DiyMobileAudio.com Car Stereo Forum banner

141 - 160 of 222 Posts

·
Registered
Joined
·
708 Posts
I use BootCamp with a legit, but unregistered, copy of Win 10 Pro. I also have VMWare set to virtualize the BootCamp partition. Most things will connect to Windows running in VMware just fine, but on occasion I'll actually need to boot into Windows.

And by "most things" I don't just mean tuning stuff, I also have to plug in professional video processors and other random programs I might need in my line of work.

I also have a VM of older MacOSes and a few Linux distros to use in a pinch.
 

·
Registered
Joined
·
4,120 Posts
Maybe it does do some gating...

You get an impulse response and you see the delayed comb of the reflections.
Say you cut them off to minimispze echos.

Then you do an inverse FFT of that response, and you get the FIR taps for Dirac.. without the reflections... phase and amplitude.
(Or maybe there is some normalisation to add in the house curve and account for the spectrum of the (pink?) noise?)

There is probably more happening, but at a kindergarten level it is close.

Yeah kindergarten is right. I wish I was smart enough to figure all this out


But yeah you get it, from what I have tested in my car and what I think is happening based on my own experiences and with the use of linearized crossovers and without and with pre eq and without , listening and measurements just comparing as much as I can to try to know what the algo is made up of.

Using minimum phase eq before the calibration obviously will change the way it makes its correction. That part is a definite. I’m just suggesting it’s not limited to the frequency domain and it does in fact make a difference for better or worse.

It’s just in my own experience that changing things up using peq before the calibration changes the way the calibration sounds and it makes it sound even more processed and the calibration seems loose a sense of space in the recording.

Being that the algo is not having us take close mic measurements and is not having us take measurements with no crossovers leads me to believe that adding any eq before the calibration would just make it more difficult for it to know what is going on in the natural space and with how the speaker is actually behaving.

When I listen to the measurements being taken the are a few spots in the measurements and it’s random that I can hear the measurement being played
With a lot more energy compared to some of the energy from the other measurements. Seems like I can hear like two or three of the 27 sweeps come out with more energy being applied to the measurement. As if the algo has some control over the sweeps energy content. Meaning not all the sweeps come out exactly the same.

I think it’s doing something with the power response in Some of the measurement points and it’s random or calculated based on the 1st measurement (IDK) but I definitely hear it sometimes.

It seems like it does look at a lot of different things. One thing I’ve been able to determine is I look at my own moving mic averages with plain old RTA in rew with pink or PN and the magnitude sometimes isn’t what Dirac’s magnitude shape resembles. I’ve tryed different measurement points (like moving the mic a few inches from the box maybe in or out) and have gotten the Dirac average magnitude or “avg spectrum” to better resemble what my RTA averages looked like, when I do that the EQ work that Dirac does turns out much more faithful to my target when I validate my target response with RTA
 

·
Registered
Joined
·
4,587 Posts
I took my car to a home audio dealer friend that I highly respect for his ear. I did this the day before leaving for Finals. I had a heavy PEQ tune and a tune with no PEQ and a simple Dirac curve.

He liked the staging / imaging work Dirac did when he toggled Dirac on/off using the remote, but he was not happy with the overall tonality of the system - thin, artificial, processed.

He said he would love to hear it without all of the PEQ, and I informed him I had a preset for that. We hit the other preset, and he said it was an improvement with no PEQ and DL on. He actually liked the tonality the best with only crossovers, TA, and levels applied - though it had some definite problem areas.

He stated from his home experience with customers running these types of algorithms - that the software is intended to improve staging / imaging without drastically changing the tonality. He said if I could strike that balance, I’d have a very good reference car.

On the drive home, I could hear what he was saying - particularly on a Dave Matthews track - my heavy PEQ preset - the guitar sound was only the strings, the no PEQ/DL - one could hear the strings AND the body of the guitar.

So I have been trying various things, but I’m not ready to take it back over He said he would love to spend a few hours with it - which would be phenomenal.


Sent from my iPhone using Tapatalk Pro
 

·
Registered
Joined
·
4,120 Posts
I took my car to a home audio dealer friend that I highly respect for his ear. I did this the day before leaving for Finals. I had a heavy PEQ tune and a tune with no PEQ and a simple Dirac curve.

He liked the staging / imaging work Dirac did when he toggled Dirac on/off using the remote, but he was not happy with the overall tonality of the system - thin, artificial, processed.

He said he would love to hear it without all of the PEQ, and I informed him I had a preset for that. We hit the other preset, and he said it was an improvement with no PEQ and DL on. He actually liked the tonality the best with only crossovers, TA, and levels applied - though it had some definite problem areas.

He stated from his home experience with customers running these types of algorithms - that the software is intended to improve staging / imaging without drastically changing the tonality. He said if I could strike that balance, I’d have a very good reference car.

On the drive home, I could hear what he was saying - particularly on a Dave Matthews track - my heavy PEQ preset - the guitar sound was only the strings, the no PEQ/DL - one could hear the strings AND the body of the guitar.

So I have been trying various things, but I’m not ready to take it back over He said he would love to spend a few hours with it - which would be phenomenal.


Sent from my iPhone using Tapatalk Pro

Interesting, yeah it’s too bad the car adds so much to the sound that makes it so hard to get the tonality like a home system. Our speakers sound so much better than they do out of the car with no eq than in the car with even DL.


So a post Dirac no pre eq tonality, what is it that seems wrong.

Has anyone tried diffrent measurement points so the average matches a moving mic pink noise average? That definitely helps


Or the other way to do it is do your calibration, than go to rew and take a moving mic 32 average (maybe 1/6th or 1/12th) and look at the shape of magnitude compared to your target ,

Whatever the deviations are use the Dirac target and compensate for the difference using your Dirac target.

Both ways work generally the same to my ears.

The 1st way in my car the left and right forward measurements are more directly in front of me (about 6” apart) instead of having them really far apart)
So the two forward high measurements are almost next to each other
That seems to get my Dirac average to match rew pink noise averages enough for me to be happy. It’s not like exact but the dips and peaks are at least in the same spots. I think that’s what it’s really about
I’ve had some Dirac avearges have completely different peaks and dips than using RTA with noise......


Getting the response in DL to validate against REW moving mic has made my tonality much better.

Have you tried ?
 

·
Premium Member
Joined
·
265 Posts
I sent the MiniDSP support team an email about running Dirac Live on Mac, and they said it works perfectly fine. They actually said they run it on mac, along with about half their customer base. So that's good news. :)
 

·
Premium Member
Joined
·
265 Posts
More on the Mac issue:

. . . we’re aware of a Dirac live bug with older Mac running newer OS (Catalina). You simply just run Dirac 2.2.2 (older version) and it solves the issue. Maybe that was the case here?
 

·
Registered
Joined
·
556 Posts
I sent the MiniDSP support team an email about running Dirac Live on Mac, and they said it works perfectly fine. They actually said they run it on mac, along with about half their customer base. So that's good news. :)
Great news. Thanks for the update.
 

·
Premium Member
Joined
·
3,011 Posts
...
Using minimum phase eq before the calibration obviously will change the way it makes its correction. That part is a definite. I’m just suggesting it’s not limited to the frequency domain and it does in fact make a difference for better or worse.
...
The Dirac delta function describes the time domain response.
And most of the suggest that Dirac Live is operating in the time domain.

It sort of makes sense then that they maybe using some gating to only get the speaker corrected with Dirac, and use gating to ignore the room reflections.

Of course those room reflections may bleed in and affect the frequency response.

And the amplitude (REW) should be able to be simultaneously corrected with the FIR that is doing Dirac.


Yeah kindergarten is right. I wish I was smart enough to figure all this out.
...
The fact that some people do 4+ years of college study and can find it challenging, sort of suggests that it is potentially complicated...
 

·
Premium Member
Joined
·
265 Posts
Shipped from China on Monday. Received in California on Tuesday. That's pretty intense.
 

·
Registered
Joined
·
4,587 Posts
My thought exactly - it didn’t seem like I paid for 2nd day air in the amount they charged for shipping, but I was very pleasantly surprised to get it so quickly. Bonus!


Sent from my iPhone using Tapatalk Pro
 

·
Registered
Joined
·
4,120 Posts
The Dirac delta function describes the time domain response.
And most of the suggest that Dirac Live is operating in the time domain.

It sort of makes sense then that they maybe using some gating to only get the speaker corrected with Dirac, and use gating to ignore the room reflections.

Of course those room reflections may bleed in and affect the frequency response.

And the amplitude (REW) should be able to be simultaneously corrected with the FIR that is doing Dirac.




The fact that some people do 4+ years of college study and can find it challenging, sort of suggests that it is potentially complicated...

Dirac transfer function is time and frequency.
It’s just a transfer function, there’s other mathematical ways of generating an impulse.

But yeah it would have to gate the responce to know the difference between a reflection and direct sound to calculate a correction curve , as it’s not having us do any close mic measurements. It is the only way to take time of flight out of a measurement that I know of. Unless there’s something I’m missing.


”And the amplitude (REW) should be able to be simultaneously corrected with the FIR that is doing Dirac.



You lost me there , not sure what you meant sorry
 

·
Premium Member
Joined
·
3,011 Posts
Dirac transfer function is time and frequency.
It’s just a transfer function, there’s other mathematical ways of generating an impulse.

But yeah it would have to gate the responce to know the difference between a reflection and direct sound to calculate a correction curve , as it’s not having us do any close mic measurements. It is the only way to take time of flight out of a measurement that I know of. Unless there’s something I’m missing.


”And the amplitude (REW) should be able to be simultaneously corrected with the FIR that is doing Dirac.



You lost me there , not sure what you meant sorry
Basically... The FIR filter corrects both REW and phase.

I don't think that the gating would work... the reflections are just too close in time.
 

·
Registered
Joined
·
22 Posts
Guys, they ship from HK, right?

Can you recommend c-dsp 8x12 with or without Dirac over helix dsp3 ?

Especially in the tuning aspect.
Thank you.
 

·
Registered
Joined
·
1,322 Posts
Guys, they ship from HK, right?

Can you recommend c-dsp 8x12 with or without Dirac over helix dsp3 ?

Especially in the tuning aspect.
Thank you.
I’d say the Helix Pro DSP and C-DSP 8x12 are similar though it seems most give a slight edge to the Helix DSP on software though I’ve been using the MiniDSP software for 4+ years now and think it is very good with solid support. The Helix has the advantage of add-in modules like the Bluetooth module. But really you can’t make a bad choice between either of them.

Where the C-DSP 8x12 separates itself from the Helix is the option of Dirac, which is a very high quality tuning “easy” (once you learn it) button. Essentially the results you get from Dirac are better than most amateur and even pro tuners in the business. Now there are definitely some true expert tuners that could provide better results than Dirac but that would take hours of tuning and expert tuners are few and far between.
 

·
Registered
Joined
·
22 Posts
This is what I wanted to hear.
I'm 99% sure that I won't be able to tune the 3 way properly. No matter how many hours I will spend on it.

And with constant gear testing and DIY install - it's too expensive to hire pro tuners for this job continuously.

If Dirac can help with that and minidsp/helix sound almost identical - then it's worth it. You can "hire" Dirac many times and then even re-sell the licence.


Where the C-DSP 8x12 separates itself from the Helix is the option of Dirac, which is a very high quality tuning “easy” (once you learn it) button. Essentially the results you get from Dirac are better than most amateur and even pro tuners in the business. Now there are definitely some true expert tuners that could provide better results than Dirac but that would take hours of tuning and expert tuners are few and far between.
 
141 - 160 of 222 Posts
Top