Jump to content
IGNORED

Dirac and time alignment of the speakers.


Recommended Posts

Hey!

 

Been reading a lot on various DRC solutions and Dirac Live seems to be the most promising at this point. Especially for a guy like me, who does not have enough knowledge in the whole theory and just want a "system" to take care of the sound.

 

Now, i am considering buying DDRC-22A with Dirac Live from miniDSP and i am going to install it in the car. In my car i have a 3-way front plus sub setup. The sound is extracted from the head unit via MOST->SPDIF converter (moBridge) and then goes to Audison BitOne. Here the signal is crossed over and fed into 4 different amplifiers. I have mid-bass drivers installed in doors mid-range and tweeters are in the bottom of a-pillars, the sub is in the trunk.

 

The idea is to install the DDRC-22A in between moBridge and BitOne and run the correction.

 

 

Now, the question is - will Dirac take care of the phase shifts that occur due to the distance difference between all the speakers? Although BitOne does provide an ability to do time-alignment of speakers, i couldn't get it perfect. So, what if i just turn that feature off in BitOne and hope that Dirac takes care of different timing?

 

I tried Dirac on my home computer in the room with no treatment and i really love the results. The booming is gone, sound became precise and transparent. Transients are so much better. I really hope i can have this kind of a result in the car.

Link to comment

Mmm... this is interesting :)

 

You probably already know that Dirac Live is used on Bentleys, Rolls-Royce, some top BMW models and now Volvo, but there is a lot of work and tuning involved as cars are very different from the two channel room experience as they almost always have an upmix to more speakers than input channels, something that is never done in hifi.

 

On top of this the listening positions in relation to the speakers are a mess compared to rooms, so gain and delay values need to take car occupancy into account and it is not unusual to have almost no direct wave at all as for example we listen to the first reflection in the wind shield...

 

I have no experience using the consumer version of Dirac Live in cars but I heard that the stereo version has been used successfully.

I suggest some experiments to see what mic positions give the results you like the best so I would start with the chair style measurement but you may also force Dirac Live to use one single measurement assuming that you will compensate for the driver seat only.

 

Ciao, Flavio

Warning: My posts may be biased even if in good faith, I work for Dirac Research :-)

Link to comment

Hey Flavio,

 

Thanks for taking time to answer my questions.

 

Yeah, i've heard that Dirac is used in some of the luxury cars. So that is why i thought that it could be a good choice for my application.

 

I think i can consider both stereo and Live version. In order to use a multichannel version i will have to install a Carptuter with Dirac Live installed on it and some audio interface to output those channels. But this is really cumbersome i think.

 

I would like to try stereo version. To do that, i will need DDRC from miniDSP which i can install before crossovers and see how it does.

 

Here is a simple diagram of how my system looks:

 

Screen Shot 2014-05-23 at 12.50.29 AM.png

 

My question is, though:

 

Given that i use a stereo version of Dirac in a 3-way system plus a sub, how will Dirac compensate for the fact, that each speaker of this 3-way system is located at a different distance from a listener? So, basically, sound from no speaker comes at the same time with the other speakers. Does Dirac Stereo somehow compensate for that? I know, that in a 2-channel set-up it can set a delay and|or gain for one of the channels, but we have a pseudo 7 channel setup here running through a stereo. I can use a DSP inside the Audison BitOne which currently does the x-over - i can set up delays for each channels. I just thought that since delay and phase are closely related and Dirac somehow fixes phase anomalies (or does it?), i can switch off the time alignment function in the BitOne and allow Dirac do its magic.

 

Another option is to time-align each speaker with the help of BitOne (using software like REW or Holmimpulse) and let the Dirac process the rest.

 

Which is better?

Link to comment

Dirac Live will compensate for the fact that each speaker is located at different distances from the listener, I do not think you need other corrections...

so if you are thinking about a correction for the driver's position only you may even try the radical approach of correcting for a single point in space by using one mic position only.

 

Things get more complicated if you want to correct taking car occupancy into account using multiple measurements, that's the reason why I said that I have no hands on experience and you will have to experiment to see what mic positions give the results you like the best.

 

I'm sorry I cannot say more than that :-(

Flavio

Warning: My posts may be biased even if in good faith, I work for Dirac Research :-)

Link to comment

Flavio,

 

Well, i do not mind having one sweet spot - in a driver's position :)

 

As far as i understood, Dirac uses the first measurement as the main one and then adds the rest to smooth the result. Will taking a few measurements near the drivers position improve the result? I am talking about a box of about 50-60cm radius.

 

And we are discussing a stereo version of Dirac, correct?

Link to comment

I've never tried this, so here are some ideas that could be good or bad:

 

1) Use a minidsp or other device to do a matrix mix and time align every amplified channel.

2) Put Dirac stereo minidsp before it, so that the correction can be applied while it's only 2 channels.

 

alternately:

 

1) Use a multichannel minidsp to matrix mix and time align everything.

2) Use audiolense xo to generate fir filters for the minidsp for room correction, phase alignment, crossovers, or whatever you want to tweak.

 

I've never tried Audiolense XO, but it seems like it's designed for something like this.

Juice HiFi

Link to comment

My guess is:

 

Dirac can only "fix" time alignment when it has complete control of a particular driver, e.g. in the middle of a driver's passband. Then it can advance or delay the signal. In the crossover frequencies, it can not because it is the xo circuits that determine that.

 

However, it can still use it's phase and amplitude fixing powers to mostly patch it up.

 

I think the general advice is to try to do the time alignment first outside of Dirac.

Link to comment
My guess is:

 

Dirac can only "fix" time alignment when it has complete control of a particular driver, e.g. in the middle of a driver's passband. Then it can advance or delay the signal. In the crossover frequencies, it can not because it is the xo circuits that determine that.

 

However, it can still use it's phase and amplitude fixing powers to mostly patch it up.

 

I think the general advice is to try to do the time alignment first outside of Dirac.

Why wouldn't Dirac work for crossover frequencies?

Link to comment

Hello StabMe,

 

Dirac Live will work with your stereo configuration with downstream crossovering to several drivers in different locations and it will correct the delays up to a maximum of 10 milliseconds wich is more than required in a car :)

 

Ciao, Flavio

Warning: My posts may be biased even if in good faith, I work for Dirac Research :-)

Link to comment
Huh?

 

Room correction only corrects what comes after it, not what comes before it.

 

Here is an example of what Dirac can't do:

 

Suppose you have a 500 Hz crossover between woofer and midrange. Somewhere within that crossover range, let's say 460 Hz, the woofer and midrange are 180 out of phase, (time alignment is wrong) so they cancel and you get a null. This behavior is totally set by the passive crossover components of the speaker. Dirac can send a delayed 460Hz signal, or it can send a 460Hz signal of boosted amplitude, but it still can't change the fact that the woofer and midrange are still 180 out of phase and you are getting a null.

 

This is just an example with made-up numbers.

 

Dirac would need active control of each of the crossovers to be able to "fix" this problem.

Link to comment
Here is an example of what Dirac can't do:

 

Suppose you have a 500 Hz crossover between woofer and midrange. Somewhere within that crossover range, let's say 460 Hz, the woofer and midrange are 180 out of phase, (time alignment is wrong) so they cancel and you get a null. This behavior is totally set by the passive crossover components of the speaker. Dirac can send a delayed 460Hz signal, or it can send a 460Hz signal of boosted amplitude, but it still can't change the fact that the woofer and midrange are still 180 out of phase and you are getting a null.

 

This is just an example with made-up numbers.

 

Dirac would need active control of each of the crossovers to be able to "fix" this problem.

True, it can't do anything with a null, and there are other things it can't do, like nonlinear distortion or temperature sensitive changes to drivers or crossovers.

 

In a real-world crossover, though, if it's 180 degrees out of phase, one side will have the polarity inverted to avoid a null. It's hard to imagine a crossover that is designed to create a null at the crossover frequency. In general, I think Dirac's phase correction should be beneficial for real crossovers.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...