Rational Acoustics



0675477985
February 9th, 2011, 08:59 PM
Hello, i am trying the Demo of Smaart 7, and i see the new possibility to filter the Impulse Response. I hear people saying that it's a new way to align speakers in a multi-way system, specially Tops (mains) and Subs.
From what i have understand the filtering allow to have the same spectral content for the two IRs, and so have the same energy for detecting PRECISELY the start of the peaks.

So it seems to be the way to have the EXACT delay of the sub, which was not possible with Smaart5/6.

Can someone know a little more or can explain to me this function ?

Thanks a lot,

Patrice

Arthur Skudra
February 10th, 2011, 03:51 PM
Hello, i am trying the Demo of Smaart 7, and i see the new possibility to filter the Impulse Response. I hear people saying that it's a new way to align speakers in a multi-way system, specially Tops (mains) and Subs.
From what i have understand the filtering allow to have the same spectral content for the two IRs, and so have the same energy for detecting PRECISELY the start of the peaks.

So it seems to be the way to have the EXACT delay of the sub, which was not possible with Smaart5/6.

Can someone know a little more or can explain to me this function ?

Thanks a lot,

Patrice
Hi Patrice,

I like your way of thinking! Certainly it's something worth exploring to at least get in the ballpark, provided that the filter is centered around the frequency of overlap between devices, and you can live with the rounded peaks in the impulse response that you would typically see with large mass low frequency drivers. Personally I still prefer using the phase curve for sub to mains alignment since I can see polarity and timing issues at the same time. Maybe it's just that I have become accustomed to doing it that way for so long! I'll have to check your method out the next time I'm out aligning a system.

Cheers!

Arthur

Rasmus Rosenberg
February 11th, 2011, 02:33 PM
Hey Pepe,
What was the different delay times? Are you sure that the impulse response for the sub is valid? it that really the first peak? In the impulse alignment did you use the first high peak for the "low" driver or the bump that comes after it as time zero? Also what settings did you use for the impulse response?

0675477985
February 16th, 2011, 10:40 AM
Hi, thanks for all your replies.

Pepe, if i have understand your Pics, you achieve a better Transfer Function using phase (instead of Impulse).

I just read the famous article of Bob MacCarthy (one of the SIM creator) who explain why we should NOT using impulse for align Main & Sub (alignment with impulse is for speaker with same Frequency Band).

All people concerned about time alignement must read this article :

http://bobmccarthy.wordpress.com/2010/02/08/phase-alignment-of-subs-why-i-dont-use-the-impulse-response/

So, i understand that the function of filtering the IR can't help us for align Main with Subs.

But if you have read the paper 'Subwoofer alignment with Full Range System" from Charlie Hugues (Ease/Easera) you know that some people are using GROUP DELAYS for aligning Subs with Mains.

Is there anybody who has already use this method ? Thanks .

merlijnv
February 17th, 2011, 04:58 AM
Bob has also written a very goog book, 2nd edition by now. You can get it at the rational store.

Arthur Skudra
February 17th, 2011, 08:46 AM
Now that we have ruled out the impulse response as a possible method, there's several methods in using the phase/group delay to find the sub to main alignment that essentially achieve the same goal, some better than others. Use the one you feel most comfortable with and helps you achieve the most accurate results. Jamie wrote an awesome post regarding the use of group delay that's worth repeating here:
http://www.rationalacoustics.com/forums/showpost.php?p=96&postcount=14

Personally I feel most comfortable using the phase curve method, getting the phase curve flat at the frequencies of interest using the delay spinners in Smaart. I find it most helpful to have at least 1 second averaging, and apply more smoothing to the phase curve than I do for the magnitude (1/12th octv for phase, 1/48th octv for magnitude).