[pika_]

Lol that made no sense at all …. Can you say again sorry

Let me try to rewrite that.

• Choose one Dirac channel to use for rear channels
• Route left and right channels to this Dirac channel. Ensure the level is -6 and one of them is out of phase, ie:

• On the Mixer, map that Dirac channel to the rear channels. I'm using two, but more could be added.

• Now I've changed the phase on Rear R because I want the output to be -R+L, but I've used -L+R on the routing for the Dirac 1.

I wouldn't even run Dirac on this Dirac channel, I would keep it flat to ensure the same signal is then sent to Rear L and Rear R.

 

[oabeieo]

Let me try to rewrite that.

• Choose one Dirac channel to use for rear channels
• Route left and right channels to this Dirac channel. Ensure the level is -6 and one of them is out of phase, ie:

View attachment 364496

• On the Mixer, map that Dirac channel to the rear channels. I'm using two, but more could be added.

View attachment 364497

• Now I've changed the phase on Rear R because I want the output to be -R+L, but I've used -L+R on the routing for the Dirac 1.

I wouldn't even run Dirac on this Dirac channel, I would keep it flat to ensure the same signal is then sent to Rear L and Rear R.

It won’t work

1. You can’t do L-R before runningDirac.

2 the L-R has to be done inn there Dirac tab not the mixer

Dirac changes phase, and the L-R is coherent phase dependent… so it has to be implemented on the signal before Dirac

You could run Dirac on both possibili and then do the L-R after Dirac is done and in the Dirac routing

Altho the timi mg would be weird

I would still only do the left rear , then for the right use the signal from front Doors to do the R-L then manual TA

Again , because the right will match the door response very close. Where left won’t

 

[pika_]

It won’t work

1. You can’t do L-R before runningDirac.

2 the L-R has to be done inn there Dirac tab not the mixer

Dirac changes phase, and the L-R is coherent phase dependent… so it has to be implemented on the signal before Dirac

You could run Dirac on both possibili and then do the L-R after Dirac is done and in the Dirac routing

Altho the timi mg would be weird

I would still only do the left rear , then for the right use the signal from front Doors to do the R-L then manual TA

Again , because the right will match the door response very close. Where left won’t

Yes, you're right; that's why I recommend not running Dirac measurement for this specific Dirac channel; that way can:

1. Add L-R before running Dirac (well, you won't run Dirac so problem solved )
2. L-R is done on the Dirac mixer
3. no phase corrections will be applied because we're bypassing Dirac at all

These plus:

View attachment 364497

• Now I've changed the phase on Rear R because I want the output to be -R+L, but I've used -L+R on the routing for the Dirac 1.

Would make it work

I'm not 100% sure about the benefits of having one Dirac channel per speaker;
This is my current FR, measured with a moving mic on the listening position. My target curve has a small dip around 200 and should start rising +- at 5k onwards.

I'll merge Woofers + Midranges, run Dirac, set crossovers, measure FR and see if it's worth it

 

[oabeieo]

Yes you're right; that's why I recommend not to run Dirac measurement for this specific dirac channel; that way can:

1. Add L-R before running Dirac (well, you won't run Dirac so problem solved )
2. L-R is done on the dirac mixer
3. no phase corrections will be applied because we're bypassing Dirac at all

These, plus:

Would make it work

The problem is you HAVE to give a channel to a Dirac route…. All 12ch must come from Dirac

Unless you can turn it off in Dirac channels configuration menu then maybe ??

 

[pika_]

The problem is you HAVE to give a channel to a Dirac route…. All 12ch must come from Dirac

Unless you can turn it off in Dirac channels configuration menu then maybe ??

Correct, but you don't need to run Dirac corrections. So it's L-R -> Dirac 1 -> mix to Rear L and Rear R

 

[pika_]

@oabeieo regarding setting the crossovers before/after running Dirac

This is a simulation of my crossover points per speaker, assuming a flat house target and 24 slope LR filters.

Tweeters HPF 2500hz LR24
Mids BPF 250 to 2500 LR24
Woofers BPF 80 to 250 LR24
Front Sub BPF 60 to 80 LR24
Back Sub LPF 60 LR24

Let's assume that I've calibrated Dirac having each speaker playing full range (when possible).
In theory, After running Dirac I should be able to set the previous crossover points and get a flat line right? LR24 phase sums up perfectly so...
Let's see what happens:

Ok, that's not what we were expecting right? ..but why?
Because the phase from the front sub is being affected by both LPF and HPF. To get it perfectly blending, I'll need to compensate for it with time, and ensure that I either have phase slopes aligned or maximum phase at the crossover point.

Ok this looks much better! So to have a perfect blend, I actually need to delay my back sub by around 6ms!
Let's add woofers now and see what happens. I'm going to create a copy of this sum and use it as a reference from this point onwards.

Hum.. not flat..but not bad I guess? Can it get better?

That looks a bit better. So i'll need to increase 1ms to both subs to get a good blending between the three groups of speakers.. now i'll add another copy of the cumulative FR and add the midrange to that:

Out of phase?!

You see where i'm going right? adding the crossovers is not as simple as we whish for..
To be honest, since Dirac is applying phase (and timing) corrections, the only way to actually set it right is with rew.. or using one dirac channel for Left, another for Right and another one for sub(s). This way Dirac will ensure that the FR is corrected for the crossover points as well. Sub channel will always have to be aligned manually.

 

[oabeieo]

@oabeieo regarding setting the crossovers before/after running Dirac

View attachment 364782

This is a simulation of my crossover points per speaker, assuming a flat house target and 24 slope LR filters.

Tweeters HPF 2500hz LR24
Mids BPF 250 to 2500 LR24
Woofers BPF 80 to 250 LR24
Front Sub BPF 60 to 80 LR24
Back Sub LPF 60 LR24

Let's assume that I've calibrated Dirac having each speaker playing full range (when possible).
In theory, After running Dirac I should be able to set the previous crossover points and get a flat line right? LR24 phase sums up perfectly so...
Let's see what happens:

View attachment 364785

Ok, that's not what we were expecting right? ..but why?
Because the phase from the front sub is being affected by both LPF and HPF. To get it perfectly blending, I'll need to compensate for it with time, and ensure that I either have phase slopes aligned or maximum phase at the crossover point.

View attachment 364786

Ok this looks much better! So to have a perfect blend, I actually need to delay my back sub by around 6ms!
Let's add woofers now and see what happens. I'm going to create a copy of this sum and use it as a reference from this point onwards.
View attachment 364787

Hum.. not flat..but not bad I guess? Can it get better?
View attachment 364788

That looks a bit better. So i'll need to increase 1ms to both subs to get a good blending between the three groups of speakers.. now i'll add another copy of the cumulative FR and add the midrange to that:

View attachment 364791
Out of phase?!
View attachment 364799


You see where i'm going right? adding the crossovers is not as simple as we whish for..
To be honest, since Dirac is applying phase (and timing) corrections, the only way to actually set it right is with rew.. or using one dirac channel for Left, another for Right and another one for sub(s). This way Dirac will ensure that the FR is corrected for the crossover points as well. Sub channel will always have to be aligned manually.

Yeah , that’s excellence, that’s exactly what I was talking about

You are on the right track , god job man, that’s lookin great !

 

[oabeieo]

And yes , the sub plus or minus 4-10ms it’s usually somewhere in there….

Sorry I said forward, sometimes it is forward , it all depends where Dirac puts the delays on things….

And if your bass amp can’t turn crossovers off , this needs to be done every time

Yeah that’s pretty much exactly what I was talking about …. The sub is spread out a lot, over time…… the filters change things and the acoustical space changes things (box, 1:4space loading etc )

So , so does it sound great now or what?

 

[pika_]

Sound it's pretty good! Bass is finally where it should be. On my previous tune with a different dsp I've set time alignment based on the impulse response, so it was damn close. I wasn't achieving the same bass response with Dirac until I actually set timing accordingly.
Now I need to find my preferred curve 🤪

 

[pika_]

And yes , the sub plus or minus 4-10ms it’s usually somewhere in there….

Sorry I said forward, sometimes it is forward , it all depends where Dirac puts the delays on things….

And if your bass amp can’t turn crossovers off , this needs to be done every time

Yeah that’s pretty much exactly what I was talking about …. The sub is spread out a lot, over time…… the filters change things and the acoustical space changes things (box, 1:4space loading etc )

So , so does it sound great now or what?

The graphics shown above are simulations. I took rew, got a flat frequency response, applied crossovers and started to see how the drivers interacted with each other.

So the example above applies to all cars with bandpass filters where LPF and HPF are close enough to change the expected phase response from it.

On top of that, we still need to add the car effect and Dirac corrections 🤣

 

[oabeieo]

The graphics shown above are simulations. I took rew, got a flat frequency response, applied crossovers and started to see how the drivers interacted with each other.

So the example above applies to all cars with bandpass filters where LPF and HPF are close enough to change the expected phase response from it.

On top of that, we still need to add the car effect and Dirac corrections 🤣

Yeah but your on the right track, your thinking is exactly what I was trying to convey

Keep in mind as long as you tell minidsp the sub is a sub in Dirac channels it will center the impulse diffrent and get you way closer… if not just perfect , but again , amp has to be able to have all filtering turned off…. And probably a pretty low inductance sub…

I’m so happy to see these results…. It’s awesome when your midbass and sub growl together…. It’s just the coolest ever

 

[pika_]

That's good input on how Dirac handles subwoofers! I guess the next step is to measure what is going on per speaker and align it as close as I can. I need to get an RCA cable to Minidsp, my head unit doesn't have an AUX input.

 

[Sonnie]

Clarification on what some of you are doing on the BM and APF part of the process.

I get setting the sub LPF and mid-bass HPF in BM and not setting those in outputs after running DL.

I read this:

"For All Pass filters:
Tweet has the BM APF and the tweeter crossover APF
Mid has the two crossover APF plus the BM APF
Midbass has the BM APF and the 1 APF for its LP
Sub has only the BM APF"

Example: intended crossover on 3-way with sub... 70/300/5000

On ALL outputs before DL is run, set the following in the PEQ section using the ALL-PASS setting:

70Hz APF representing the BM LPF for the sub and BM HPF for the mid-bass
300Hz APF representing the LPF for the mid-bass and HPF for the mids
5000Hz APF representing LPF for the mids and HPF for the highs/tweeters

Therefore, there will be 3 APFs on each of the 7 output channels.

On the 5000Hz mid to tweeter crossover... not sure what I might use yet... will experiment. Most of my systems have been 2-way with a mid-bass in the door and wide-band dash speaker (last one was HAT Legatia x3). This system will have the tweeters, however, the mids have excellent extension out to 27kHz. I was thinking about trying 10kHz as the LPF and let the tweeters be more like super tweeters above 10kHz. Thoughts?

BLAM Signature MS 3.55

 

[oabeieo]

Clarification on what some of you are doing on the BM and APF part of the process.

I get setting the sub LPF and mid-bass HPF in BM and not setting those in outputs after running DL.

I read this:

"For All Pass filters:
Tweet has the BM APF and the tweeter crossover APF
Mid has the two crossover APF plus the BM APF
Midbass has the BM APF and the 1 APF for its LP
Sub has only the BM APF"

Example: intended crossover on 3-way with sub... 70/300/5000

On ALL outputs before DL is run, set the following in the PEQ section using the ALL-PASS setting:

70Hz APF representing the BM LPF for the sub and BM HPF for the mid-bass
300Hz APF representing the LPF for the mid-bass and HPF for the mids
5000Hz APF representing LPF for the mids and HPF for the highs/tweeters

Therefore, there will be 3 APFs on each of the 7 output channels.

On the 5000Hz mid to tweeter crossover... not sure what I might use yet... will experiment. Most of my systems have been 2-way with a mid-bass in the door and wide-band dash speaker (last one was HAT Legatia x3). This system will have the tweeters, however, the mids have excellent extension out to 27kHz. I was thinking about trying 10kHz as the LPF and let the tweeters be more like super tweeters above 10kHz. Thoughts?

BLAM Signature MS 3.55

View attachment 367908

Tweet gets 2 APs
Mid gets 3 APs
Sub gets 1AP


It’s an exact simulation of every crossover each speaker will see in its signal path


I have had really good results keeping the phase in the midrange to midbass on some cars , and only doing the AP for the BM

It works nice. In that case

Tweet gets 1
Mid gets 1
Sub gets 1

All at same frequencie.

The phase wrap in the midbass to midrange sometimes can make stage sound farther away

 

[teh_squirrel]

@oabeieo do you use caps on your tweets for the dirac sweeps? How do you feel about it?

 

[Sonnie]

Tweet gets 2 APs
Mid gets 3 APs
Sub gets 1AP


It’s an exact simulation of every crossover each speaker will see in its signal path


I have had really good results keeping the phase in the midrange to midbass on some cars , and only doing the AP for the BM

It works nice. In that case

Tweet gets 1
Mid gets 1
Sub gets 1

All at same frequencie.

The phase wrap in the midbass to midrange sometimes can make stage sound farther away

You mention Tweet, Mid, and Sub... but left out mid-bass. Any reason?

So you are saying that in some cases you don't use APFs for any driver except the crossover set in BM?

 

[oabeieo]

You mention Tweet, Mid, and Sub... but left out mid-bass. Any reason?

So you are saying that in some cases you don't use APFs for any driver except the crossover set in BM?

That was a mistake

Midbass gets 2

 

[adamj.life]

Thanks for this guide it has made my first tune a bit of a breeze compared to the learning curve I had with Helix.
Though I am now trying to tune the rest of my setup, as the first tune was just with the standard 3way and sub setup.
I now want to add in my secondary subs and rear fill.
I have been following what pika has done to free up the extra Dirac Chanel’s by grouping tweets and mids, but then it got real technical and to be honest I lost what changes he was making and why. But from what he originally had, so in essence having.

Tweets and Mids - Dirac 1&2
Midbass - Dirac 3&4
Rearfill - Dirac 5&6
Front Subs (50-80hz) - Dirac 7
Main Sub (10-50hz) - Dirac 8

is this the best approach? Would grouping the tweets and mids affect staging?

Or should rearfill come off the midrange channels in the mixer.

Or am I totally off track and there is a better approach to achieve what I need to with this setup?

 

[oabeieo]

Thanks for this guide it has made my first tune a bit of a breeze compared to the learning curve I had with Helix.
Though I am now trying to tune the rest of my setup, as the first tune was just with the standard 3way and sub setup.
I now want to add in my secondary subs and rear fill.
I have been following what pika has done to free up the extra Dirac Chanel’s by grouping tweets and mids, but then it got real technical and to be honest I lost what changes he was making and why. But from what he originally had, so in essence having.

Tweets and Mids - Dirac 1&2
Midbass - Dirac 3&4
Rearfill - Dirac 5&6
Front Subs (50-80hz) - Dirac 7
Main Sub (10-50hz) - Dirac 8

is this the best approach? Would grouping the tweets and mids affect staging?

Or should rearfill come off the midrange channels in the mixer.

Or am I totally off track and there is a better approach to achieve what I need to with this setup?

That sounds perfect to me

I like mids and tweets grouped together

Some of the guys say sub and midbass is also good .

 

[adamj.life]

That sounds perfect to me I like mids and tweets grouped together Some of the guys say sub and midbass is also good .

How would I go with the crossovers here for my setup. The thee way setup in the guide says to add them to the route table for lpf and hpf and not to put a hpf on the midbass in the outputs either. How would that work with the two sets of subs? Little lost on how to scale the three way setup for what I have.