DiyMobileAudio.com Car Stereo Forum banner
161 - 180 of 321 Posts

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #161 ·
And btw 3.1.2 sucks
It works very very slow it took like 10 minutes just to resume my project

and the disable gain and delay button didn’t even work it crashed the software so I downgraded back to 3.1.1
 

·
Registered
Joined
·
505 Posts
The fact that the crossovers are turned off and targets made would definitely improve things as it would smooth out the response, it’s how there summing..... it seems to me still flat is the way , no default targets.... no tilt..... I don’t see it working optimally any other way .... yet ;-)

Dirac wasn’t made to tune a multi-way, were just utilizing it this way , so it needs to be proper to have the best results, so default target means nothing in this type of setup
Do you agree with Ryan's method of grouping the mid-basses/sub together? I found that makes a really big difference, and am definitely a fan of that approach. I tried that at the same time as doing the Dirac curve. I would like to try Ryan's method of grouping MBs/sub together along with your flat curve approach. But I likely won't be able to until the Meet itself as I'm feeding my other car addiction next weekend (and weekend after is the meet).
 
  • Like
Reactions: Truthunter

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #163 ·
Do you agree with Ryan's method of grouping the mid-basses/sub together? I found that makes a really big difference, and am definitely a fan of that approach. I tried that at the same time as doing the Dirac curve. I would like to try Ryan's method of grouping MBs/sub together along with your flat curve approach. But I likely won't be able to until the Meet itself as I'm feeding my other car addiction next weekend (and weekend after is the meet).
yeah I do agree w Ryan , as doing anything to get the speaker to behave flat throughout the crossover massively helps.

I see why he’s adding the midbass and sub together, maybe he’s trying to get the phase rotation out of the bass management.... I can see that working yes , I also believe all of what he’s doing will definitely sound good.... and it might sound better .... only one way to find out

But as far as getting the crossovers themself to behave the most textbook at your position and knowing there’s going to be phase wraps , but they will sum properly, my way is preferable...

do I think it honestly comes down to how well the initial delays are done. And a few other considerations, like the GD in where midbass and subs are grouped together. It may work on some and may nit on others

doing everything on its own channels in Dirac
Tweets get a paired group
Mids get a paired grouped
Midbass get a paired group
And sub gets it’s own single group

dirac runs before bass management. So the crossover in bass management is off when Dirac is running sweeps anyways

for me it’s about proper summing on all the crossovers.... so I wouldn’t group anything different. And I wouldn’t add a tilt to any targets....except the sub.

after Dirac, the sub is separated from all the highs in bass management. The bass manager will give the midbass , midrange and Tweets the inherent phase shift from midbass crossover in bass management

that leaves the sun on its own separate thing as bass management LP side only goes to sub so a tilt on the sub won’t affect the highs... and it’s the highs that need proper summation more then the lows.....

As far as the sub tilt , turn on a LR4 and in the minidsp you can put the mouse over the shape on the crossover screen , it will tell you how many dB down the slope is on the crossover picture in minidsp app.
So look where the top of the knee of the crossover for sun goes to 0db and start your tilt up to 10hz from that point ... that way you don’t affect the knee of the crossover and everything stays in relative gain

my way is repeatable for everyone ..... at least knowing how textbook crossovers work
 

·
Registered
Joined
·
319 Posts
The only thing about using a tilt in the response is... you risk having errors in the crossovers.....

all of the crossovers in the minidsp are calculated for flat magnitude. That means the high pass and low pass need to sum together to fill the hole caused by a -3/-6 dB down..... filling the hole isn’t the problem.

it’s the phase tracking that needs requires both sides of a crossover to be of equal amplitude and the same shape for the crossover to work properly

As the crossover attenuate the hi pass and low pass are drifting to -180deg (there completely out of phase with one another) and that 180deg mark is the crossover point for a LR and the 90deg mark for BWs .... the two speakers one at -180 and the one at +180 sum to be zero degrees. And more importantly throughout the leg of the crossover it has to match the amplitude and shape of the adjacent crossover to sum properly. So if theres an amplitude change (from a tilt or a peak or dip in the response without the crossover being engaged) it runs the risk of destructive interference. Which is the number 1 reason I stopped using the 8x12. So all the coefficients need to be there before any tilt is applied.

so I need to think this one through all the way..... the logic of adding a tilt may or may not work.... I mean, it sounds like it’s working for you guys... I suppose if the tilt is at the same Rate then maybe .... all except the logarithmic way frequencie and phase interact.
Phase gets smaller as frequencies rise. So the tilt would have more turn on the low pass then the high pass

The way you guys are turning off crossovers and grouping I need to think this through.... I’m not fully on board yet, but it might work , I’m not an expert, but I am a thinker , just sometimes a slow thinker. :)
This was my initial worry all along... I just do not prefer a ruler flat response in a vehicle. But, until I've actually heard a system setup this way I cannot say it's "wrong". Sounds like everyone whose done it thinks its an improvement so there's got to be something to it, just not sure what it is.

I wonder if sending this method over to the support folks at Dirac or MiniDSP would get us any response? Maybe they could explain why this is or is not a good way of setting things up?
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #165 · (Edited)
This was my initial worry all along... I just do not prefer a ruler flat response in a vehicle. But, until I've actually heard a system setup this way I cannot say it's "wrong". Sounds like everyone whose done it thinks its an improvement so there's got to be something to it, just not sure what it is.

I wonder if sending this method over to the support folks at Dirac or MiniDSP would get us any response? Maybe they could explain why this is or is not a good way of setting things up?
they told us this is the way to do it years ago, I just didn’t understand it at the time

And no one is saying ruler flat is how you listen that’s just how you measure it initially you add your tilt with a shelving filter after Dirac

That tundra I did sure did sound good flat tho ..... and a tilt on sub only .....

i use flat in my car on most of my drivers and you definitely want it flat from 300hz-3.5khz,

which brings up a whole Nother topic but the way that Direc does a flat response with no tilt sure does sound good..
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #166 ·
The “problem” doing it this way is Dirac won’t undo the phase twist caused by crossovers.... here’s the catch

what sounds better .... reduced phase twist with crossovers that sum pretty badly. OR

excellent summing crossover and GD removed from each passband individually by drivers

experience has shown us , keeping the phase twist and having better summation in crossovers sounds arguably better.

in a 2ch Dirac tune, a super solid pre tune could also sound excellent.

it’s just a different style of tuning... and I think that’s why we have four presets

do a different style On each preset
 

·
Registered
Joined
·
2,941 Posts
Anyone have any suggestions on how to set up rear fill while using the multi channel Dirac? Here is what my routing tab currently looks like.

Rectangle Font Parallel Pattern Screenshot


7 x Dirac already in use, for rear fill though I need to have 2 x Dirac channels available in order to set it up properly. If I assign rear fill to Dirac 7 and Dirac 8 like so:

Product Rectangle Font Parallel Pattern


The problem as you can now see, my subwoofer on Bass Mgt is also on Dirac 7, so this will not work. As is, I cannot think of a way to make it work without removing the subwoofer from Dirac and manually tuning that instead. Maybe that could work, but at the moment my subwoofer / midbass blend is better than I could ever get manually, so I am reluctant to pull the subwoofer out.

I thought about maybe putting the subwoofer in with the midbass (since they are getting grouped in the Dirac software curve) but not sure that will work either as it then gives issues with Bass Mgt since I don't want that applied to the midbass.

So, other than MiniDSP giving me a Dirac 9, can anyone see a way to keep multi channel and add rear fill?
 

·
Registered
Joined
·
505 Posts
Do you agree with Ryan's method of grouping the mid-basses/sub together? I found that makes a really big difference, and am definitely a fan of that approach. I tried that at the same time as doing the Dirac curve. I would like to try Ryan's method of grouping MBs/sub together along with your flat curve approach. But I likely won't be able to until the Meet itself as I'm feeding my other car addiction next weekend (and weekend after is the meet).
Btw, I realized I could actually try grouped MBs/sub + Flat curve pretty easily, using my existing Dirac measurements...basically just needed to open up Dirac again and edit the curves in there (no new measurements needed). So I spent 10 min doing that during lunch today. It sounds pretty good. I can't really say whether it's better or worse than Ryan's approach. I will say that I definitely think grouping the subs and mid-basses together works really damn well, so any 7ch tunes I do in the future will be using that grouping approach.

Seems to me like it doesn't really matter between drawing flat vs. using Dirac curve, so long as you apply proper shelves afterward to hit your target. But inconclusive (for me) for now.
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
1,832 Posts
Anyone have any suggestions on how to set up rear fill while using the multi channel Dirac? Here is what my routing tab currently looks like.

View attachment 314149

7 x Dirac already in use, for rear fill though I need to have 2 x Dirac channels available in order to set it up properly. If I assign rear fill to Dirac 7 and Dirac 8 like so:

View attachment 314150

The problem as you can now see, my subwoofer on Bass Mgt is also on Dirac 7, so this will not work. As is, I cannot think of a way to make it work without removing the subwoofer from Dirac and manually tuning that instead. Maybe that could work, but at the moment my subwoofer / midbass blend is better than I could ever get manually, so I am reluctant to pull the subwoofer out.

I thought about maybe putting the subwoofer in with the midbass (since they are getting grouped in the Dirac software curve) but not sure that will work either as it then gives issues with Bass Mgt since I don't want that applied to the midbass.

So, other than MiniDSP giving me a Dirac 9, can anyone see a way to keep multi channel and add rear fill?

My first thought is to pretune the mids/tweets the "old fashioned" way and put them on two Dirac channels (L/R). So CH1 is L tweet/mid, CH2 R tweet/mid, CH3-5 midbass and sub, 7 & 8 Rear fill.
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #170 ·
Anyone have any suggestions on how to set up rear fill while using the multi channel Dirac? Here is what my routing tab currently looks like.

View attachment 314149

7 x Dirac already in use, for rear fill though I need to have 2 x Dirac channels available in order to set it up properly. If I assign rear fill to Dirac 7 and Dirac 8 like so:

View attachment 314150

The problem as you can now see, my subwoofer on Bass Mgt is also on Dirac 7, so this will not work. As is, I cannot think of a way to make it work without removing the subwoofer from Dirac and manually tuning that instead. Maybe that could work, but at the moment my subwoofer / midbass blend is better than I could ever get manually, so I am reluctant to pull the subwoofer out.

I thought about maybe putting the subwoofer in with the midbass (since they are getting grouped in the Dirac software curve) but not sure that will work either as it then gives issues with Bass Mgt since I don't want that applied to the midbass.

So, other than MiniDSP giving me a Dirac 9, can anyone see a way to keep multi channel and add rear fill?
Do it from the mixer instead ????

it seems to me I remember being able to do it from the mixer instead of the routing. I’m pretty sure you can right click and do the invert on that screen as well

I think you can do it on any of the screens actually I think you can even do it on the input before base management... It’s pretty versatile but I think for what you’re saying I think the mixer would be correct if I remember correctly

doing it from the mixer would be after Dirac... Which wouldn’t cause any problems with the sub
 

·
Registered
Joined
·
2,941 Posts
My first thought is to pretune the mids/tweets the "old fashioned" way and put them on two Dirac channels (L/R). So CH1 is L tweet/mid, CH2 R tweet/mid, CH3-5 midbass and sub, 7 & 8 Rear fill.
Eww the old fashioned manual way!! That may be the only option though.

it seems to me I remember being able to do it from the mixer instead of the routing. I’m pretty sure you can right click and do the invert on that screen as well
I'll check, but I'm not sure you can do it on the Mixer tab. Would be nice if you could, I'll mess with it some tomorrow. At the moment, I don't have the rears back in the car, just trying to see if it's feasible to include them as based on past experiences, adding them into the multi channel tune, should at least in theory, make the stage huge.
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #172 ·
Eww the old fashioned manual way!! That may be the only option though.



I'll check, but I'm not sure you can do it on the Mixer tab. Would be nice if you could, I'll mess with it some tomorrow. At the moment, I don't have the rears back in the car, just trying to see if it's feasible to include them as based on past experiences, adding them into the multi channel tune, should at least in theory, make the stage huge.
yeah I’m like 99% sure you can do it anywhere in there..... even in the input section.... let us know tho to validate
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #173 · (Edited)
Just did a full blown retune on my car and was like really careful and looked to see what was minimum phase and what wasn’t... could only really tell under 1k

anyways , I turned all my crossovers off and did peq tune on most of the peaks and just did eq on left and right together on all drivers and made them as flat as I can.

I applied eq to both left and right bit only played noise on the left, leaving right muted. Then played right and compared.

for the right I only tamed any massive peaks and removed eq where it wasn’t needed not adding any more peq bands under 1k

above 1k was sane technique except I added more peq band to right to make it match left.

all done moving mic RTA pink PN 16k samples

Then turned back on crossovers and ran Dirac 2ch

it works very good.

so this is completely different then Dirac 1.7 , I could not do a peq pre tune in 1.7
So I always preached no pre tune.... 3.1.1 is doing something different

after experiencing this new 8x12 thing and reading all the enhancements and upgrades on Dirac webpage as it goes from 1.7 to 3.1 there’s many times it said something of “improve phase response between left and right”

so...... a pre tune works now. And works really good. So ...... maybe . We need to do what we’re doing in the 7ch tune but with peq and manual tuning then run a 2ch Dirac ......
volunteers??

Edit , I think the goal with a pretune is Trying very hard to make left and right have the same EQ bands in the same cuts and boozTrying very hard to make left and right have the same EQ Bands as much as possible.... can’t just start over completely on the right after doing left. That will definitely create a De-correlation filter. I only added like 1PEQ per pair and got it pretty dam flat. +/-3db
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #174 · (Edited)
So speaker tuning (crossover tuning) absolutely should be done at the listen position.....

so, awhile ago I was talking to thomas drugeon aka pos (creator or rephase) on the rephase thread on Diy. He said do crossover tuning at the speaker..... but he also said , you can do it at the listening position but it’s only good got one spot...

im beginning to feel it’s the best spot (for us) in a car.

close mic measurements do reveal measured phase very nice... super easy to read and no FDW needed. For us tho.... at our seat and a FDW works much better.

i went on a listen and it’s way way way way better.... my tweeters don’t sound brash no more .

so we’re on to something, I think we can make a 2ch tune be pretty dam bad ass if we do the 7ch method manually and do a 2ch Dirac......

@Ryan,
I understand your method now..... can you try and do your method but do a manual tune between sub and midbass pre Dirac , and do peq crossover tuning? Moving mic averages, and lmk..... I bet it’s good

you actually got me thinking, I was trying to figure out why you were doing it like that.... I get it now. Would be cool to know what you think like this
 

·
Registered
Joined
·
505 Posts
So speaker tuning (crossover tuning) absolutely should be done at the listen position.....

so, awhile ago I was talking to thomas drugeon aka pos (creator or rephase) on the rephase thread on Diy. He said do crossover tuning at the speaker..... but he also said , you can do it at the listening position but it’s only good got one spot...

im beginning to feel it’s the best spot (for us) in a car.

close mic measurements do reveal measured phase very nice... super easy to read and no FDW needed. For us tho.... at our seat and a FDW works much better.

i went on a listen and it’s way way way way better.... my tweeters don’t sound brash no more .

so we’re on to something, I think we can make a 2ch tune be pretty dam bad ass if we do the 7ch method manually and do a 2ch Dirac......

@Ryan,
I understand your method now..... can you try and do your method but do a manual tune between sub and midbass pre Dirac , and do peq crossover tuning? Moving mic averages, and lmk..... I bet it’s good

you actually got me thinking, I was trying to figure out why you were doing it like that.... I get it now. Would be cool to know what you think like this
Maybe I'm missing something Andy, but aren't you just talking about the "old" way, which is outlined in Mini's docs? The old way is to handle PEQ, XOs, and TA before Dirac, and then running 2ch Dirac.
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #176 ·
Maybe I'm missing something Andy, but aren't you just talking about the "old" way, which is outlined in Mini's docs? The old way is to handle PEQ, XOs, and TA before Dirac, and then running 2ch Dirac.
yeah , but a combo of both....

it might screw things up, I need to get another 8x12
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #177 ·
So my drive into work morning today ,
Tonality is great but imaging is wandering all over the place.....

so, to disappoint everyone I’m going to have to retract what I just said about separate left and right EQ working now it doesn’t... it created a de-correlation filter. And Direc corrected some of it but it did not fix all of it....

So I will be changing it and I will EQ the left side and the right side equally and just find the spot that has the flattest response for both without separate EQ.

last night in the garage it sounded just fine, coming in today it still sounds good but there’s small errors and it’s definitely EQ in pretune causing it.

boo :-(
 

·
Registered
Joined
·
5,121 Posts
I completed a from scratch tune this morning using the 7 channel method - protective crossover on tweeters only, 7 channel Dirac, then Auto EQ in REW to get overall curve and to get low shelf - sounds terrific to me. I’m extremely happy with this tune.


Sent from my iPhone using Tapatalk Pro
 

·
Registered
Joined
·
5,003 Posts
Discussion Starter · #179 ·
I don’t miss all those measurements, man that’s like over 90 sweeps or something..
I completed a from scratch tune this morning using the 7 channel method - protective crossover on tweeters only, 7 channel Dirac, then Auto EQ in REW to get overall curve and to get low shelf - sounds terrific to me. I’m extremely happy with this tune.


Sent from my iPhone using Tapatalk Pro
I don’t miss all those sweeps ..... like over 90 or something crazy....

Did you merge your sub into your midbass group ? It makes sense to do that , but also seems it would be the same thing on its own.. except tilt.

so you did 7 separate channels and 4 groups ?
 

·
Premium Member
2006 Hyundai Sonata V6
Joined
·
385 Posts
Tried this method, although since I have a cap before the tweeters (and mids, at least for now), I had all crossovers disabled for Dirac. Did not run any REW, as I didn't have time, and simply copied RAC's specs for a shelf filter. So still more to come, but so far, so good. With only the arbitrary shelf filter (applied to all outputs), and more crossover settings to try, I'm getting the best blend between sub and woofers I've had so far.
 
161 - 180 of 321 Posts
Top