TOPIC:

Dave,

if I could hit the thank you button 10,000 times then I would!!

that file works!!! oh snap !! I’m so excited now !!

your a good dood I appreciate it a lot 

So what happens if I change the Q in the spreadsheet?

i thought 1st order was always q.5 and can’t be changed 


thanks 

Andrew 

Try it and see what happens.

Dave.

Nothing happens …

okay okay I didn’t want to break it 


so , this moving b0 and b1 to reverse polarity….


so in the case of a 12db filter , to simulate a 12 db filter , i should be able to add the normal polarity AP1, to the low and high pass , run Dirac , take off AP1s , add the 12db XOs , and reverse polarity on the low pass……


meaning , if Dirac makes flat an AP1 , a reversed polarity LPF that’s 12db will match its shape 

Im not understanding why I would want to reverse the polarity of the AP1? 

I suppose if I did that , and ran Dirac I wouldn’t need to flip it on the LPF xo after Dirac because it would match its shape and make that linear 

hummmm …..

why would anyone want to reverse the polarity on an AP1……? 

I'm unclear on what issue you're trying to resolve with this procedure.  You're adding excess phase to existing phase distortion, doing some manipulation, and hoping the result coming out the other end is linear phase???  I'm not all that familiar with Dirac capabilities, but I don't believe it can perform magic.

The practical usage of all-pass filters has been rendered largely irrelevant by current DSP platforms that can implement frequency-independent delays.  In the old days of analog crossovers they had usefulness in creating relative delays, but with the resultant added phase distortion.
Mostly, now, AP's seem to be used for academic purposes only.

Dave.

 

Oh it’s magic lol 


so , it’s simple 

on the multichannel Diarc boxes (8x12DL and ddrc88) when using a multi-way 


procedure to get linear phase (and it works I’ve measured it) 

turn off all crossovers in dsp (put a small protection cap on tweeter perhaps 20-30mfd hi volt polypropylene) otherwise all crossovers off. 

turn on corresponding APF that matches the phase shift of intended crossover….and set the frequencies 

so if your crossovers are all LR4 and there at 80-300-3000 let’s say for example. And don’t forget to add the APF to each driver in bass management! 

so on the output of sub , use AP2-80hzQ.7

midbass AP2-80hzQ.7 and AP2-300hzQ.7 

on midrange do AP2 80hzQ.7 and AP2-300hzQ.7 and AP2-3000hzQ.7 

on tweeter AP2-3000Q.7  and AP2-80hzQ.7 

so now you have added a phase shift that matches exactly what the crossover will do and BM or LFE 

now run multi channel Dirac on all separate drivers with crossovers OFF  and APFs ON 


in Dirac DEFINE YOUR TARGET that you want , THEN divide the groups into pairs of speakers 

group 1-2 tweets 
3-4 mids 
5-6 midbass 
7-8 sub (subs) 


then pull curtains on each pair to the limits (knee) of the natural response of each pair, example tweeter may play to 700hz so you pull curtain to 700hz on tweeters. So on an so forth. Do that to all pairs… don’t change target shape, just move curtains. 

pulling the curtains way beyond with crossovers off will make the drivers responce flat and give excellent crossover behavior and summing. 
madding the APF Dirac will see the phase shift mid band where crossovers will be on each pair and correct it and make the phase flat with the magnitude. (That’s what Dirac does) 


then after Dirac is done, exit and go to plug in. And turn OFF (yes turn off the APFs) and turn ON the crossovers. (The pre defined crossovers because now you can’t change the crossover, unless you remeasure) 

whollah , linear phase crossovers that sum perfectly.


it basically does what 2ch Dirac does in multichannel mode. It’s just tricking it and yes it works, it’s repeatable, measureable, and Dirac does it very nice.  Although it is only good for one seat!  But that’s OK! 


so I’ve got LR4s at Q.7
BW18 as Q1
LR8 as two APF one Q.5 and 2nd one Q.7 
and LR2 1st order APF 


altho you could use a 1st order APF on a BW12 also as it’s close enough…. Not exact but better then nothing 

so if we can get APFs that match more crossovers that would be fantastic!!!

like I would love to know more! 


anyway hope that helps 
it act does work quite nice 

andrew 

What/who was your inspiration for that procedure??

Dave.

I came up with it actually , as fr as what inspired me? Just doing sims on rephase and seeing some APs match crossovers exactly. So I tried it and it works. 

I shared it with my friends on DIYma fourm 

about 20 competition cars are all using the same technique now and everybody agrees this is the way to go. 


one of the members Anu was kind enough to make a step-by-step manual that adopts this method. 

Comparing it to my fir crossovers, I cant tell the difference it pretty much is the same exact thing. 

I brought an 8x12 into my home (where I can make measurements without all the reflection) to verify the method and it works great the phase is smooth, clean impulse, fast step …. It works wonderfully.  


better then moving mic method with peq then 2ch Dirac. 


andrew 

If it's innovative you should write a white paper on it.

Dave.

I’ve never thought of a white paper on a product.

i thought that was just for like , theory and philosophical audio talk stuff, like the engineering stuff, for actual smart people 

but thank you, that’s very kind , yeah maybe I can do a write up and show some measurements, I think I still have them. 

andrew