DiyMobileAudio.com Car Stereo Forum banner
61 - 80 of 236 Posts

·
Registered
Joined
·
33 Posts
I'm also rather confused by all this. I realize it's a big ask, but would anybody who has given this technique a try be willing to take some screenshots to illustrate exactly what you're doing?
I was thinking about doing that. I am going to conduct another run maybe Sunday evening since I didn’t have a totally quiet environment during my last couple of measurements thanks to the cooling fan for the instrument cluster display kicking on (smh).

If by chance you are confused about what the OP meant by ungrouping target curves in the DL software like I was, the manual explains exactly how to do this. I searched ‘group’ in the pdf file and I think it is the 2nd entry that takes you to the right spot. Once I got through that step, I was able to follow the rest of OP’s instructions.
 

·
Premium Member
2006 Hyundai Sonata V6
Joined
·
317 Posts
I was thinking about doing that. I am going to conduct another run maybe Sunday evening since I didn’t have a totally quiet environment during my last couple of measurements thanks to the cooling fan for the instrument cluster display kicking on (smh).

If by chance you are confused about what the OP meant by ungrouping target curves in the DL software like I was, the manual explains exactly how to do this. I searched ‘group’ in the pdf file and I think it is the 2nd entry that takes you to the right spot. Once I got through that step, I was able to follow the rest of OP’s instructions.
Thanks for the offer and the pointer on finding it in the manual!
 
  • Like
Reactions: dmparker5725

·
Registered
Joined
·
462 Posts
I tried OP's method just now. It's definitely very quick and efficient.

My results are mixed. The crossovers all seem clean, though I haven't measured yet.

Flat is a bit bright for me; tried applying a -5dB high shelf across the board (Q=1, F=4000), which helped. Not sure that I like this better than the usual 2ch tune, but I haven't played with it a whole lot.

Note: I have a capacitor inline with my tweeters. I would have been too nervous to do this approach without a cap. I think OP had done his approach on TBMs, which supposedly are very hard to blow.
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
1,820 Posts
Spent some time tonight experimenting with this new method. But I did a couple things different.

First off I don't have caps on the tweets so instead of risking blown tweeters I set a 800hz BW2 HP & filpped polarity before Dirac. And I kept this HPF applied after so as not to mess with how Dirac corrects for it. To apply my typical 3500hz LR4 HPF after - I used the advanced xover function to add biquads.

Second, I used the Bass Management function to apply the Subs LPF and Midbass HPF. These filters aren't seen by Dirac as they are before where the test signal is injected into the signal path within the DSP. This also allow the midbass HPF to cascade to the mids & tweets.

Third, again using the advanced xover function I added biquads I added a LPF to the Sub output to match the midbass LPF which virtually cascades that filter down to the sub. [EDIT: Actually didn't have to use the advanced function here because the standard sub LPF was set in the Input/BM screen so the sub output tab xover was free to apply this filter.] Why do this "cascading"? This explains the theory: Linkwitzlab - 4 Way topology issues & cascading

Forth, I kept the Sub & both Midbass in the same group. My thinking is assigning channels to a single group instructs Dirac to match phase so why not keep those three in the same group.

The only setting in the output tabs I changed before Dirac was reducing the level of the sub by 4.5db to prevent clipping while measuring in Dirac. All other levels, eq, delay were zero'd out.

I also left the "Default" Dirac curve in for all groups instead of loading a flat curve.

Here are some screen shots:

Sub/Midbass:
Font Slope Screenshot Software Technology



Mids:
Rectangle Font Line Screenshot Slope


Tweets:
Slope Font Line Screenshot Software


My first impression is YES the xover regions do sound more cohesive. Some quick measuring shows almost textbook summing between the sub & midbass. And the acoustic xover (73hz) was very close to the electronic (70hz).... it was for all the xovers points actually.

Rectangle Slope Plot Font Line



And look at the match on the L/R midbass above. I've never been able to get measurements that good with other methods. Notice the typical ~80hz null is gone from the L midbass too. It doesn't measure like that when doing a 2ch Dirac tune - The null is still there and I believe that's what causes a phase misalignment with the sub.

I did add a 100hz, 15db, 0.8Q low shelf to all output eq's in order with get some more low end which worked great but was a bit too much for my tastes. Tonality still needs some massaging but this newly discovered method to implement Dirac definitely has some major potential. I attempted to use AutoEQ in REW to shape to my desired final curve but I couldn't get it to boost the areas that needed it.... need to read up on why.

Here is before & after adding the shelf filter:
Rectangle Slope Plot Font Line
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
462 Posts
Thanks Ryan. I'm going to give that a go today. Seems to me like you custom designed that curve for me (15db bass boost). :)

Do you think using Dirac default curve versus flat matters? I used flat yesterday. My main takeaways from your post were to group MBs/Sub together, use bass mgmt (plus an additional LPF on the subs to match the MBs), and use biquads instead of gain to get a proper tilt on the bass region.

To get my downward tilt, I had done something like a high shelf @ 370Hz, -2dB, q=.5, and 3000Hz, -3dB, q=.7
 
  • Like
Reactions: tittysprinkles888

·
Registered
Joined
·
1,820 Posts
Do you think using Dirac default curve versus flat matters? I used flat yesterday.
I do think so only because, in my observations over some time now, Dirac will always adjust the relative levels between Dirac channels to the "Default" curve. This is the curve that always shows up upon opening the filter design screen. It has a slight downward slope.. like only 4-5db from 20-20k.


My main takeaways from your post were to group MBs/Sub together, use bass mgmt (plus an additional LPF on the subs to match the MBs), and use biquads instead of gain to get a proper tilt on the bass region.
All that is right except the last part. In this attempt I only used biquads to add xover filters where more than 1 was of the same type needed on the same channel (the tweeters in my case). But yeah - definately don't use gain/level to make the tilt because that will throw off the xover region phase coherence.
Though - eq biquads can be generated with REW AutoEQ to shape the final response which I attempted but couldn't get it to boost where needed under 100hz & above 3khz as shown here:

Rectangle Slope Font Plot Parallel
 

·
Registered
Joined
·
462 Posts
Makes sense to me; thanks for the edit on your post. So I ran Truthunter's process, and it sounds fantastic. I did not try to run Auto EQ afterward; instead I applied a high shelf (2500Hz, -4dB, Q=.7) to get closer to my target. I measured afterward, with seat leaned back and waving mic (red is everything, blue is sub only, green is front stage without sub):
Rectangle Slope Plot Font Line


Couple questions:

1) On the Inputs/Bass Mgmt tab, I have the LPF outputs at 0dB each; am I supposed to attenuate those at all to account for 2 channels being summed? Or does dirac take care of that already? I'm asking partially because it's a tad bass heavy for me, and trying to figure out the most appropriate way to attenuate. I suppose I could adjust the low shelf a bit.
2) Are you putting any gap between the electrical crossovers in Bass Mgmt? Currently I have both the HPF and LPF at 90

Thanks!
 
  • Like
Reactions: Truthunter

·
Registered
Joined
·
3,638 Posts
On my own recent tune. Inside a 2 channel. However, I was struggling with the idea of just having parametric EQ shape my entire roll off.

For instance, I have horns and I was like okay. I can electronically set them to have a 2000hz hpf with a 6db shallow electronic slope and no lpf. That way my actually wanted acoustic hpf of 1200hz could be shaped via parametric and jazzis spreadsheet to a perfect 1200hz, 24 db crossover slope.

That way I had protection below driver excursion frequencies and I had an exact textbook 24db acoustic slope on my horns.

Rinse and repeat on all drivers to give them a perfect acoustic roll off.

Run Dirac sweeps. Everything sums EXACTLY.

Sent from my LM-Q730 using Tapatalk
 

·
Registered
Joined
·
1,820 Posts
Makes sense to me; thanks for the edit on your post. So I ran Truthunter's process, and it sounds fantastic. I did not try to run Auto EQ afterward; instead I applied a high shelf (2500Hz, -4dB, Q=.7) to get closer to my target. I measured afterward, with seat leaned back and waving mic (red is everything, blue is sub only, green is front stage without sub):

Couple questions:

1) On the Inputs/Bass Mgmt tab, I have the LPF outputs at 0dB each; am I supposed to attenuate those at all to account for 2 channels being summed? Or does dirac take care of that already? I'm asking partially because it's a tad bass heavy for me, and trying to figure out the most appropriate way to attenuate. I suppose I could adjust the low shelf a bit.
2) Are you putting any gap between the electrical crossovers in Bass Mgmt? Currently I have both the HPF and LPF at 90

Thanks!
I'm glad you had success! Thanks for posting your experience here. The more people that try it and report back... the more we can confirm this is a good/bad idea 🙂

1) That is an issue when using a single Dirac channel for sub(s) and having two outputs feeding a sub amp for a single sub. In your case: On the mixer tab, set the two sub channels to -3db.

2) No gap needed in the xover because the raw response is being corrected to the default curve on all channels. That makes for the acoustic slopes lining up very close to or precisely where the electronic filters are set. Just don't adjust relative driver levels after and the acoustic xover won't change. That is also why it's important to apply any post Dirac EQ filters to all channels so the phase changes they create applies to all drivers and relative phase between them remain intact.

Hope that all makes sense.
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
462 Posts
I'm glad you had success! Thanks for posting your experience here. The more people that try it and report back... the more we can confirm this is a good/bad idea 🙂

1) That is an issue when using a single Dirac channel for sub(s) and having two outputs feeding a sub amp for a single sub. In your case: On the mixer tab, set the two sub channels to -3db.

2) No gap needed in the xover because the raw response is being corrected to the default curve on all channels. That makes for the acoustic slopes lining up very close to or precisely where the electronic filters are set. Just don't adjust relative driver levels after and the acoustic xover won't change. That is also why it's important to apply any post Dirac EQ filters to all channels so the phase changes they create applies to all drivers and relative phase between them remain intact.

Hope that all makes sense.
Yep, that all makes sense, thanks! I had each sub output set to -6dB on the mixer tab, prior to running Dirac. Now, I don't think I can change that without running Dirac again. In theory, I think, it shouldn't even matter how I set it on the Mixer tab since that would have been accounted for in any Dirac measurements, and therefore Dirac would have adjusted gain accordingly. That's kind of why I was asking specifically about the Inputs / Bass Mgmt section, since that would all be outside of Dirac's measurements.

Hope that made sense :)
 

·
Premium Member
2006 Hyundai Sonata V6
Joined
·
317 Posts
I've never used bass management, but I'll give it a go. So, set crossover there without a gap, then lower both sub channels -3dB in mixer. I assume if I experiment with BM crossover point, I'll need to run a new complete Dirac run each time I make a change? Installing mid & tweeter protection caps today, before making any attempts at this new way of Dirac.
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
1,820 Posts
I've never used bass management, but I'll give it a go. So, set crossover there without a gap, then lower both sub channels -3dB in mixer. I assume if I experiment with BM crossover point, I'll need to run a new complete Dirac run each time I make a change? Installing mid & tweeter protection caps today, before making any attempts at this new way of Dirac.
The LP/HP filter in BM can be changed after Dirac with out issue as long as both are changed to the same frequency... they should still sum correctly... again as long as relative levels between drivers is untouched.
 

·
Registered
Joined
·
4,980 Posts
Discussion Starter · #73 ·
I had a lot of reading to catch up on

well this is really fantastic!!!

i’m loving all the feedback.... good and bad. I wish I had more time to mess with it myself I was up against the clock on that install

Yes I did use base management between the sub and highs also , we wanted that filter to be inherent.

keep it up it sounds like we’re are on to something really good. I remember on his truck I was able to cross the 3” at 200 Hz we’re before 200 was getting way too much power, and it was interacting with the door very very good..

I couldn’t believe how much better the crossovers were interacting

This is really good, I remember someone on the mini DSP forum telling me something about octave spacing using shelving filters with a q.5

so that’s not an PEQ, as someone mentioned.... it’s a shelving filter....it’s sorta the same except it’s not.

but it makes sense that it would have to be a global change and it can’t be a change that doesn’t affect all other channels equally otherwise it will simply mess up the crossover alignment...
So that makes using levels to make your target sheet not an option.... idk , we’re gonna have to try I really like how truth Hunter is taking a good approach.
Although I can’t see your screenshots
 

·
Registered
Joined
·
4,980 Posts
Discussion Starter · #74 ·
I've never used bass management, but I'll give it a go. So, set crossover there without a gap, then lower both sub channels -3dB in mixer. I assume if I experiment with BM crossover point, I'll need to run a new complete Dirac run each time I make a change? Installing mid & tweeter protection caps today, before making any attempts at this new way of Dirac.
you definitely want BM! Absolutely!
Keep your HP crossover on your mid base turned off... use only the crossover in BM

that will make an inherent shift across all of the drivers so they’ll pick up and inherit that phase ****, they also have a hi pass that is way out of band, but it will allow you to cross your mid range lower because it will have two filters, end it will have the exact same phase fingerprint if you will on all channels
 

·
Registered
Joined
·
4,980 Posts
Discussion Starter · #75 · (Edited)
I've never used bass management, but I'll give it a go. So, set crossover there without a gap, then lower both sub channels -3dB in mixer. I assume if I experiment with BM crossover point, I'll need to run a new complete Dirac run each time I make a change? Installing mid & tweeter protection caps today, before making any attempts at this new way of Dirac.
you definitely want BM! Absolutely!
Keep your HP crossover on your mid base turned off... use only the crossover in BM


I'm glad you had success! Thanks for posting your experience here. The more people that try it and report back... the more we can confirm this is a good/bad idea 🙂

1) That is an issue when using a single Dirac channel for sub(s) and having two outputs feeding a sub amp for a single sub. In your case: On the mixer tab, set the two sub channels to -3db.

2) No gap needed in the xover because the raw response is being corrected to the default curve on all channels. That makes for the acoustic slopes lining up very close to or precisely where the electronic filters are set. Just don't adjust relative driver levels after and the acoustic xover won't change. That is also why it's important to apply any post Dirac EQ filters to all channels so the phase changes they create applies to all drivers and relative phase between them remain intact.

Hope that all makes sense.
you got it ! That’s perfect...

this is so awesome , I think I want an 8x12 again now

I think it’s just one of many ways to use the 8x12 and another preset to make out of the many different styles of tuning

although for me , it’s wouldn’t be a choice, it would be the only way for a 7ch Dirac tune, if I was using an 8x12 ......

**** I want one again ...... I miss that dsp
 

·
Registered
Joined
·
462 Posts
you definitely want BM! Absolutely!
Keep your HP crossover on your mid base turned off... use only the crossover in BM

that will make an inherent shift across all of the drivers so they’ll pick up and inherit that phase ****, they also have a hi pass that is way out of band, but it will allow you to cross your mid range lower because it will have two filters, end it will have the exact same phase fingerprint if you will on all channels
With this in mind, would we also want to apply the midrange's HPF to the tweeter? I'm guessing that tweeter is far enough out of band that it wouldn't make a meaningful difference, but since the last several posts are basically all about keeping phase correct on all drivers, I thought it would be worth asking.
 

·
Registered
Joined
·
2,896 Posts
1) On the Inputs/Bass Mgmt tab, I have the LPF outputs at 0dB each; am I supposed to attenuate those at all to account for 2 channels being summed? Or does dirac take care of that already? I'm asking partially because it's a tad bass heavy for me, and trying to figure out the most appropriate way to attenuate. I suppose I could adjust the low shelf a bit.
Interestingly when I tried using the Input/Bass Mgmt tab with it enabled I had no low end at all. I ended up disabling the bass management and then adding a low shelf filter to get any kind of output. I might re-enable it and with the low shelf in place see how that sounds, still trying to feel my way around this method.

I need to measure again though, I kept the crossover on my tweeters (no caps, did not want to blow them) and ended up with a bump right at that crossover where if I turn the volume up it starts to sound really harsh. Lower levels it is fine, but with the volume up.... ooofff.

Edit: Looked at it again and realized my error with regards to Bass Management... on the routing tab, I left my subwoofer on inputs 7/8 (TOSLINK) instead of moving it to the Bass Management line at the bottom. I updated that, enabled Bass Management again and have a really solid low end. It's sounding pretty great already, wish I had time today to run a new set of sweeps with a safe crossover point, but my regular crossover disabled, on the tweeters.
 

·
Registered
Joined
·
462 Posts
Interestingly when I tried using the Input/Bass Mgmt tab with it enabled I had no low end at all. I ended up disabling the bass management and then adding a low shelf filter to get any kind of output. I might re-enable it and with the low shelf in place see how that sounds, still trying to feel my way around this method.

I need to measure again though, I kept the crossover on my tweeters (no caps, did not want to blow them) and ended up with a bump right at that crossover where if I turn the volume up it starts to sound really harsh. Lower levels it is fine, but with the volume up.... ooofff.

Edit: Looked at it again and realized my error with regards to Bass Management... on the routing tab, I left my subwoofer on inputs 7/8 (TOSLINK) instead of moving it to the Bass Management line at the bottom. I updated that, enabled Bass Management again and have a really solid low end. It's sounding pretty great already, wish I had time today to run a new set of sweeps with a safe crossover point, but my regular crossover disabled, on the tweeters.
I've made that same mistake before. Glad your bass is solid now. There's nothing quite like that feeling :)
 

·
Registered
Joined
·
4,980 Posts
Discussion Starter · #79 ·
Interestingly when I tried using the Input/Bass Mgmt tab with it enabled I had no low end at all. I ended up disabling the bass management and then adding a low shelf filter to get any kind of output. I might re-enable it and with the low shelf in place see how that sounds, still trying to feel my way around this method.

I need to measure again though, I kept the crossover on my tweeters (no caps, did not want to blow them) and ended up with a bump right at that crossover where if I turn the volume up it starts to sound really harsh. Lower levels it is fine, but with the volume up.... ooofff.

Edit: Looked at it again and realized my error with regards to Bass Management... on the routing tab, I left my subwoofer on inputs 7/8 (TOSLINK) instead of moving it to the Bass Management line at the bottom. I updated that, enabled Bass Management again and have a really solid low end. It's sounding pretty great already, wish I had time today to run a new set of sweeps with a safe crossover point, but my regular crossover disabled, on the tweeters.
hey what’s up mang !!! :)
 

·
Registered
Joined
·
4,980 Posts
Discussion Starter · #80 · (Edited)
With this in mind, would we also want to apply the midrange's HPF to the tweeter? I'm guessing that tweeter is far enough out of band that it wouldn't make a meaningful difference, but since the last several posts are basically all about keeping phase correct on all drivers, I thought it would be worth asking.
I did that in a mosconi aerospace, it didn’t do a lot actually, although it did do some.
It’s the 10 some odd ms group delay/ phase shift that’s the big issue with the 70-80hz crossover .... a 400hz crossover is almost 4ms exactly..... and Dirac can make up for that without any smear.....

Minidsp did it right, they only put in what we need , the 2nd cascade honestly is not important.....

Edit , and another thing , even if something is out of band.... the shift is still there , and adds the delay to the entire magnitude... like example , you can go to rephase and sim this, let’s say you have a tweeter, crossover is at 3000hz , now go and add a minimum phase peq anywhere in the tweeter magnitude (the lower the frequencie the more the shift) you’ll see the minimum phase peq make a small ripple in the flat phase response.
then add a 2nd HPF at 80hz and watch that ripple in phase at the peq band move quite a bit.

the 1st HPF at 3000hz would make that 80hz HP way way way out of band... but it still affects all the phase all the way to nyquist.

so the tweeter and midrange get the phase shift from the bass management crossover.... it really pulls everything into place
 
61 - 80 of 236 Posts
Top