[oabeieo]

This is pretty cool man, thanks for spending the time I reposted it on the Mini DSP forum , hope you don’t mind it’s really we’ll done

and seriously thank you for spending the time for us. I will be referring a lot of people to use that that’s a really good write up. Awesome contribution thank you

Curious if there are any plans to share this with miniDSP

Already did

this will help a lot of folks

I shared it here

Minidsp thread

 

[oabeieo]

I would like to know where people are placing their microphone in relationship to their drivers seat or relationship between passenger and driver if that’s better


Sent from my iPhone using Tapatalk

The first position should be directly between your ears

The rest of them are kind of whatever start with a small box and make it bigger and find a averaging cube that you like

The smaller the box the more tight the average will be, the bigger the box the more lively the sound will be at the end because it will get more of the room

so you’re still looking for a dead accuracy go with a small box if you’re looking for overall better sound quality for yourself and everyone else go with a bigger box

That is without explaining things to technically

 

[oabeieo]

maybe shoot minidsp an email, as I suspect the higher-ups aren't that interested in the forum on their site, tho devteam certainly pitches in

This is kind of a big deal and needs to have a link to the document on the product page, possibly the manual? a link that is, live document and all


Bravo on the guide otherwise, really clever work in there.

Also, in 6.b.ii.4.a of the guide, "do not apply a hpf" .... is that to say, bypass the hpf we set in 4.e.iii.1.b ?

they watch the Fourn , and I have a connect w mini and they have reached out

they are slow to do things , but when they do they go all in …..

They also have to adhere to Dirac “recommendation” so ….. it’s more a Dirac thing actually and Dirac is developing a multi-way tune
So that’s that I suppose

 

[oabeieo]

Makes sense thank you. The one thing I’m confused about is I put it 2-4 measurements behind the chair a little based off the illustration because it seemed like it was saying do some behind you, it didn’t sounds good so I deleted those measurements

Yeah the box kind of shows a 3-D looking thing where the rear ones look behind you but they should be right above your head and right at your shoulders

I will make my rear box smaller ( head size ) in my front box about 16” ,

I like to use the couch with tightly focused

and the two side measurements I get into the passenger area for the right, and then for the left directly on the glass one low and one high

 

[oabeieo]

Greatly appreciate the guide and everybody's contributions to it.

I have a few questions:

Why the do you recommend a 1.25 x fs BW12 protective crossover. Why BW12 over LR12?

How will the protective crossover BW12 affect the tweeters phase and actual crossover if using LR24 electrically?

With an Audiofrog GB15, this puts the protective crossover at about 1480hz really close to the recommended minimum crossover of 1800hz.

Can I just use a sharper slope at one octave lower then desired crossover frequency if the measurements are not to loud?

I ask all this because it was my impression from the manual that Dirac cannot correct for phase through the crossover when using separate Dirac channels per driver vs 1 channel left drivers and 1 channel right drivers.

I figured two different crossover types BW12 and LR24 on the same driver at different frequencies would mess up phase.

Thank you.

I don’t think anyone is recommending a BW or a LR, I think it all depends on what your system goals and speaks and power are…

The LR12 or BW12 are fantastic Crossovers, I love both! (In a complementary setting)

I would stick with LRs in a multi-way, except in the case of a 4 way active I would use one (ONLY ONE) BW alignment.

Quadratic crossover s are good for highly reflective environments. You don’t want 90-180-270 tho , so keep it LRs except one . So I would use one between the midbass and midrange…. The low pass is lagging so you’ll have to accommodate for the 90deg in the sub (or the tweet depending which polarity invert you use if 2nd order) , but that’s easy peasy (it’s the lowest speaker in the chain so simple delay you can eat up the group delay of the shift of the adjacent crossover and get back to 0 (or 180)

Or just use all LRs

 

[oabeieo]

Actually Andy at AF ( @GotFrogs ) has has a fantastic explanation, here is a link… his explanation is simple and eloquent… I absolutely love how he explains it here

He doesn’t go into the GD that is in the low pass, but if you pay attention you should get it, because he does say enough to put one n one together

Crossovers: How They Work and How to Choose Them - Audiofrog

There are lots of choices and even more opinions about which crossovers are the best. Arguments for and against the various slopes and alignments often include all kinds of crazy suppositions about the effects of phase shift, arguments that purport a gradual roll off to be better because they...

www.audiofrog.com

 

[oabeieo]

Will be interested to see where this goes.

I've tried both methods. For my setup, a 3 way tune works better. I'm using 1 channel for L, 1 for R, and 1 for Sub.

Then again, the setup I have is actually pretty flat / close to target curve with minimal EQ to start. I also have pretty good measurements for delay in the setup.

I'm in a "whatever works" kind of mode, so if this works for the car and creates the best results? Awesome. For me following MiniDSP and getting a "pre-tune" close (timing alignment, Xovers, etc) where I want has had better results.

Would love to see more thoughts on "box" size vs results. Also, I have not tried, but would be interested to see how much using 13 or 17 points would "help". To do list never ends...

I’ve always thought you can’t argue with results

If you like it , then that’s a win

Get this , in my shop , I have some hangs with big horns and 10” midbass

In the sweet spot it sounds great , directly under and behind the speakers they sound (to me) even better, and it’s like 120deg off axis…. So strange , but I love the response. The detail is amazing and midbass thumps….

The responce there shows a bump at 150 and a wide dip from 500-4Khz , idk , I like it….

 

[oabeieo]

Ide like to give one big thank you to everybody who contributed to this guide. 👏👏👏

Words cannot describe how beautiful the sound is emitted from the speakers. There are no more speakers. 🤩 @oabeieo hit the nail on the head with his spaceship comment in the older thread. 👽

If you are lazy like me or do not have time for all the pre-eq required for a 2-channel tune, this Dirac channel per driver method is for you. Blows 2 channel no pre eq out of the water.

You can do all the plugin setup at home and do the measurements somewhere quiet without having to rush.

Followed the guide to a t.

Used my amplifiers high pass filter at 12db/octave at about 1000hz to protect the tweeters.

Removed all crossovers in the plugin except the subsonic filter set on the sub output channel.

Left all drivers in one group, dragged the curtains on the sub to the roll off along with the right tweeter.

Loaded the Dirac curve to slot 2 and the Audio Frog curve to slot 1.

Left bass management at 80hz high and low pass.

Enabled all crossovers including the woofer's low pass filter on the subwoofer output (which I do not fully understand yet but believe it has to do with phase?)

Never knew a car could sound anything like this.

Great job everybody!

Eventually will try putting each pair of drivers in separate groups and report back.

Somebody had mention drawing the nulls in separate left and right curves for a 2-channel tune, but with this method I think we would need each driver in its own group to do that.

Anybody drawn in the nulls using this method with success yet?

yes!!! You know what it is when tge tune is so dam good and everything is just so amazing….. and it’s flat out Spaceship!!! Yes!!! Hahahahaha!

if listening to your system makes you feel like if you walked out in your front yard and saw this, then it’s awesome! Gloryhole spaceship mode is my fav

 

[oabeieo]

I was talking with @Anu2g about doing an amendum for rear fill. I think it’s a fantastic idea because there’s some things that should be added for setting it up properly most importantly doing the L minus R in the DIrac Routing (NOT THE MIXER)

and also, don’t take Dirac measurements with the L minus R in place, that must be done after all of your tuning is complete, and the fronts are exactly where you like them , then you manually work on the rear fill.

and while you’re working on the rear Fill doing things like adjusting the PEQ at the outputs of the rear channels and levels so that The minute and I mean ultra minute phase differences between front and rear get fixed even though you may be botching the response and deviating from your target. Because the use of PEQ Will make very small frequency dependent phase changes, and it will basically fine-tune like ultra fine-tune the summation between the rears and the fronts , I mean I would do the checklist like this

do anu2gs steps and for rear speakers treat them just like any other full range dirac channel,

After Dirac get fronts made perfect with your target etc etc with rears off after measurement are done and auditioning target responses.

after fronts done turn back on the rears, go to the Dirac routing tab. Implement L minus R

then go to the output PEQ and gain for rears,
Set signal delay between eight and 12 ms. Just enough to break away from the fronts spectrally.

adjust levels between left and right and find the sweet spot

and then listen to dynamic music and this is the last part but it’s the icing on the cake you have to be a good listener and listen to your music as you linked together the EQs for the rear outputs then start in the 80hz to 2.5k range and move different frequencies up and down using boost or cut and different Qs and listen for the rears to be 100% coherent with every frequency coming from the fronts! This is the fine-tuning stuff that makes the rear fill ultra bad ass. Remembering you’re not too worried about the frequency response of the rears as much as you want the timing to be corrected using PEQ to adjust phase

different spacing between the low doors in the rear or the C pillars or Ds Will require different delay settings, between eight and 12 ms.

and ideally having the speakers up high are better. if you can cut your rear fills into the roof above your head you’re a stud and that **** will be so spaceship.

 

[oabeieo]

Lol @ "gloryhole spaceship mode" as a descriptor for what's going on here

ya know I’m pretty sure spaceships really do come out and hover over your car, especially after the very first really good Dirac calibration you’ve ever heard in your life.

It’s like the feeling when you stick your wiener in a Gloryhole (I’ve never actually done that but can relate) and then spaceships are landing all around you…. It’s like the best ever feeling, and the most confusing feeling merged together.

Lmao!! Sort of like aliens walking in on you while you’re jerking off. Hahaha

 

[oabeieo]

This is how I would describe a really bad tune

oh it was so funny I’m still laughing..
(as my head hurts from being hysterical for the last 20 minutes)

Okay okay time to be serious, I don’t want to mess up a good thread anymore

did you have a chance to think about doing a rear fill section…..

We’re asking so much. Lol , the can o worms is spilled and word is out on your manual… now we want more dammit! Lol (jk) 😁

I’ve tuned nearly 30 8x12s now and all except 2 had rears….. I really do think it would be a great add on….

 

[oabeieo]

Sanity check question -

This process will not work with a DDRC-22D that is hooked up before an existing DSP (like a Zapco HDSP-V or whatever) right?

nope ,

although, it’s a good idea to do a solid L and R tune using moving mic averages with crossover off using the driver eq on your helix or zapco

then sum them together (playing with levels) starting with tweet to mid and adjust levels so they sum with what will look like about a half oactave bump during overlap, then engage crossovers and should sum flat, then midbass to mid and then sub to midbass

all while thinking about your target shape

you will definitely create a bunch of decorlation between left and right (doing separate left and right eq) , Dirac will fix it in the sum and it will image good so worry not. And that is new to 2.0 and 3.0, in 1.7 it wouldn’t fix the left and right phase errors…. It only made transfers match. So I know I preach no separate L and R eq (or preach with caution as the center will surly decoralate. With a ddrc22 upstream, worry not….. make your drivers flat and stack the levels to start your target tilt. Most ppl automatically assume the midbass needs a lot more power , that is false. It entirely depends on the efficiency of your system as a whole. Just stack your flattened responses to begin your shape, not by much tho. A little goes a long way, like make the midbass 1db above the mids and mids and tweets the same level. It’s important that you keep the crossover shape mostly as they need to sum properly and have phase tracking. If you over gain the midbass, you’ll overshoot your crossover point ot is won’t sum properly (and that defeats the entire point. So just use good judgment think about what you’re doing and the end result… know your system limitations etc. ideally tho you would sum to as flat as possible and build your target into Dirac curve…. All depends on your gain and what kind of power you have. And remember, if your driver tuning measures flat and the sum of them is a mess , Dirac will fix that, you just want the overall to be flat (ish)

if you have big dips when turning crossovers on and speakers all on together, well that’s an indication you need to reverse the polarity on a set and think about your filter order. With even order butts you should expect a small bump in the crossover area. That is ok!! Leave it! Let Dirac fix the sum… on LRs and odd butts they should sum flat. Listen to your crossovers also… make sure nothings fighting….like for example, having a mid and a tweet very close by and using different time delay settings, that can cause a dip at the crossover. I would try using a common time delay for both and somewhere in between unti you have good reinforcement. Somewhere in there should be the best possible. And acoustically, they may just push on eachother. What you don’t want is using delay outside of the actual distance.

Personally I like to start with a common distance of both speakers mounted next to eachother. (Example if tweet is 28” and mid is 26” I use 27 for both) It’s sounds the most natural post Dirac, and in my oppinion and on most cars I’ve tuned. Of course some exceptions. Have both speakers interacting and having cone movement at the same time when there mounted close together is a good thing. Because remember phase is Angular! Just because a delay scheme may be fitting to your nose, it’s quite a mess at let’s say all the outside Dirac box measurement points. And that “perfect “ time delay is no longer valid.

so save yourself from trying to align impulse peaks gang. It’s a worthless effort. The average distance of two drivers mounted close together is almost always going to be better sounding. May not measure 📐 as precise at one point in space, but my head is bigger then a mic capsule. And when I drive I move my head around, lean on center consul armrest , etc etc.

 

[oabeieo]

The default target ive tried several times now
It works really well….

On my CM-10 mic it sounds so good. It sounds like it fits the room and the responce actually sounds like I would expect from the target. The CM-10 I believe gives Dirac that last bit that makes it what I believe it was always meant to be. Especially in tonality! Hands down cm-10 wins (big time)


on the umik1 it has too much midbass and is muddy imo (ever so barely I might add)

on my UMM6 same thing as umik1 except the highs are brighter a little so it makes up for what sounds muddy to me

the umik1 and UMM6 on a perfectly flat target sound proper and sound good….

I’ve tried this on 2, 8x12s, a flex, and a ddrc22
All same results, all different “rooms”
The flex is in a big room and it also shared the same attributes as the others that were in car.

ive ordered a umik2. We will see how it measures.

 

[oabeieo]

@Anu2g - did you ever use a UMIK-1? If so, any comments on it vs. the UMIK-2 in terms of tuning results?

@oabeieo - how are you connecting the CM-10 to your PC?

with a tascam us-366
But my Roland just showed up… super excited.

i wouldn’t mess with the tascam, it’s junk imo

this is the hotness, and an earthworks mic (oh anu2g I’m so jel)

Roland Octa Capture

 

[oabeieo]

Are getting a pile of mics with that?

Nope , I bought 7 more cm10 s

but Im also getting the jl max at the shop, that will have a pile of mics. The shop is buying that (hopefully soon)

 

[oabeieo]

Wow man I can't thank all of you guys enough. Amazing!!!! I have the C-DSP 8x12DL installed and I'll absolutely be using this guide.

ive heard rumors of tonal issues with Dirac
Remember it’s false

Once you get the hang of it and find your target you will never go back to standard dsp

 

[oabeieo]

Yeah you can do Dirac 7 and 8 on separate subs and still use BM as a source feed

you’ll just have to use a y cord for the amps

you have 2 bass amps I presume

 

[oabeieo]

Perfect. Yes 4 channels on 2 amps each bridged to mono, so I'll need the Y cables.

wait …. For subs ?
two 4ch

so you have 4 subs ?

 

[oabeieo]

Yeah that will work awesome

i have that on a customer truck it’s pretty dam sick

 

[oabeieo]

I've been playing around with this for a bit now. Only issue having is that my response post tune is quite hot on the top end (subjectively). If I do a pink noise measurement post tune I have rising response above 5khz. Pink noise vs a sweep might be responsible here, which suggests I need to use a different target curve. Any similar observations from you guys?

absolutely yes

then I use a cm-10mic and it’s consistent every time… tonal issues resolved

but yes your target needs changed

what I’ve found with Dirac is you just have to find your target….

it becomes a thing where “who cares what harman says” lol

i think the inconsistencies between mics and measurements swing as much as 4db!!!

i also found Dirac with any calibrated mic to work best with a smooth curve…..

Especially above 250hz

like a flat line (whatever tilt isn’t what I’m talking here but a flat shape). Tilt , no tilt , a tilt down then a tilt up (like I do rn) as long as there’s no abrupt changes…. No more then 1db per octave on the target above 250hz itseems to work it’s best

I have a flat response with 4 anchors to 600hz then a 1 db drop with no anchors after that to 2500…. Then a tilt back up 1db to 9k then a 2db tilt up to 20k

makes it crisp, smooth , the 20k is not louder then 600….. it’s my curve with a cm 10

on my umik , no tilt back up , in fact a 1db tilt down and a 2db drip at 4K instead of 2500

so different mics , different curves , the generic targets all sound good initially, but sorta blow after a min

find what you like , small variants… start with all the way flat and work from there