DIYMobileAudio.com Car Stereo Forum banner

MiniDSP C-DSP 8x12 with Dirac Live

91K views 484 replies 58 participants last post by  oabeieo 
#1 · (Edited)
#6 · (Edited)
Erin, Thanks for taking the time to provide us this detailed write up :thumbsup:

I have a few questions:

This gets me to #3: Equalizing the driver response. Now, to be honest, one should also take the time to pre-measure each active driver’s response and then use broad strokes to flatten out any large peaks or anomalies seen. However, I didn’t do this my first go-round because, frankly, I was too giddy at the thought of this new toy and I went straight to Step #4.... ...
With all of this in mind, you can understand why MiniDSP encourages us to pre-correct the response of the "speaker" before we run DL, especially when used in the car environment.
It's unclear, to me, if you ever went back and did some broad stroke driver correction?



Step #4: Set up my crossovers for each speaker driver... A portion of Step #4 includes setting time delay....
Did you go as far as to align acoustic xovers up, at the listening position, so as to minimize phase issues in the xover regions?

Did you adjust levels and apply xovers so as to maintain a flat response from 20-20khz or to loosely follow the target curve you had in mind?

Also, when running the Dirac channel per driver measurement in order to obtain the delays - Were all the xover filters engaged?



Here's the same screenshot but of only the DL targeted - modified Wisdom curve (orange) and the optimized response (green):

Would be interesting to see the actual measured response and how it aligns with the theoretical "optimized" response. Did you happen to take any after measurements?
 
#7 ·
Erin, Thanks for taking the time to provide us this detailed write up :thumbsup:
Thanks, Ryan. :)


It's unclear, to me, if you ever went back and did some broad stroke driver correction?

Did you go as far as to align acoustic xovers up, at the listening position, so as to minimize phase issues in the xover regions?

Did you adjust levels and apply xovers so as to maintain a flat response from 20-20khz or to loosely follow the target curve you had in mind?

Also, when running the Dirac channel per driver measurement in order to obtain the delays - Were all the xover filters engaged?

Would be interesting to see the actual measured response and how it aligns with the theoretical "optimized" response. Did you happen to take any after measurements?


I did not. But I'm glad you asked. I meant to bring that back up again in my final assessment, particularly because I did little to optimize the response before DL had it's turn. What I did do was set crossovers on my compression drivers so that the horn loading effect was flattened out, and I did add a broad filter on my midranges around 800hz (a shelf would have been more effective but I just didn't do that), and I set my T/A based on DL's result of each active speaker. I failed to mention this I believe but I also used DL's gain adjustment values from the active driver sweeps to set the levels before I re-ran DL in 2-channel stereo mode. I'll need to update my OP to add this info.

But, overall, I did little before I ran DL. I think that speaks even more to the benefit of DL's capabilties based on the results I got.

I also should have mentioned this in the review, because I did mention it in another thread, that it would make more sense to me to not work to flatten the response too much in areas where you plan to have a curve that already mimics the natural response. For example, if you have a high frequency response that is falling, and you plan to have a target curve that is doing roughly the same, I wouldn't try to pre-correct the signal to flat; I think that would just be wasting time and DSP resources. But that's just a suggestion.



Would be interesting to see the actual measured response and how it aligns with the theoretical "optimized" response. Did you happen to take any after measurements?

I did this really early on but didn't think to grab screenshots. The response I measured was practically perfectly matched to what DL targeted and resulted in. The higher frequencies didn't match as well but that's more a function of me using a single mic measurement when I re-measured again. I think most of us know that it's imperative to take an average of measurements over an area because even moving the mic one-quarter of an inch can alter the measured the response. This same question came up in a thread on one of the home audio forums, though, and the folks who replied with data showed that DL's corrected response was indeed what it said it was. So, I wouldn't personally spend time trying to verify if DL actually did what the software says it did. Just my $0.02.
 
#10 ·
Excellent write-up Erin... appreciate your time and information.

I am sure it would be super helpful to a LOT of people to have a step by step guide on using this system from beginning to end... particular on how to setup the routing and mixing tabs.

1. Do this
2. Do this
3. Do this
etc., etc.

The manual is mind-boggling to me... way more info than is needed and let direct to the point instructions on how it is to be setup. I realize there are a lot of options, but similar to what you suggested... most people buying this product are doing a 2 or 3-way active front with a subwoofer somewhere behind them... and maybe rear fill.
 
#11 ·
great write up. Much appreciated. I always wanted there to be a great auto eq option like this.

As of what I tried (pioneer auto-eq, JBL ms8, Alpine Audessey) the kicker key had the only auto eq results that didnt require major tweaking but its a low power amp. Super impressive and Id like this option on their larger amps. This is wayyy more flexible and tweak able.

To be clear the unit is $500, then to be able to use DL will cost about $400 more. So roughly $900 to give it a go.

I am very intrigued about using this with a 2 way in front and rears + sub.

Is there surround sound processing of some sort or any effects for rears in this unit?
 
#12 ·
great write up. Much appreciated. I always wanted there to be a great auto eq option like this.



As of what I tried (pioneer auto-eq, JBL ms8, Alpine Audessey) the kicker key had the only auto eq results that didnt require major tweaking but its a low power amp. Super impressive and Id like this option on their larger amps. This is wayyy more flexible and tweak able.



To be clear the unit is $500, then to be able to use DL will cost about $400 more. So roughly $900 to give it a go.



I am very intrigued about using this with a 2 way in front and rears + sub.



Is there surround sound processing of some sort or any effects for rears in this unit?
The unit plus software is 900. I think the umik comes with it and you have to use the Umik-1 mic with it.

There's no upmixing for center or 7.1 or anything like that but it does have the ability to do rearfill.

I promise. Its worth every dime.


Sent from my XT1710-02 using Tapatalk
 
#14 ·
I have been reading the posts on here and the MiniDSP forums about the live units... and have wondered about getting the 24hd version and using it along with my helix mini dsp, I have a apl1 now which I have had since 2015 and don’t think I have used it to its full potential. Would it be worth it it, as in not using the xover, time alignment part of the MiniDSP running it after the head unit into the helix just for the eq ?
 
#15 ·
Are you asking if you should buy a 2x4HD and put it inline before the Helix? Or are you asking if you should buy a CDSP 8x12DL and put it inline before the Helix? Regardless, the answer is no to both.

If you have an APL and Helix Mini then you basically have the same functionality as the unit that is reveiwed above albeit less channels of Dirac and an additional AD/DA conversion.
 
#22 ·
So does anyone have any videos or screenshots of using this software? Even if it doesn’t have the Dirac software? This review really peaked my interest in the minidsp but I literally can’t find anything about the software and there is nothing to download unless you purchase a unit.

Any help would be appreciated.


Sent from my iPad using Tapatalk
 
#24 ·
Yeah I'm well aware of the product page. What I'm saying is, most other DSP programs you can download and try in a demo mode. With MINIDSP, I don't believe that's available. I was wondering if anyone had taken any video of them tuning, configuring, or a tutorial of sorts.

Sent from my SM-G965U using Tapatalk
 
#34 · (Edited)
I upgraded my 8x12 v2 to DL in June. I've tried 4 different tunes with it and have not been fully satisfied. Up until this weekend I had only used the time delay Dirac calculated & combined that with a manual tonality tune... I could never get a full tune utilizing Dirac to sound tonally pleasing. But, I will say that just using those Dirac calculated delays (which were much different then delays by distance alone) with a manual tonality made a big improvement in cohesion between drivers.

This Saturday, I worked on a complete ground up tune fully utilizing the Dirac live correction. Like Erin, I ran it through in multichannel first to obtain delays & levels which I plugged in to the output tabs and then ran it again through a 2ch Dirac (L/R). Unlike him, I spend a couple hours pre-tuning basically doing some broad stroke EQ & adjusting xover filters on the outputs and to get all drivers roughly following individual driver target curves I had generated with Jazzi's spreadsheet. The results were glorious!

Yesterday, I competed in the MECA NE 4x with that tune. It was the first competition for which I was not in my car for 1-2hours, before being judged, waving a mic around my head to make last minute improvements because it didn't sound "just right". The full Dirac tune was that good! I placed 1st in Mod-Street (3 competitors total) & 5th out of 19 in the money round. My drivers are in stock speaker locations - midbass in doors, mid in dash corners bouncing off windshield, tweeters in sails, trunk baffle sub.

Possibly the first car utilizing Dirac Live in the competition circuit?

It eliminated the midbass nulls I had that eq wouldn't touch. Playing by themselves: Left had a deep half octave 12db null centered at 160hz; Right had a deep half octave null centered at 180hz. These nulls combined to a 10db half octave null around 170hz. They are all gone!

Now that the groundwork is laid, it's literally takes 5min to change the target curve, re-optimize and load it into a new preset. I'm in love with this thing!
 
#35 ·
Congratulations!! That's awesome. I think some of us are starting to now really tap into the potential of the DL unit's and figuring out the best way to utilize them.

Quick question for you, since if I remember correctly you have a single subwoofer. How are you assigning the sub in the Mixer / Routing tabs in order to get the L/R tune?
 
#36 · (Edited)
Thanks Ian!

Yes, I think it's really coming together now. I'm considering writing up a little tutorial.

For the multichannel (7ch) Dirac config: Dirac7 was routed to Output 8 (Sub) in the mixer tab.

Then for the 2 channel Dirac config: Inputs 1/2 (L/R) were routed to Dirac1 & Dirac2 which were then both routed together at 0db to Output 8 (Sub) in the mixer tab.

My amps input switch is set to only use L input for the bridged channels which is coming from Output 8 on the DSP.

I did not transfer the gain (-13.5db) for the sub channel to the output tab from the results of the 7ch config optimization. I tried it with that gain but sub level was too low... so I tried again keeping it at 0db and sub level was perfect.
 
#40 · (Edited)
Yes learning what pre and post tuning is important with Dirac live

And everyone’s cars will be a little different , one car may need a small level ajust on maybe a left mid turn down by a dB or two, and post tuning is mostly all just fine tuning levels never any eq work, at least for me.


Dirac will “undo” some of the “pretune” you have and it’s all wasted coefficients and added filter ringing at that point. So knowing what needs pretune is part of the curve in learning. And So knowing what to expect after every trial it gets better. The nice thing is it’s only a few try’s and you get it.

For anyone that’s on the fence. It takes a few go’s at it. Once you get the hang of it and know what it’s doing that only a couple times will teach hands-on your cars will be amazing. And Dirac will grow on you and quickly become something you won’t want to live without. There’s a psychological process definitely as some of your old tunings may be what your used to and like more. But that is temporary and going back to your old tuning will reveal that this really is better than what you could do on your own.
 
#41 ·
As previously mentioned, I tried to use DL to measure and optimize each individual driver to a specific target curve a number of ways. The truth is, it just didn’t work well for me. My best results were met by using DL in stereo mode with the subwoofer getting half of the summed signal as it was summed by combining Left+Right in the routing tab. DL suggests this and my results back this as being the best method. But, again, it might be useful for you to first run DL on each individual driver to let DL tell you the time delay and level values to set for the corresponding output channels in the MiniDSP. Then re-run DL in 2-channel mode after you’ve made those adjustments.
Erin... this somewhat confusing for me in what you are doing here. For clarification ... miniDSP/DL recommends measuring the subwoofer separate from the other speakers. The 2-way or 3-way system with the frequency response outside of the subwoofer range is what they recommend measuring as one speaker (supposedly if they are not too far apart). This is similar to the home... we measure all the full range speakers and the subwoofer separately.

Can you share your XML file.... or post a screenshot of your routing and mixer tabs?

I wouldn't mind trying what you are suggesting to see how it sounds, but it seems like it would confuse the bass management.
 
#42 ·
Erin... this somewhat confusing for me in what you are doing here. For clarification ... miniDSP/DL recommends measuring the subwoofer separate from the other speakers. The 2-way or 3-way system with the frequency response outside of the subwoofer range is what they recommend measuring as one speaker (supposedly if they are not too far apart). This is similar to the home... we measure all the full range speakers and the subwoofer separately.



Can you share your XML file.... or post a screenshot of your routing and mixer tabs?



I wouldn't mind trying what you are suggesting to see how it sounds, but it seems like it would confuse the bass management.
So the unit has 12 outputs. 8 of which can be used as dirac. We've found through testing that its actually best to only use 3 or those outputs in the whole system, not considering rearfill and the like.

So basically. The first time you run dirac. Use as many channels as you have speakers.

You have a simple 6.5 and tweeter and a sub woofer? Thats 5 channels for 5 speakers. So do the 9 measurements and then send it into the box. What you're after here isn't actually a tune but the delay and levels that dirac does. You can see this via the Dirac tab in the plug in.

Now, write those down. And make a 2nd tune on the 2nd tab of the plug in. Same crossovers, same pretune eq if u choose too, but this time input the levels and delay that dirac used for tune 1 on the first tab.

Then run Dirac once more but instead, use inly 3 channels of the 12 total.

One for left. One for right. And one for subwoofer.

Its works a lot better.

Sent from my XT1710-02 using Tapatalk
 
#46 ·
Looks like about the only difference in yours and mine is I use the Bass Mgt tab on the routing per the recommendation of miniDSP. But I do link my two sub channels at output in the Mixer.

You do eliminate two of your input channels on the Routing tab, but that won't make a difference as the fronts are all getting the left and right input either way. The Mixer tab will determine how they are output for Dirac measurements.
 
#48 ·
I'm not sure if anyone has mentioned this idea yet or not, but I was brainstorming about process flow and came up with an idea which I will be trying out when I get my next build finished (where my 8x12DL is going).


My plan is to attempt the following:


1. Set delays in the plugin using SysTune impulse response, like I always do.

2. Measure each driver individually using the 5 mic array for spacial averaging.

3. Export the measurements from SysTune and import them to REW.

4. Use REW to build EQ biquads and export those to miniDSP. This can be done to a flat target curve, or to whatever curve we want.

5. Finish plugin base tune by adding crossover points.

6. Run Dirac for 3 channels, left right and sub.
 
#49 ·
I'm not sure if anyone has mentioned this idea yet or not, but I was brainstorming about process flow and came up with an idea which I will be trying out when I get my next build finished (where my 8x12DL is going).





My plan is to attempt the following:





1. Set delays in the plugin using SysTune impulse response, like I always do.



2. Measure each driver individually using the 5 mic array for spacial averaging.



3. Export the measurements from SysTune and import them to REW.



4. Use REW to build EQ biquads and export those to miniDSP. This can be done to a flat target curve, or to whatever curve we want.



5. Finish plugin base tune by adding crossover points.



6. Run Dirac for 3 channels, left right and sub.
Did that but with Rew. Dirac doesn't average the measurements. So the pre eq you do is pointless

Sent from my XT1710-02 using Tapatalk
 
#52 ·
Lately I’ve been doing the rear measurements
About 6” apart top to bottom, at top my shoulders with mic pointing out from my ears and at top of my ears and directly touching my head

The outermost forward measurements at my knees and the top ones R in passenger seat and L about 6” in front of my 1st measurement point. (Parallels)
And the 1st measurement at my nose.

I likeee so far.

Having the rear measurements pulled a little tighter seems to get what’s going on at my ears more true to the target and what the avg shows compared to my moving mic RTA

So far so good, my target seems to be easier to deal with and I have nice smooth curve that sounds good
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top