Large 2.2 system set-up assistance

[Ed Zeppeli]

Using the file from post-16, I get the best tracking with the woofer inverted and delayed 0.4 ms.
I used a 6-cycle FDW.]

Cheers. Do you mean post #19?

So rather than invert the woofer would I be better off putting the horn back in positive polarity and then delaying the woofer your set amount or would that maybe throw us off in a different direction?

 

[jtalden]

Warren,
The timing Post-19 was good, but was 1/2 cycle from the one in Post-24.
Either one would serve well in terms of SPL support.

If you are wanting the get a cleaner picture the phase charts measure with the mic roughly 50 inches from the baffle. That reduces the impact of the room on the measurements and provides a cleaner picture of the direct sound phase. It's still best to put the mic at your ear level when listening, 36-38 inches high.

 

[jtalden]

Cheers. Do you mean post #19?

Yes, I meant Post-19.

So rather than invert the woofer would I be better off putting the horn back in positive polarity and then delaying the woofer your set amount...?

Yes

 

[jtalden]

Regarding JohnM's Post-23:
This is not what I expected. Visually the rollup of the peak energy through the XO range seem about right for the 'Both' and too much for the 'A+B' which I had adjusted for good phase tracking. I am not sure what to make of this.

For a IIR XO with good phase tracking we expect a smoothly increasing delay moving to the lower frequencies. The question is how much is correct for any given XO situation. Will we be able to see detail timing or is it just a quick check for a discontinuous alignment. We could sure see the problem at the 6kHz XO before the HF horn was dropped. A closer mic distance would in this case also.

 

[Ed Zeppeli]

Here's the actual measurement showing Phase with the horn polarity flipped back and woofer delayed .4ms.

 

[Ed Zeppeli]

Data file for all of today's sweeps with the bottom three being the most recent that include the processor adjustments we've come up with.

 

[jtalden]

That look great for timing!

 

[Ed Zeppeli]

Thank you for all your detailed replies and instruction. Hopefully others will be able to make use of it here as well.

Now, onto integrating the subs; which I presume I can use this phase technique for as well.

 

[jtalden]

Thanks,
Yes,...as you probably know:

• SW positioning can greatly influence the process and quality of the results.
• The mic must be at the LP for distributed SW locations.
• First set delay on the closer SW so the 2 SWs IRs are aligned in time and the sound arrives together.
• Indeed, the process is the same as above the for XO delay setting. [It is just usually a little more difficult to get a clean direct sound phase chart as there are usually 1-3 room modes in that XO frequency range.]

 

[Ed Zeppeli]

Here's the sub on sub phase graph. Crossovers set to Low Pass 70Hz LR24. Of course this isn't showing the crossover with the woofers but the way the subs are interacting with each other.

Things get real slow down here. I delayed the rear sub 16.73 ms and had to flip it's polarity old school. I got my hands dirty on this one and it was quite distressing....such manual labour in a digital world.

 

[jtalden]

Good first step. The phase tracking looks really good. No apparent issues with room modes in either SW >40 Hz Much better than the others I've seen. If the SPL looks reasonable to EQ within a 12dB or so range, then you have a good situation for the SWs.

Unfortunately the modal response of the woofers may still make the XO range difficult. That's the next step though.

 

[Ed Zeppeli]

I'm a little concerned that when I implement EQ that their inherent delay will change all this phase timing work. Do you revisit the crossover phase graphs after equalization and re-adjust?

 

[jtalden]

I may have mislead you. No need for EQ work at this point. It was more a comment to suggest that if the SPL is wildly deviating then consideration may be given to moving one or both SWs to find a better situation. If EQ is applied now it is best to stay below the XO range as the EQ in range will very likely be impacted by the addition of the woofers. It probably would not hurt the process to EQ now, but there is no need to do the job twice.

More Generally:
EQ phase impact is relatively minor compared to the XO filter impact and thus the timing is not effected significantly. EQ does impacts the phase, but that normally is good as the phase is corrected as the SPL is corrected in the minimum phase portions of the response.

We can always revisit the XO timing after the final EQ is established. I do this, and using modest EQ, have not found need to change the timing.

The big issue here is that this XO is in the modal range. The standing waves cause major disruption and severely limit EQ control. The modes are different for the SWs vs the mains due to their being located in parts of the room. This can create a rather chaotic phase interaction in the XO range. Determining the best timing then can be problematic due to these interactions. In my room it is best to just set the timing for the direct sound and accept the modal disruption that occurs. There were tradeoffs in the other timings I experimented and none really improved the overall situation significantly. I cannot say for sure if there may be cases where another delay timing may actual result in better EQ control and smoother SPL.

 

[Ed Zeppeli]

More Generally:
EQ phase impact is relatively minor compared to the XO filter impact and thus the timing is not effected significantly. EQ does impacts the phase, but that normally is good as the phase is corrected as the SPL is corrected in the minimum phase portions of the response.

We can always revisit the XO timing after the final EQ is established. I do this, and using modest EQ, have not found need to change the timing.

Good to know. Thanks

Here's where I'm at today. Since the subs are synced up (and to bring the levels back up to approximately match the LF Woofs) I decided to run them together when aligning phase with the left woofer. To get an acoustic crossover of about 70Hz I ran the subs Low Pass at 45Hz LR24 and the Woofer High Pass at 50Hz LR12dB. The impulse responses, left woofer versus subs, looks ugly but here's the phase unwrapped as estimated in REW.

 

[Ed Zeppeli]

Here's the actual after running another few sweeps. What's that all about? Room mode?

 

[John Mulcahy]

Data file for all of today's sweeps with the bottom three being the most recent that include the processor adjustments we've come up with.

Here is the spectrogram for 'both'.

 

[jtalden]

Since the subs are synced up (and to bring the levels back up to approximately match the LF Woofs) I decided to run them together when aligning phase with the left woofer.
To get an acoustic crossover of about 70Hz I ran the subs Low Pass at 45Hz LR24 and the Woofer High Pass at 50Hz LR12dB. The impulse responses, left woofer versus subs, looks ugly but here's the phase unwrapped as estimated in REW.

The result looks great. I assumed you will use them mono and thus it's the combined measurement that counts.

Here's the actual after running another few sweeps. What's that all about? Room mode?

No problems with phase there. Was that question followed with a big , or was that a true question on some point I'm missing?

 

[Ed Zeppeli]

The result looks great. I assumed you will use them mono and thus it's the combined measurement that counts.



No problems with phase there. Was that question followed with a big , or was that a true question on some point I'm missing?

I'm quite happy with the curve except for that purple (woofer) phase drop which corresponds with a suckout at about 60Hz in the SPL graph.

 

[jtalden]

The phase 'drop' at 60Hz is really a very slight phase 'rise' of a few degrees. The trace fell above 180° in that range so it was wrapped to the bottom of the chart for that short segment above 180°. The 2 traces are thus actually very close to each other in that range.

 

[jtalden]

Here is the spectrogram for 'both'.

View attachment 3820

John, Is there a message in the chart that we are overlooking?

It looks a lot cleaner than my setup at the LP and of most others I have looked at.
The appearance suggesting some excess delay timing around the 600Hz XO, but assume that it would look more normal if the mic was closer to the speaker and thus the room influence was less.

I'm thinking the 400hz null is floor/ceiling bounce or some other strong spectral interference.

 

[Ed Zeppeli]

I do love the new graphic John and am looking forward to trying it out for myself. Thanks so much for previewing it here using my data. Also, this form software is a real step up from VBulletin; which I'm used to.

As I dig into REW and start learning more of its intricacies the only thing I find odd is the repetition necessary to relatively align multiple sweeps in the Impulse windows when seeking a 'near zero' IR overlay. It would be nice to be able to Shift All or Shift All Selected measurements the same amount with one command. Also, a software recording of what shifts have taken place - either manually or from the Impulse Estimate IR Delay would be handy. When I've been playing with charts for over an hour - despite taking notes - I find myself re-doing sweeps because I'm unsure of what my initial offset was.

To end on a positive; Thank you so much for providing this amazing piece of analysis software for us. I'm putting that Donate button on my to do list.

Cheers,

Warren

 

[John Mulcahy]

John, Is there a message in the chart that we are overlooking?

No, just providing it as a comparison of the peak energy time based view.