Measurement Question

B

Brent Venter

Freshman
Mar 31, 2011
22
0
1
Hi Guys,

So I got to tune an install over the last two days and because I had extra time I played around to get the maximum out of the experience. While aligning the under balcony speakers to the main hang I came across something I hope to get some insight on.

In the attached photo's there are two traces, purple is the under balcony speaker and blue live trace the under balcony summed with the main hang. In the polarity correct measurement you can clearly see the summation around 200Hz-300Hz (under balcony high-passed), however the coherence and HF in the 10kHz-20kHz range is poor. You can see in the Phase trace the correct polarity in this region as well.

In the polarity incorrect photo it is very clear the non summation in 200Hz-300Hz, however in the HF region the coherence is much better and summation in this region compared to the polarity correct measurement. Did this summation show up because the coherence improved and was there all along? and why would the coherence improve relevant to the polarity?

I know I didn't show the measurement of the main hang separately, I ran out of trace memories however I did confirm polarity.

PS. polarity incorrect means under balcony reversed compared to the main hang and vice versa.

Thanks

Polarity_Correct_SFN.pngPolarity_Incorrect_SFN.png
 
Re: Measurement Question

It looks like you do not have your delay set properly in the measurement system-there is a lot of wrapping going on.

It that is off-then all sorts of other things will be off as well-specifically in the high end.

How did you go about setting the delay time? There are a number of different "schools of thought" on that one.
 
Re: Measurement Question

Ivan Beaver said:
It looks like you do not have your delay set properly in the measurement system-there is a lot of wrapping going on.

It that is off-then all sorts of other things will be off as well-specifically in the high end.

How did you go about setting the delay time? There are a number of different "schools of thought" on that one.
Click to expand...

Hi Ivan,

I set the receive delay using the delay finder and then confirming with the impulse response to verify polarity and delay. I will sometimes adjust receive delay to flatten the passband that I might need to work with, however normally prefer the flat part in the HF.
 
Re: Measurement Question

Brent Venter said:
Hi Ivan,

I set the receive delay using the delay finder and then confirming with the impulse response to verify polarity and delay. I will sometimes adjust receive delay to flatten the passband that I might need to work with, however normally prefer the flat part in the HF.
Click to expand...
My approach for delays is basically this.

First get a reference measurement from the center or main part of the main PA coverage. Everything else is referenced back to this response.

Next measure in the delay coverage zone. See what is "missing" from the main PA. Put a highpass filter and eq the delay zone (main PA off) till it has a reasonably flat response.

Find the delay time to the measurement location from the main PA (delay off), then turn off the main PA and find the delay time to delay (with no delay inserted in the DSP).

Subtract the 2 and enter the difference in the DSP.

Turn the main PA on and the delay and see how well they sum together. Adjust delay level and possibly eq as needed.

Do not get to hung up on a single measurement position. Look at other positions in the coverage area. Remember that the distance is not going to change much from the main PA-but will change quite a bit from the delay speaker-so averages is what we are looking for here.

You can only get it exact for a single seat-every other seat is going to be "off".

Measurement is the fine art of compromise and knowing what to "ignore" or what not to worry about.
 
Re: Measurement Question

Ivan Beaver said:
My approach for delays is basically this.

First get a reference measurement from the center or main part of the main PA coverage. Everything else is referenced back to this response.

Next measure in the delay coverage zone. See what is "missing" from the main PA. Put a highpass filter and eq the delay zone (main PA off) till it has a reasonably flat response.

Find the delay time to the measurement location from the main PA (delay off), then turn off the main PA and find the delay time to delay (with no delay inserted in the DSP).

Subtract the 2 and enter the difference in the DSP.

Turn the main PA on and the delay and see how well they sum together. Adjust delay level and possibly eq as needed.
Click to expand...

Thanks for the response. This is exactly my workflow and what I have done here. I was about to include an attachment of my on-axis reference measurement and just figured out my platform had defaulted to my previous session ughhh...

Anyway I used my ears in the end to make the final decision and am not second guessing it, I was just hoping to clarify this phenomena as I have not come across it before.

Soundwaves-1 Brent-0

Thanks again.
 
Re: Measurement Question

Tim McCulloch said:
Not pretty but that's okay.

Did you apply the under-balc HPF filter after using delay finder?
Click to expand...

Hi Tim, yes not pretty as this venue is far from anechoic, however I had on-axis reference traces (not shown) to compare against and help guide the tuning strategy.

Yes I had applied the HPF after using delay finder as I measured both systems to see where the HPF should be set to best achieve my reference.
 
Re: Measurement Question

Hey Brent,

Nice messy measurement, always challenging when you're looking at graphs like these.

Looks to me like the LF is out of phase by 120° or so when the HF polarity is "correct". Probably the two speakers have very different phase slopes, or perhaps just as a result of your HPF on the fills. You could confirm with a nearer measurement that had better signal to noise. IMHO, the phase/polarity of the HF is unimportant compared to lower frequencies. If the HF is all hashed up then you will have very localized interference that your ears will largely ignore, as usually each ear will be in a different acoustic environment. Mess up 400Hz though and the response nulls and peaks will be much physically larger, if I had to pick I would go for good alignment across a wider area in the mid frequencies and let the above 8kHz stuff slip out of alignment.

Your real problem is likely the different slopes of the phase trace in each speaker system. Can you find some way to make them closer? This is one reason it's nice to have fills built with good hardware, as there are a few tricks I can pull to get them to match the phase of the mains. I always prefer to make compromises in the fills rather than add phase rotations to the mains. The more the two traces differ the narrower a bandwidth you're going to have to hit with your alignment.
 
Top Bottom