to solution to visualize delays in our loudspeaker systems.
Second, let’s see on hardest case – heavy filtered subwoofer.

Let’s start with 3 way, 2nd order crossover.  (picture 01_CR3w_12.png)
It is creating just 0.9 ms delay on 100 Hz. (picture 02_CR3w_12okTDApl.png)
Let’s tune out delays of crossover bands. The relative delay of each band is easy readable from the graph.
(picture 03_CR3w_12_noTDApl.png)

What about 3 way, 8th order (48dB/oct) crossover. (picture 04_CR3w_48.png)
As you see, it is introducing serious amount of delay. (picture 05_CR3w_48okTDApl.png)
Let’s tune out delays of crossover bands. (picture 06_CR3w_48noTDApl.png)

And now, 2 way system with a heavy filtered subwoofer. (picture 07_CR2wSS_48.png)
Its delay picture  08_CR2wSS_48okTDApl.png
And impulse response of the crossover subwoofer band - 09_IR_LF40_100.png
Let’s introduce some additional delay   -  10_CR2wSS_48noTDApl_1_2.png

And solution.
The best delay tuning result, you can get, will be just the crossover delay response.
For such subwoofer case, as described, it is impossible to read small delay step on crossover frequency point. Therefore, you must introduce significant (30 ... 40...50 ms) additional delay to separate bands in graph. Than measure real delay step by use of data cursors (picture 10_CR2wSS_48noTDApl_1_2.png) and subtract this result from your initially introduced additional delay. You may need final tweaking in range - 90 degrees (-2.5 ms for 100 Hz) from such result by observing AFR at the crossover frequency.

But, why do we need so heavy filtered subwoofer with such huge timing (delay) distortions.
This problem is more serious for band pass subs... You can see even 150 ms of delay.
Of course, it is ok for a loudspeaker that is intended to reproduce just low frequency sound effects.

BR,

Raimonds
Title: Re: Phase align subs to mains
Post by: Raimonds.Skuruls on March 01, 2014, 06:56:18 AM
next pictures
Title: Re: Phase align subs to mains
Post by: Timo Beckman on March 01, 2014, 07:38:52 AM
So i did a blog post a few days ago where i explained a little about how you might eq a system using smaart7.
Since i also measured the output of the processor you can see the effect eq filters have on the live ir.
The measurements on the low driver in that particular system and it's processor output show the ir being stretched more in time with every additional eq filter i put in (as expected) and at the last screen shot on the low driver it moves because i implemented a low pass filter as was also expected.

http://timobeckmangeluid.wordpress.com/2014/02/12/eqin-a-system/

Now would i use the ir for the final alignment : no i would have to phase align the whole thing in the end.
This was a low driver and as you can see the lp filter was set at 1200Hz. If you implement a low pass at say a 100Hz on a sub-woofer the same thing will happen (only the values will be a bit more then at 1200Hz:-).
Title: Re: Phase align subs to mains post
Post by: Ivan Beaver on March 01, 2014, 10:08:07 AM
Hi Folks,

Thanks for the great info. I was tying to find delay/propagation times for the different sub-systems I was measuring, hence the use of Impulse Response. I understand now why this is an inappropriate approach, as the energy from the sub is so spread out in time it's very difficult to see what's actually going on in IR. Thanks again for your help.

Which would be more appropriate for measuring phase and slope between these pass bands? Noise/sine/program? Sine wave at the crossover freq in the processor? Or pink noise in the octave 63-125hz, covering the crossover area? Sub heavy music?

Cheers,
Joey

Sine waves will only be one freq at a time-Very little to see there.

Swept sine takes awhile to see results of changes.

Music will not energize all the freq-and it doesn't matter if it is bass heavy or not.

Pink will energize all freq and is the best to use in my opinion.
Title: Re: Phase align subs to mains
Post by: Tim McCulloch on March 01, 2014, 11:07:04 AM
Hi Folks,

I'm hoping someone may help me learn to use Smaart to align  subs to mains systems.

As I understand it, we can’t time align a subwoofer to the mains, as subwoofers are stretched over time per frequency. So we cannot use the Impulse Response function in Smaart 6 to accomplish this. However, when attempting to measure with Impulse Response, the arrival times of the mains and subs at the mix position, I ran into some issues I hope someone can help me understand.

Measuring the arrival time of the mains at the mix position seemed to be very easy. I simply set a level, chose the IR Analysis page, started the Signal Generator using pink noise, and started the IR Analysis. The resulting trace was an easily discernible trace showing arrival time and positive polarity. I utilized the following settings in IR analysis: Type – Linear IR, FFT – 64k, TC (ms) – 1365, Avg – 4.

I then attempted to measure the subs arrival at the mix, with the same settings and method, but failed miserably. I tried increasing the FFT size and averages with no success. The trace that resulted was a nearly flat line, with no way to determine polarity or arrival time.

I hope someone who sees this post will know why I was unsuccessful in my first attempt. I must say I am not completely surprised at the result, as the more I attempt to come to terms with this very powerful tool the more I realize how little I know regarding its use.

Best Regards,
Joey

edited for spelling

Joey, how many threads on how many different sites?
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 01, 2014, 01:43:58 PM
Joey, how many threads on how many different sites?
You know that way you get all different kinds of answers-so it makes it confusing to "piece them al together".
Title: Re: Phase align subs to mains
Post by: Ryan Hammond on March 02, 2014, 08:41:35 AM
I use a polarity tester from JL Audio Tools.
Title: Re: Phase align subs to mains
Post by: Mac Kerr on March 02, 2014, 10:37:44 AM
I use a polarity tester from JL Audio Tools.

You might be able to match the polarity of the subs to the mains, but a polarity checker will not help at all in aligning the phase.

Mac
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 02, 2014, 10:55:32 AM
You might be able to match the polarity of the subs to the mains, but a polarity checker will not help at all in aligning the phase.

Mac
Phase-polarity-it is sadly the same to most people :((((

YES THERE IS A DIFFERENCE-AND it is IMPORTANT!

Think of it this way.  2 cars (lets use NASCAR for example) cross the finish line at the same time.  They are in "polarity".

Which one wins?  The one that is has the most "phase shift" (laps around the track).  The losing car may be laps behind-but they are in the same physical position (polarity).

I guess it "doesn't matter" to some people-but it SURE DOES to the guy who gets the check!!!!
Title: Re: Phase align subs to mains
Post by: Cailen Waddell on March 02, 2014, 11:19:23 AM

Phase-polarity-it is sadly the same to most people :((((

YES THERE IS A DIFFERENCE-AND it is IMPORTANT!

Think of it this way.  2 cars (lets use NASCAR for example) cross the finish line at the same time.  They are in "polarity".

Which one wins?  The one that is has the most "phase shift" (laps around the track).  The losing car may be laps behind-but they are in the same physical position (polarity).

I guess it "doesn't matter" to some people-but it SURE DOES to the guy who gets the check!!!!

So Ivan - a somewhat ignorant question.  If both cars cross the finish line at the same time (amplifying each other) but one is a lap behind, how do you see that on measurement software?  On an RtA it would follow that things 'look right'. Yes?  Or am I missing something critical (I probably am).

Title: Re: Phase align subs to mains
Post by: Mac Kerr on March 02, 2014, 11:26:33 AM
So Ivan - a somewhat ignorant question.  If both cars cross the finish line at the same time (amplifying each other) but one is a lap behind, how do you see that on measurement software?  On an RtA it would follow that things 'look right'. Yes?  Or am I missing something critical (I probably am).

On a dual FFT analyzer like Smaart, or SIM, or SysTune, you see a phase trace that is frequency vs phase. To align different subsystems you look at their phase trace and get them to match at the frequencies of interest.

In Ivan's race car analogy the cars represent a sine wave. In real audio it is more complex, but still understandable, measurable, and alignable (over a limited range of frequencies).

A simple RTA cannot show this information.

Mac
Title: Re: Phase align subs to mains
Post by: Jerome Malsack on March 02, 2014, 11:33:36 AM
My example to your question would be Latency.  Because my signal goes threw a system and becomes late but arrives in polarity with the signal that is on time does not present it self in photo.  Our ears will know that the sound is not on the correct time.   The Kick drum with a significant amount of smack from the beater in the 4 k area is heard and the sub 80 hz segment is not there.  The next beat you get beater and sub together.  Until the song ends and the beater is missing but you get the 80hz kick extra after??
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 02, 2014, 12:19:53 PM
On a dual FFT analyzer like Smaart, or SIM, or SysTune, you see a phase trace that is frequency vs phase. To align different subsystems you look at their phase trace and get them to match at the frequencies of interest.

In Ivan's race car analogy the cars represent a sine wave. In real audio it is more complex, but still understandable, measurable, and alignable (over a limited range of frequencies).

A simple RTA cannot show this information.

Mac
Exactly-you would see a 360° phase shift.

I would also add that it is kinda easy to get 2 traces to align at one freq.

But one should ALSO look at what is happening on a octave on either side of crossover.

This is something that the simple "RTA and sinewave-adjust till max cancellation-then flip polarity method) does not show.

You could end up with cancellations on either side of crossover if the time is not correct.

So you have to look at the phase-the amplitude (over a range of freq) to see if you have got it correct.

Save trace-adjust time-what happened? adjust again and so forth.
Title: Re: Phase align subs to mains
Post by: Jay Barracato on March 02, 2014, 12:23:17 PM
The critical point is that you are not looking at what the phase is, but how it is changing with frequency. The slope of the phase trace is the critical information.

Sent from my DROID RAZR HD using Tapatalk

Title: Re: Phase align subs to mains
Post by: Thad Kraus on March 02, 2014, 04:24:26 PM
So i did a blog post a few days ago where i explained a little about how you might eq a system using smaart7.
Since i also measured the output of the processor you can see the effect eq filters have on the live ir.
The measurements on the low driver in that particular system and it's processor output show the ir being stretched more in time with every additional eq filter i put in (as expected) and at the last screen shot on the low driver it moves because i implemented a low pass filter as was also expected.

http://timobeckmangeluid.wordpress.com/2014/02/12/eqin-a-system/

Now would i use the ir for the final alignment : no i would have to phase align the whole thing in the end.
This was a low driver and as you can see the lp filter was set at 1200Hz. If you implement a low pass at say a 100Hz on a sub-woofer the same thing will happen (only the values will be a bit more then at 1200Hz:-).

Just read through this article and it was quite helpful. How does one handle a reference signal when not measuring after the DSP? For example if using crowns internal DSP on their amps I am unable to take a measurement after processing.
Title: Re: Phase align subs to mains
Post by: Tim McCulloch on March 02, 2014, 04:55:16 PM
Just read through this article and it was quite helpful. How does one handle a reference signal when not measuring after the DSP? For example if using crowns internal DSP on their amps I am unable to take a measurement after processing.

Timo's article was for demonstration purposes, to show the absolute affect of filters on time by comparing the output of the DSP to the acoustic output of the loudspeaker.  For the purposes of aligning subs to tops, you don't need to measure the DSP as all we care about is how the acoustic crossover works out.

Jay B. says "The slope of the phase trace is the critical information."  I agree, and add to it - the other critical info is that the phase traces of the tops and subs are overlapping or at least parallel and not separated by more than 90° (less is better), and the overlap should extend at least 1/2 octave (a full octave is mucho better) on either side of the acoustic crossover.
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 02, 2014, 05:10:30 PM
Just read through this article and it was quite helpful. How does one handle a reference signal when not measuring after the DSP? For example if using crowns internal DSP on their amps I am unable to take a measurement after processing.
The reference signal is gotten from the output of the interface and looped back into one channel.

When you measure the loudspeaker, you are measuring everything inbetween the noise source and the loudspeaker-processing-amplifier-mixer-room-air etc.

So you adjust as needed/desired to get the response that you want.
Title: Re: Phase align subs to mains
Post by: Cailen Waddell on March 02, 2014, 05:26:47 PM

The reference signal is gotten from the output of the interface and looped back into one channel.

When you measure the loudspeaker, you are measuring everything inbetween the noise source and the loudspeaker-processing-amplifier-mixer-room-air etc.

So you adjust as needed/desired to get the response that you want.

So - I have a lot of questions now...

One - if you are comparing your reference signal to what is coming out of the speakers, there is a time (phase?) difference due to processing and the distance from the mic to the speaker(s).  Do you correct for this?  Does it matter?

Two - when looking in smart at the phase response, do you just slowly nudge your delay on your subs until it looks right?   Surely I am over simplifying...

Three - multiple microphone - because a singal mic represents a single seat, so how many places do you measure from?


Thanks all... This is still a bit of a dark art for me, though I understand the approach is actually quite scientific, logical, and the art perhaps comes in when knowing what compromises to make where....
Title: Re: Phase align subs to mains
Post by: Mac Kerr on March 02, 2014, 05:36:45 PM
So - I have a lot of questions now...

One - if you are comparing your reference signal to what is coming out of the speakers, there is a time (phase?) difference due to processing and the distance from the mic to the speaker(s).  Do you correct for this?  Does it matter?
Yes it matters, and the offset is done in the analyzing software so that it is comparing 2 signals in time with each other.

Quote
Two - when looking in smart at the phase response, do you just slowly nudge your delay on your subs until it looks right?   Surely I am over simplifying...
Yes you do, but it may be that you will have to delay your mains to the subs.

Quote
Three - multiple microphone - because a singal mic represents a single seat, so how many places do you measure from?
Unless your mains and subs are very close to each other you will only be able to get it right for a limited number of seats, pick the good ones. If your mains and subs are very close together, the area of good phase match will be much larger, and it won't matter much where you measure, pick good seats.

Mac
Title: Re: Phase align subs to mains
Post by: Tim McCulloch on March 02, 2014, 05:47:24 PM
So - I have a lot of questions now...

One - if you are comparing your reference signal to what is coming out of the speakers, there is a time (phase?) difference due to processing and the distance from the mic to the speaker(s).  Do you correct for this?  Does it matter?

Two - when looking in smart at the phase response, do you just slowly nudge your delay on your subs until it looks right?   Surely I am over simplifying...

Three - multiple microphone - because a singal mic represents a single seat, so how many places do you measure from?


Thanks all... This is still a bit of a dark art for me, though I understand the approach is actually quite scientific, logical, and the art perhaps comes in when knowing what compromises to make where....

ONE - Yes, there is a time difference.  You use the Delay Finder in Smaart to apply delay to the reference signal.

TWO - Uh... only if you know for sure that your subs are AHEAD of the mains, time-wise.  My experience is that it ain't necessarily so.  You may end up delaying the tops to the subs and then aligning the rest of the system to the tops.

THREE - It depends on what you're measuring and why.  If you're doing a sub/top alignment Mac's advice is spot on.  If you're doing more, you need to measure where various subsystems (mains v. ff, balcony/under-balcony fills etc) have overlapping coverage within 6dB of each other.

The art of measurement is in several areas:  interpretation of displayed results and developing an intuitive grasp of what you can align and what you can't.  Based on science but executed in a more artisan way.
Title: Re: Phase align subs to mains
Post by: Ryan Hammond on March 02, 2014, 05:49:14 PM
Wow... it all makes sense, but I never gave it a second thought... I just assumed that they were in phase... we've all heard the obvious phase issues with mains, but it never occurred to me that the subs and mains could be off. This is why I have only been concerned with polarity and not the phase relationship to mains. Our subs have always been very close to the mains. ... it all "sounds" in line throughout the venue.

Title: Re: Phase align subs to mains
Post by: Mac Kerr on March 02, 2014, 05:54:11 PM
Wow... it all makes sense, but I never gave it a second thought... I just assumed that they were in phase... we've all heard the obvious phase issues with mains, but it never occurred to me that the subs and mains could be off. This is why I have only been concerned with polarity and not the phase relationship to mains.

No, they will still need to be phase aligned, but the alignment will work over most of the coverage area. The phase of the signal coming out of a speaker is affected by crossover design, box design, driver selection, and electronic treatment. The idea is to use the electronic treatment to correct the errors induced by all the other influences.

Mac
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 02, 2014, 06:03:09 PM
If your mains and subs are very close together, the area of good phase match will be much larger, and it won't matter much where you measure, pick good seats.

Mac
Like where the guy that writes your check sits------------------------
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 02, 2014, 06:13:11 PM
. Our subs have always been very close to the mains. ... it all "sounds" in line throughout the venue.
But simply "being physically close" DOES NOT mean that they are in phase.

What Mac was saying was that if the subs and the mains are close together-then the relative distances between the two will be the same over a wide area.

If the subs are on the floor and the mains flown-almost everywhere you move you will end up with a different delay time-hence the need for the "good seats".

When the subs and tops are properly aligned-there will be more impact-clarity etc to the music.

AND REALLY IMPORTANT-that MAY MEAN that they are out of POLARITY!!!!!!!

This if often the case.  Other times it is not-it really depends on the particular cabinets used-the processing that is on them (crossover filter types/slopes/freq etc).  They all have an effect on the actual arrival time.  Steeper slopes cause more "delay".

Hence the reason audiophiles like first order filters-because of the gentle slopes and it is much easier to align different passbands with gentle slopes.  But the gentle slopes don't offer much protection to the drivers-so it is a trade off.

It really does not matter what the polarity is-it is the PHASE RELATIONSHIP that you are interested in.  And different cabinet combinations will have different relationships.

Just the simple action of putting a low pass filter on the subs will throw their "arrival time" off-even if they are sitting right next to the mains.

And if you are using horn loaded subs-there will be additional delay inside the cabinet.

To be honest I never worry about whether the subs are in polarity with the mains-it is simply the phase that is most important.

And with the phase trace you can VERY QUICKLY tell if they are out of polarity-much faster than with a "clicker" or "polarity checker".
Title: Re: Phase align subs to mains
Post by: Jay Barracato on March 02, 2014, 06:14:10 PM
Measurements can give you answers, but you have to know what questions you are asking.

The best measurement system can only suggest the results of choices you have made. Experience tells you what those choices might be.

I think negative information is also important. A lot of times you will hear experienced system techs talking about things that can't be fixed by alignment or eq. If you find yourself working in a circle and redoing over and over what you already changed, odds are you have stumbled against something that can't be fixed. Also, another piece of negative information is that you must be able to get consistent measurements before you can make consistent predictions based on the measurements.

Beware on anything with the word "wizard" in its name. It is the antithesis of effective measurement/testing.

Sent from my DROID RAZR HD using Tapatalk

Title: Re: Phase align subs to mains
Post by: Ryan Hammond on March 03, 2014, 09:27:14 PM
No, they will still need to be phase aligned, but the alignment will work over most of the coverage area. The phase of the signal coming out of a speaker is affected by crossover design, box design, driver selection, and electronic treatment. The idea is to use the electronic treatment to correct the errors induced by all the other influences.

Mac

Well, to be honest, I've never had to mess with it. ... and I'm glad I never had to; however, I have checked the polarity of our subwoofers just to try out the app. As for phase, it's never been an issue; although our system is an install with custom DriveRack settings.
Title: Re: Phase align subs to mains
Post by: Ivan Beaver on March 03, 2014, 09:43:31 PM
Measurements can give you answers, but you have to know what questions you are asking.

The best measurement system can only suggest the results of choices you have made. Experience tells you what those choices might be.


And the exact same thing goes for prediction programs. 

They only show you the RESULT of your "design".  They do not show or tell you what to put in.

It is your EXPERIENCE AND KNOWLEDGE that determines what might or might not work.

There is no replacement for experience.
Title: Re: Phase align subs to mains
Post by: Cailen Waddell on March 04, 2014, 08:51:33 AM
So...  Maybe my last question, when working with a system that is L/R with a mono center ground stack sub cluster, would you run pink through the whole system or just left and the subs then right and the subs.  In my head, if the subs are equidistant from the left and right, it would seem to be that if the relationship between left and sub is correct the relationship between right and sub is correct?

I am also assuming that this is a proper l/r system where each side fully covers the audience. Which maybe leads to a second question - in a stereo system, there is inherently a lot of phase interaction between sides due to spacing - yes?  Do you try to fix that or just accept that it is a disadvantage of stereo versus a mono center cluster....

I'll add that this is mostly an academic interest for me.  I know enough to know I don't know enough to do this, and with our installs especially, it's better to let  people who are good at this due system tuning. That said, playing around with our portable B right better understand the concepts is something I might do...
Title: Re: Phase align subs to mains
Post by: Tim McCulloch on March 04, 2014, 10:06:22 AM
So...  Maybe my last question, when working with a system that is L/R with a mono center ground stack sub cluster, would you run pink through the whole system or just left and the subs then right and the subs.  In my head, if the subs are equidistant from the left and right, it would seem to be that if the relationship between left and sub is correct the relationship between right and sub is correct?

I am also assuming that this is a proper l/r system where each side fully covers the audience. Which maybe leads to a second question - in a stereo system, there is inherently a lot of phase interaction between sides due to spacing - yes?  Do you try to fix that or just accept that it is a disadvantage of stereo versus a mono center cluster....

I'll add that this is mostly an academic interest for me.  I know enough to know I don't know enough to do this, and with our installs especially, it's better to let  people who are good at this due system tuning. That said, playing around with our portable B right better understand the concepts is something I might do...

You want the fewest variables in each individual measurement, so you'd start by using just the Left or Right and the subs.  Decide which seat in the house is the alignment point, put your measurement mic there (ground plane, or at least a big, flat surface like a console lid, sheet of plywood or even a couple of banquet tables laid over the seats).  I start with an impulse response measurement for the tops, and then the bottoms.  This is to show if the tops or subs are "leading" in time.  You can use Smaart's Delay Finder, but I like to look at the IR and coherency trace to see if there is anything funny going on (reflections, ground bounce, etc) that might confuse the DF.  Once I'm confident that the pass bands are relatively close I'll move to the phase alignment.

If your install is symmetrical then you can *presume* L/R to be the same, but I think it prudent to prove it to be so.

With a multiple loudspeaker system there will be time issues because every audience member occupies a different place relative to the loudspeakers.  You accept this and move on, unless there is a problem with a mono source (like the pastor's mic).  Such issues will not be eletronically corrected without creating new issues in other parts of the room, and likely indicate flaws in the basic design concept.

Set up a small rig and experiment.  Seriously.  It's the only way you'll get to understand both measurement and alignment.  After you make a couple hundred invalid measurements and recognize them on the screen, you'll begin to move much more quickly.  Trust me, I made hundreds of stupid mistakes while learning measurement; fortunately my ears told me to ignore the pretty picture on the computer...
Title: Re: Phase align subs to mains
Post by: Jay Barracato on March 04, 2014, 03:42:26 PM
Or take measurements of systems you already know are aligned/ that you like the sound of. After a while you will start to recognize what good sound likes.

When I first started measuring on a regular basis, I know I probably amused some system techs by approving the system based on playback and then setting up to take a measurement. I was looking for the attributes I liked.

Sent from my DROID RAZR HD using Tapatalk

Title: Re: Phase align subs to mains
Post by: Franz Francis on March 19, 2014, 12:32:41 PM
I’ve had the experience where the phase alignment done in the DSP with delay look good in Smaart but did not sound pleasing to the ear, I had to physically move the subs to do the alignment which also turned out to look good in Smaart but sound way better to my ears.
Title: Re: Phase align subs to mains
Post by: Raimonds.Skuruls on April 01, 2014, 09:09:17 AM
... If both cars cross the finish line at the same time (amplifying each other) but one is a lap behind,
how do you see that on measurement software?

... The phase of the signal coming out of a speaker is affected by crossover design, box design, driver
selection, and electronic treatment. The idea is to use the electronic treatment to correct the errors
induced by all the other influences.
Mac

But simply "being physically close" DOES NOT mean that they are in phase.
...  They all have an effect on the actual arrival time.  Steeper slopes cause more "delay".
...Just the simple action of putting a low pass filter on the subs will throw their "arrival time"
off-even if they are sitting right next to the mains.

Measurements can give you answers, but you have to know what questions you are asking.
The best measurement system can only suggest the results of choices you have made. Experience tells you
what those choices might be.
...If you find yourself working in a circle and redoing over and over what you already changed, odds are
you have stumbled against something that can't be fixed.
Also, another piece of negative information is that you must be able to get consistent measurements
before you can make consistent predictions based on the measurements.

Not only a PA needs alignment.
John asks me to help with his Quested V3110 studio monitors.
The results you can see in the next topic
http://forums.prosoundweb.com/index.php/topic,148306.msg1368687.html#msg1368687

BR,
Raimonds
http://aplaudio.com
Title: Re: Phase align subs to mains
Post by: Chris Johnson [UK] on April 15, 2014, 05:31:13 PM

http://timobeckmangeluid.wordpress.com/2014/02/12/eqin-a-system/

Great article Timo. Nice explanation all the way through. I'd never thought about working on the system with an inverted trace of the processors output visible. It sure makes calculating filter widths a lot easier!

Here's a question for you though: What was the physical setup here? I notice you have a wonderfully flat coherence trace in the region of interest in all your measurements (>= 95% I'd say). I presume this wasn't an "in-the-trenches" (on-site) thing, but in somewhat "lab conditions"?

One of the challenges in the real world, where lampies love to drive cherry pickers around at the moment you wan't to tune the system, is getting solid coherence on your measurements. Obviously we'd all love to work in a nice quiet room. If only gigs were like that :-). Do the 2 options for Mag Averaging have any effect on the coherence math?

Thanks,

Chris
Title: Re: Phase align subs to mains
Post by: Timo Beckman on June 14, 2014, 06:02:44 AM
Great article Timo. Nice explanation all the way through. I'd never thought about working on the system with an inverted trace of the processors output visible. It sure makes calculating filter widths a lot easier!
Here's a question for you though: What was the physical setup here? I notice you have a wonderfully flat coherence trace in the region of interest in all your measurements (>= 95% I'd say). I presume this wasn't an "in-the-trenches" (on-site) thing, but in somewhat "lab conditions"?

Thanks,
Chris

Glad you liked it. The mic's were close to the D.U.T. I try to get a semi flat response first (and if needed a similar phase response between products so i can use all together if needed).
After that's done and the speaker is used on a install/production i tune it to get a response that's both technical ok but more important: does it still sound musical.

To achieve this means i have to use a lot of mic positions and take an avarage of those positions which is coherence weighted (a option in smaart when doing averages). I still check al traces together solo on screen against the average and overlay an inverted processor trace for eq points in a room. Doing this let's you see trends in a room and with a live average on screen while doing eq points you can see if the eq points are having the effect you want.

If there's little effect with the eq points check out what maybe the cause of that. Before eq'ing a system it might also be a "cool" thing to check the splay angles of an array position of loudspeakers etc.

Eq'ing is just a small part of the outcome of a system. First verify if all speakers are working correctly (polarity connected to the right amp etc.) verify positioning and if not act accordingly time align/adjust gains on all delay's and phase align acoustic x-overs in array's. There's a lot more that can be done before you do eq's.
Title: Re: Phase align subs to mains
Post by: Timo Beckman on June 14, 2014, 06:43:02 AM
Glad you liked it. Sorry for the late response i did not see it because the topic being moved. The mic's were close to the D.U.T. I try to get a semi flat response first (and if needed a similar phase response between products so i can use all together if needed).
After that's done and the speaker is used on a install/production i tune it to get a response that's both technical ok but more important: does it still sound musical.

To achieve this means i have to use a lot of mic positions and take an avarage of those positions which is coherence weighted (a option in smaart when doing averages). I still check al traces together solo on screen against the average and overlay an inverted processor trace for eq points in a room. Doing this let's you see trends in a room and with a live average on screen while doing eq points you can see if the eq points are having the effect you want.

If there's little effect with the eq points check out what maybe the cause of that. Before eq'ing a system it might also be a "cool" thing to check the splay angles of an array position of loudspeakers etc.

Eq'ing is just a small part of the outcome of a system. First verify if all speakers are working correctly (polarity connected to the right amp etc.) verify positioning and if not act accordingly time align/adjust gains on all delay's and phase align acoustic x-overs in array's. There's a lot more that can be done before you do eq's.