ProSoundWeb Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: 1 [2] 3 4 5   Go Down

Author Topic: Eclipse Audio FIR Designer  (Read 10822 times)

Mark Wilkinson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1104
Re: Eclipse Audio FIR Designer
« Reply #10 on: September 09, 2021, 03:10:45 PM »

Thought I'd add a follow on post about how i use FirDesigner to tune speakers....and how helpful i find FirD's Auto Mag, & Auto Phase capability.
Don't mean to be contradicting Timo, who has far more experience with this sort of thing than i do. 
Just thought I'd throw out what seems to be working well for me, and ask for critique....

The Smaart screen grab below is of a 4-way DIY synergy.
It has a pair of 12"s, four 4"s, and a dcx464 coax CD.  Its straight sided conical horn is 75x60 degrees, about 34" x25".
(It's a new project that still needs a secondary flare, which  should eventually help smooth out off-axis responses further.)



First step is taking raw measurements for each driver section separately, from on-axis out to whatever off-axis angle i want to average together.
0, 10, 20 and 30 degree measurements were averaged for each section.  (Both FirD and Smaart have a great averager...i tend to use Smaart's because that automatically saves the trace with the raw measurements.)

Next step is to choose xover points and slopes based on what seen so far.  In this case, I chose a 100Hz high-pass, 12"s to cross to 4"s at 250Hz, 4"s to dcx-mid at 650Hz, dcx-mid to dcx-HF at 3800Hz.

Then I put those xovers into FirD as target curves for each section, and import each sections' response traces average. (one driver section at a time).
FirD then matches driver's response to the target curves.
Since the measurement I'm using for matching is an average, I generally import the average trace with 1/12th octave smoothing.  And then also use Auto Mag/Phase at 1/12th smoothing for auto-correction.  (If it were a single measurement, 1-12th is usually too aggressive to match response to, and more smoothing is often needed. I know Michael John cautions against overcorrection.)

Ok, FirD generates a FIR file for each section, they go into the processor, and delays and levels are all that needs to be set.
It sounds like a long process, but like anything it gets easier and quicker with repetition.  I can do the whole 4-way from scratch in around a hour or so now.

Anyway, here's 0,10,15,20, and 30 deg horizontal traces.
To me at least, they look very good...(and more importantly, the box sounds great !!)
So I'm all for the Auto Mag/Phase taking the work out of dialing in EQ's etc :)






Logged

Don Davis

  • Full Member
  • ***
  • Offline Offline
  • Posts: 162
  • So Cal
Re: Eclipse Audio FIR Designer
« Reply #11 on: September 10, 2021, 03:03:32 PM »

Thought I'd add a follow on post about how i use FirDesigner to tune speakers....and how helpful i find FirD's Auto Mag, & Auto Phase capability.
Don't mean to be contradicting Timo, who has far more experience with this sort of thing than i do. 
Just thought I'd throw out what seems to be working well for me, and ask for critique....

The Smaart screen grab below is of a 4-way DIY synergy.
It has a pair of 12"s, four 4"s, and a dcx464 coax CD.  Its straight sided conical horn is 75x60 degrees, about 34" x25".
(It's a new project that still needs a secondary flare, which  should eventually help smooth out off-axis responses further.)

First step is taking raw measurements for each driver section separately, from on-axis out to whatever off-axis angle i want to average together.
0, 10, 20 and 30 degree measurements were averaged for each section.  (Both FirD and Smaart have a great averager...i tend to use Smaart's because that automatically saves the trace with the raw measurements.)

Next step is to choose xover points and slopes based on what seen so far.  In this case, I chose a 100Hz high-pass, 12"s to cross to 4"s at 250Hz, 4"s to dcx-mid at 650Hz, dcx-mid to dcx-HF at 3800Hz.

Then I put those xovers into FirD as target curves for each section, and import each sections' response traces average. (one driver section at a time).
FirD then matches driver's response to the target curves.
Since the measurement I'm using for matching is an average, I generally import the average trace with 1/12th octave smoothing.  And then also use Auto Mag/Phase at 1/12th smoothing for auto-correction.  (If it were a single measurement, 1-12th is usually too aggressive to match response to, and more smoothing is often needed. I know Michael John cautions against overcorrection.)

Ok, FirD generates a FIR file for each section, they go into the processor, and delays and levels are all that needs to be set.
It sounds like a long process, but like anything it gets easier and quicker with repetition.  I can do the whole 4-way from scratch in around a hour or so now.

Anyway, here's 0,10,15,20, and 30 deg horizontal traces.
To me at least, they look very good...(and more importantly, the box sounds great !!)
So I'm all for the Auto Mag/Phase taking the work out of dialing in EQ's etc :)
Amazing Mark, just how many projects do you have gong on? Between you and Peter I can't keep up.
Looks fantastic BTW
« Last Edit: September 10, 2021, 04:02:40 PM by Mac Kerr »
Logged

Mark Wilkinson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1104
Re: Eclipse Audio FIR Designer
« Reply #12 on: September 13, 2021, 08:40:11 AM »

Amazing Mark, just how many projects do you have gong on? Between you and Peter I can't keep up.
Looks fantastic BTW

Hi Don,  kind words, thank you.
Hope all as been well for you.

Too many projects  haha....
Peter's design and your drawings really got me started years back.  I had so much fun with the PM90 & 60, I've just kept going with one project after another...at least a dozen.
Wish all the builds had stayed about live sound though (that's where my heart and respect for folks know-how stays; hi-fi circles are too often plain nuts, Lol).
Unfortunately, the very limited local live-sound partnership opportunities didn't pan out.

I really would like to try this latest synergy project out live though...it's about as loud as the PM boxes, weighs under 70 lbs, and is the best sounding i've managed yet.
Really rocks under some Labhorns .  (Mr Danley also surely knows how to inspire a DIYer!)
Logged

Helge A Bentsen

  • Hero Member
  • *****
  • Online Online
  • Posts: 1766
  • Oslo, Norway.
Re: Eclipse Audio FIR Designer
« Reply #13 on: October 03, 2021, 12:35:15 PM »

Got a call from the guy building my boxes, they are ready for pick-up.
Next step is measurements, just have to find some time for it. Calendar has exploded with gigs after Norway opened up a week ago.

If I understand this correctly, most of the posters here base their XO/EQ decision in FIR Designer off an average of X measurements from onax to whatever off-axis angle of choice. Pretty similar to what I do manually now. Going to try averaging in Smaart or Fir Designer when I get around to measuring.

But I have a question.
Where do you determine alignment delay between passbands? At a fixed position in the middle between them or at some other position?

I've found that with a 2-way speaker, if you place your mic on-axis at distance X in the middle between LF/HF and use this as your reference point for alignment delay, sometimes you have to adjust this delay slightly if you put the speaker on a stand above the audience for better summation in the audience area. I suspect this is a common issue for every speaker. Do you "down-tilt" the alignment delay on purpose or do you go for maximum summation on-axis and live with a bit more variation in the audience area?
Logged

Mark Wilkinson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1104
Re: Eclipse Audio FIR Designer
« Reply #14 on: October 05, 2021, 09:34:15 AM »


But I have a question.
Where do you determine alignment delay between passbands? At a fixed position in the middle between them or at some other position?

I've found that with a 2-way speaker, if you place your mic on-axis at distance X in the middle between LF/HF and use this as your reference point for alignment delay, sometimes you have to adjust this delay slightly if you put the speaker on a stand above the audience for better summation in the audience area. I suspect this is a common issue for every speaker. Do you "down-tilt" the alignment delay on purpose or do you go for maximum summation on-axis and live with a bit more variation in the audience area?

Yeah, the inevitable geometry of multiple sources, and where to time align.  Same ole dilemma of where in the audience to align subs to mains, only generally alot less of a deal.

I like to get the mic as reasonably far away as possible and i can still make good measurements, on-axis and kinda centered between acoustic centers.
Which for me is about 4m outdoors.
That has worked pretty well for keeping time alignment at distances both closer and farther, especially farther since angular summation errors get smaller with distance.

That said, if i had a speaker that couldn't be tilted, and some members of the audience will be up close, i'd probably time align to the off axis tilt.  Again with the thinking that summation errors (lobing) decrease with distance.



In the grandest scheme, I really don't think it matters too much as long as the time alignment isn't made too close....if too close it's like your eyeball muscles focused in for reading, which doesn't work for distance vision....
Logged

Helge A Bentsen

  • Hero Member
  • *****
  • Online Online
  • Posts: 1766
  • Oslo, Norway.
Re: Eclipse Audio FIR Designer
« Reply #15 on: December 08, 2021, 02:26:44 PM »

This week new restrictions was imposed on live shows because you know what. Bad for business, good for measuring speakers.

Made a DIY turntable and measured today, imported the data into FIR designer and hit a "snag".
Phase in the top end varies a bit with angle, probably because I rotated the speaker and didn't adjust Ref dly in Smaart for each measurement.

Is this normal or should I go back, redo my measurements and adjust Ref dly for each angle/position?
Or is there something else I missed in the tutorial?

Logged

Chris Grimshaw

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1822
  • Sheffield, UK
    • Grimshaw Audio
Re: Eclipse Audio FIR Designer
« Reply #16 on: December 09, 2021, 08:21:39 AM »

How much is "a bit", and how high in frequency?

Chances are I wouldn't worry about it.

Chris
Logged
Sheffield-based sound engineering.
www.grimshawaudio.com

Russell Ault

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2499
  • Edmonton, AB
Re: Eclipse Audio FIR Designer
« Reply #17 on: December 09, 2021, 10:05:01 AM »

{...} Phase in the top end varies a bit with angle, probably because I rotated the speaker and didn't adjust Ref dly in Smaart for each measurement. {...}

I could be mistaken, but wouldn't this phase shift indicate that the distance between source and microphone is changing slightly as you rotate?

-Russ
Logged

Mark Wilkinson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1104
Re: Eclipse Audio FIR Designer
« Reply #18 on: December 09, 2021, 12:12:51 PM »

I could be mistaken, but wouldn't this phase shift indicate that the distance between source and microphone is changing slightly as you rotate?

-Russ

Yeppers.  Just means the acoustic center of the CD (or whatever driver is dominating the impulse response) is not centered over the turntable's axis.

If you are using an average of a few angles generate the FIR correction, I'm not exactly sure how FirDesigner's various averaging options handle phase..over my pay grade still.
I've been using Smaart's Averager with default settings, because it makes for a little easier workflow.

If the measurements are simply for displaying polars, where you want to show phase too, I usually reset delay finder for each meas.
Logged

Helge A Bentsen

  • Hero Member
  • *****
  • Online Online
  • Posts: 1766
  • Oslo, Norway.
Re: Eclipse Audio FIR Designer
« Reply #19 on: December 09, 2021, 03:19:32 PM »

I could be mistaken, but wouldn't this phase shift indicate that the distance between source and microphone is changing slightly as you rotate?

-Russ

I'm thinking the same, it's mostly in the top two octaves of the horn.
Which makes sense since since I didn't pay attention to where the horn was relative to the center point of the turntable.

O well, learned something new :)


It's mostly visible when I load all the different measurements in supplementary view.
Logged

ProSoundWeb Community

Re: Eclipse Audio FIR Designer
« Reply #19 on: December 09, 2021, 03:19:32 PM »


Pages: 1 [2] 3 4 5   Go Up
 



Site Hosted By Ashdown Technologies, Inc.

Page created in 0.032 seconds with 25 queries.