Rational Acoustics

Jon McDowell
January 18th, 2011, 10:52 AM
I'm playing around with aligning a MTH-1 mid to a MTL-1X sub. The mid has 10ms of delay. I get overlap with the sub @7.5ms of delay (what I call first cross since thats the first time the sub trace overlaps the mid trace) and 12.77ms (what I call best fit of phase since it follows the mid trace a little better). Only one can be correct, but which? Or is there something I'm missing? Also, tried to ujpload images to the site, but keep getting HTTP 500 errors. I can upload text, but not images regardless of size or type.


January 18th, 2011, 01:29 PM
Hi Jon,

I can see the image, although it is small. It looks like the green and yellow are happiest together as opposed to the green and blue, but you should also see what sort of magnitude trace they give you when combined. Ideally you want the two phase traces to be exactly overlapped at your crossover freq and to have identical phase slopes for as much of the area surrounding the crossover freq as possible.


Jon McDowell
January 18th, 2011, 03:18 PM
Here is the combined response for both. Both magnitude traces are pretty close. The coherence trace looked a little better @ 7.5ms, but there is less ripple @12.77ms


Arthur Skudra
January 18th, 2011, 11:51 PM
What color trace is what here? Can you take individual snapshots of the sub and mains so we can see where they intersect, and let us know what the Smaart delay time was for each? Also, can you take your smoothing down to none or 3 pts? With smoothing as high as you have it, you will find it difficult to see any cancellation if there is any in the crossover because it's smoothed out! ;)

Kip Conner
January 20th, 2011, 09:26 AM
I would love to see raw responses of each driver in one screen shot first. How are the boxes to be arranged? Are they tight packed or are they spread because of one being flown?

If I zoom the screen and look at your first image I can see that you aren't done eq'ing the subs. There is stuff that is above the electronic cross over point that needs to be dealt with in the subs around 190Hz. You may want to run transfer functions of your processor and save those traces as "target curves" and lay those behind your actual response curves to help in your eq'ing. I think that you'll find that by fixing that 190hz that your coherence is going to improve right there.

Also looking at your phase repsonse, my guess is that there may be too much delay. The more delay that we add, the more wraps we are going to get thus more of those "Ns" in the phase display. Hopefully you aren't changing the delay locator in smaart as you go between drivers. I have seen phase responses like the one you have in print, but have never gotten anything like it on my own.

Jon McDowell
January 22nd, 2011, 01:07 AM
Finally got some more play time. The boxes are ground stacked high on top of low. The room is the iso room of my studio, so not necessarily the ideal situation, but it's 5 below outside. Here are more screen shots. I know EQ'ing isn't perfect, but close enough for this exercise. Crossover is a BSS FDS-355. Mic is an Audix TR-40 into a M-Audio Fastrack Pro. Here is the mid @ 10ms and the low @ 8.26ms.


Jon McDowell
January 22nd, 2011, 01:13 AM
Here is the mid @ 10ms and the low @ 13.292ms


Jon McDowell
January 22nd, 2011, 01:18 AM
Mid @ 10ms and low @ 8.26ms composite. This more closely matches the group delay trace when I measured the crossover electrically.


Jon McDowell
January 22nd, 2011, 01:21 AM
Mid @ 10ms low @ 13.292ms. There is more phase wrapping with this setting. BTW, BW 24db high pass @ 35.7Hz.


Kip Conner
January 23rd, 2011, 11:36 AM
Thanks for positing these. I think that you are on the right track, but it looks like you are still being held back in the EQ area since time=frequency=phase. I totally understand that you are having to measure under adverse conditions but the raw phase response of the subwoofer might be misleading you a little bit to force delay into the situation.

Let's call it an exercise and not a real world scenario. In the real world scenario (free field measurement) the subs may not react the way they are in your ISO booth. In your current situation the subs are doing something between 300 and 500 Hz that seems to be effecting your phase trace. The 8.26ms looks like it would be in your favor. Without seeing what's below the screen in the freq response it's hard to tell. If you could fix the little hump in the phase response then the traces would somewhat line up across several octaves.

Are you using a tight slope such as 24 dB per octave? I own a few of those FDS355's myself and by preference I almost always go with a LR24 between subs ans mids to keep them playing nice. For the exercise you might go as tight as you can with a 48dB slope just so you can clean up the EQ on both ends of the xover point.

Also, I'm going to assume that the 10ms of delay you have on the mid is in reference to the high freq driver and that you didn't add the delay to help create a favorable slope? If so, you might want to start the process in the with the mids to the highs before moving to the subs. Even if EV says it should be 10ms you have the tools to make each box correct. Drivers don't always react the same over time and as they become warn.. thus the frequency repsonse changes (which you already know), those changes affecting the phase response.

Looks like you are real close!