DiyMobileAudio.com Car Stereo Forum banner

561 - 580 of 928 Posts

·
Premium Member
Joined
·
97 Posts
The reason I think the subwoofer check box corresponds with the LFE channel is because when subwoofer is checked I definitely heard a different stimulus. Also, I don't see what the purpose of the subwoofer check box would be if not to account for this boost, again, since DL doesn't know the source signal will or won't be boosted without you telling it; and the last thing they'd want you to do is mess with the subwoofer level after the fact. I don't have anything solid to base this on, just some "if/then" assumptions that I gave. But, it's simple enough to find out by running the setup without the sub box checked. I don't care about being right or wrong. I care about getting the most out of this thing so I'm just trying to provide some options and logic as to why I'm suggesting what I am. :)


Also, Ian, I still recommend you reconsider the subwoofer target you've used. I can't help but think that DL is trying hard to get to that and just can't do it without boosting a whole lot of frequencies. And, no, you can't have my Mc amp!!!!!! :mad::D
I believe the main reason for the subwoofer option in the dlct app is for the subwoofer volume control. In subwoofer vol mode, the vol knob controls the boost Or cuts for those channels marked as subwoofer in the dlct app.
 

·
Registered
Joined
·
1,561 Posts
I believe the main reason for the subwoofer option in the dlct app is for the subwoofer volume control. In subwoofer vol mode, the vol knob controls the boost Or cuts for those channels marked as subwoofer in the dlct app.
Could be but why wouldn't the manual state that?

The manual only states that "the subwoofer checkbox tells the Dirac Live analysis algorithm to use a different method to detect the impulse on that channel... because of the limited frequency response of the subwoofer" which makes sense.

Pretty sure the remote sub level control, which has to be assigned as outlined in section 7.3 on page 55 of the manual, only affects the output levels, as it does in the Non-DL version, outside of the Dirac function.... Which is why I think they include the note in section 5.3.3 - "Check that the subwoofer volume is set to zero. (Remember that this volume is relative to the master volume. Zero is typically the best setting to use for calibration.)"

But then again, manuals aren't always %100 correct or complete so :shrug:
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #564 (Edited)
I really like the idea of only a 2 CH Dirac tune... Mostly since I only have a ddrc24, LOL....
So I would like some advise on the best way to set this up. I have a 3-way front, plus Sub. I think I need to simply have ALL left plus the L ch of the Sub on Dirac 1 and similar for the Right side. Then tune as discussed through this post, including some finer details in the below quotes... :D

I will have basic TA and XO done of course first.... Those 2 ddrc24 outputs will go to my Helix P6, but all EQ on the Helix will be bypassed. I will have the Sub XO and levels set on the P6 SUB outputs but simply passing through to the SUB AMP, right?

Oh, and would it make any difference if the pre-tune is set in the Helix OR DDRC24?


Honestly , Dirac does it’s thing the way it wants to,

Some say a pretune is the way, I find it over corrected, and after reading the diracdoc I can see why I find it over corrected as it’s meant to ignore certain things

It’s mire about retraining your ears to accept the new ways of listening
And it’s better! I went back to the old ways several times and was so disappointed once I gave it a chance and let it do it.


I do however strongly believe in a post Dirac tune for taste and for better crossover interaction, the Dirac doesn’t know how good or bad the crossovers will interact after its correction and sometimes a little delay needs to be added to a speaker after to get things staging and sounding good again. And just for personal taste


A ddrc with a helix is solid way to go. Although I hate the helix file structure, it’s still a very good dsp. The C would do as good or better


But Dirac does what it wants? Sometimes the after just needs a little love.
But even a little tweaking after Dirac is still miles better than what you do with just peq


If you haven’t bought that yes I would get the C

If you have the helix get the ddrc it will work fine


Edit:

I would still keep your helix functioning for crossovers and some tuning bound that as well. A ddrc upstream would be fantastic (I’ve done it actually)
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #565 (Edited)
I just tried 8ch Dirac
But linked all channels to single eq (so it behaves like 2ch Dirac but it does delays


I tryed it before with very screwed up results,

This time it’s good, it sounds correct

I don’t know if anyone else wants to try it also,
I think last time I was boosting somewhere ,

It shouldn’t work but it is actually pretty good.


I’ll measure the phase tonight in rew and see what it’s doing.


Minidsp has never said anything about doing it this way, I’m still a little confused why this time it’s working very decent, and last time it was a mess

:-/


Maybe last time I had gaps in crossover region and wanted to add boosts after crossover s or my target wasn’t low enough

Idk but it seems to sound a awful lot like 2ch Dirac now, your guess is as good as mine ....



I know I’ve said this is bad way,
I’m just screwing around , and ended up working now.

I’m only posting this if anyone else wants to try it

If you have some 8ch measurements, linked all to one target and load a curve
I’m just curious if anyone else’s sounds good doing this
 

·
Registered
Joined
·
2,456 Posts
OK, just got done with some measurements.... tried with the subwoofer not checked on the output tab, seemed to make no real noticeable difference (either visually or auditory), certainly not a 10dB difference if it were thinking it was an LFE channel or not.

So, went back and took new measurements with it set as a subwoofer. I had tweaked the LPF from 24db to an 18db slope. I also modified the target to more closely resemble the measured response from the sub:



Overall Dirac response:


Measured REW (1/3 smoothing, pink noise, sweeps around the head):


Looking at the REW response, I think the output of the mids could be lowered. They play 800 - 4200Hz, which seems to coincide right with that bump. That could also explain some of the harshness at higher volumes.

But.... overall it is sounding a WHOLE lot better. The subwoofer is still lacking some for me, although it does match between REW and Dirac (about a 13-14dB swing from one end to the other). I think the next time I will try to raise the subwoofer in the output tab by 2-3dB, lower the mids 4-5dB and re-measure. Might find time to do that later today, especially now I know what level on the output tab seems to work for all positions.

My only real complaint currently........ Overall volume. Again, I am not sure if this is a behaviour difference between tablet and head unit. It just seems odd to me that for it to be loud (hitting 100dB playing music) that I have to have it at 75% volume. That could absolutely just be me though, especially as it is plenty loud enough, just odd to have to turn it up so much for it to be loud.
 

·
Premium Member
Joined
·
3,563 Posts
Not with Dirac. But I can measure with Smaart afterward and get a phase plot.
I think that looking at the phase for the sub, and the mid bass separately... will show if anything 'untoward' is happening in the cross over region.
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #568
I think that looking at the phase for the sub, and the mid bass separately... will show if anything 'untoward' is happening in the cross over region.
Totally agree,

And if he can do overlays with smaart man that could help us,

I’ve only been able to compare single measurements of individual speakers and a measurement of two (or more) speakers in REW with the same mic location and same windowing methods etc.

It’s clear to me it’s not trying so much to “linearize” or make phase flat, as much as it’s trying to make the phase the same for bolth left and right , the entire transfer function. I’ve seen the phase still have some GD properties and some wraps from XOs , but left and Right were so so much alike vs with Dirac off.


With smaart he should really be able to see what’s going on every step of the way,


John please do! I’ll put my learning hat on and let you take it away.
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #569
My only real complaint currently........ Overall volume. Again, I am not sure if this is a behaviour difference between tablet and head unit. It just seems odd to me that for it to be loud (hitting 100dB playing music) that I have to have it at 75% volume. That could absolutely just be me though, especially as it is plenty loud enough, just odd to have to turn it up so much for it to be loud.


I think that’s pretty normal with a lot of eq cuts, but unfortunately it’s necessary to make big cuts or it’s just worse. You should still have enough overall volume tho


As far as the harshness, can you isolate what frequencies it sounds harsh at?

Do you think maybe (just maybe) it’s one of the speakers disagreeing with the correction,

I have extremely efficient speakers that are so capable of any kind of eq you throw at them and this sometimes can put them right up to the end of what they want to do. Meaning,

If the speaker naturally wants to play a certain way and your asking it to do something opposite of how it wants to behave it could make it harsh kinda quickly or prematurely. Follow what I’m saying...

If you can isolate the frequencies that’s making it harsh , you could try a different crossover or maybe some eq to tame it and let another speaker play the part it’s breaking up (if that’s even possible?)

Just some ideas .......
 

·
Registered
Joined
·
2,456 Posts
As far as the harshness, can you isolate what frequencies it sounds harsh at?
Not sure exactly which frequencies yet, it's towards the higher end of female vocals. Need to give it a longer listen to be sure though.

Anyway, so I knocked down the mids by 4dB and increased the subwoofer by 3dB, then re-measured. I think I am experiencing the same thing that Ryan did with his measurements and the sub...

Truthunter said:
I tried upping the level of the sub to try and meet up with the target but no matter what I did the measured response stayed the same even though I could hear the sweeps were louder which baffled me.
The sweep sounded louder, visually they looked larger on the DLCT measuring screen, I also just compared the sub response on the filter design tab and it is exactly the same, despite me turning it up. Weird.

Anyway, took new measurements, loaded my target and now making some real good progress with this. Latest REW (1/6 smoothing) response looks like this:


I'm thinking if I can cure that dip around 55Hz that I may well be completely happy with this particular tune.
 

·
Premium Member
Joined
·
97 Posts
It’s clear to me it’s not trying so much to “linearize” or make phase flat, as much as it’s trying to make the phase the same for bolth left and right , the entire transfer function. I’ve seen the phase still have some GD properties and some wraps from XOs , but left and Right were so so much alike vs with Dirac off.
Are you in on the beta 2.0? I understand that one of the main upgrades on 2.0 DL is a better phase correctiOn between pairs of speakers.

“In addition to the usability upgrades, the new Dirac Live also includes an enhanced phase correction algorithm for improved stereo reproduction. Where the previous version of Dirac Live individually measured the phase of each stereo speaker, this new version also analyzes speakers in pairs, which ensures that the pair's phase responses are matched to each other.”

I am still hoping they release it soon. So that I can stop tinkering so much in my car. Wife is going to wonder soon why I am always sneaking to my car with my laptop in tow.....
 

·
Registered
Joined
·
1,561 Posts
The sweep sounded louder, visually they looked larger on the DLCT measuring screen, I also just compared the sub response on the filter design tab and it is exactly the same, despite me turning it up. Weird.
Right... I did it over three times turning the sub up 4db at a time and saw no difference on the filter design screen. Bloody Wonky :p
 

·
Registered
Joined
·
1,561 Posts
Are you in on the beta 2.0? I understand that one of the main upgrades on 2.0 DL is a better phase correctiOn between pairs of speakers.

“In addition to the usability upgrades, the new Dirac Live also includes an enhanced phase correction algorithm for improved stereo reproduction. Where the previous version of Dirac Live individually measured the phase of each stereo speaker, this new version also analyzes speakers in pairs, which ensures that the pair's phase responses are matched to each other.”

I am still hoping they release it soon. So that I can stop tinkering so much in my car. Wife is going to wonder soon why I am always sneaking to my car with my laptop in tow.....
IDK, it seems to do a decent job of that already. I mean, on a standard DSP, if I were to set delays by distance, then eq L/R to the same target and then measure the combined response with mono pink noise - the response would be all over the place because of phase differences between the two sides. There is a popular tuning guide that instructs it be done that way but it never produced good results for me. Dirac seems to take care of those differences as the combined overall measured response basically follows the targets set for each side.
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #574 (Edited)
Not sure exactly which frequencies yet, it's towards the higher end of female vocals. Need to give it a longer listen to be sure though.

.

So somewhere in the tweeter to mid crossover, uugh yeah that’s tough to get right no doubt.


I would try starting with the tweeter channels in PEQ link them together, than set a peq to a Q of 1.7 , turn it down to -6db , than listen and move the frequency slider up/down and try to see if the mode goes away, than do the same thing to the mid and try the same thing. Try to isolate that frequency,
Moving the frequency slider trying and listening to find that spot! And which driver it’s on

When you find the exact spot turn it back up to 0 and now you know where the problem is , , now try different crossovers , steeper slopes or move it up or down and try to get the other speaker to cover that band more (if you can)

If you can’t , you might try steeper or shallower slopes to take the load off and let adjacent speaker do more there...


If all else fails , try using your target and creating a notch in that band and that should (hopefully lessen its effects) the latter will only put a small dip in response, I would start with -2db and work that area

Are you in on the beta 2.0? I understand that one of the main upgrades on 2.0 DL is a better phase correctiOn between pairs of speakers.

“In addition to the usability upgrades, the new Dirac Live also includes an enhanced phase correction algorithm for improved stereo reproduction. Where the previous version of Dirac Live individually measured the phase of each stereo speaker, this new version also analyzes speakers in pairs, which ensures that the pair's phase responses are matched to each other.”

I am still hoping they release it soon. So that I can stop tinkering so much in my car. Wife is going to wonder soon why I am always sneaking to my car with my laptop in tow.....


Yes beta .... I was able to load filters into my ddrc22d as I got a set of measurements made in the C , literally the next day they ended the beta so I got super lucky to have clean measurements and working filters that work with my C .....

The beta had ended so I can’t take any new measurements until the release




So, to clarify , the current version makes both sides have as close as possible the same impulse (within its parameters) and does nothing about how the stereo pair interact.

In 2.0 (which is hot diggity dog for us) it looks at combfiltering between the stereo pair caused by one side being farther and having to use signal delays (which cause a whole new set of comb-filters) so it addresses some of that.

It doesn't do anything about comb-filtering caused by multiple reflections and compound reflections (copies of copies of sound arriving late) because that is just not possible.

Yeah the ddrc should be first to release, I’m anxiously waiting!

It’s only a slight improvement over what we have now. But still an improvement,
And the auto target is very intuitive, the auto target is actually really bad ass, where the auto target on ours now is the same for everything and everyone’s systems, and has nothing to do with the target and how it sounds on your specific measurements. In 2.0 the auto target is the cats meow
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #575
Right... I did it over three times turning the sub up 4db at a time and saw no difference on the filter design screen. Bloody Wonky :p


I don’t think giving sub more gain and than measurements is going to change anything, it will just add or subtract gain to match the target, (unless I’m not understanding you right)


I do my sub measurements with my sub bass knob very very low (1/8th way up)

That way I have the added gain at the end,



I know you guys want to stick to a target and not deviate, but honestly,
Heat changes the sub output, and different recordings have different bass.


I know what you meant in previous posts to have good bass and not have to crank it up , I do know what you mean. I agree with that also, and I too like sq bass , I’m not a bass head , I just like to hear it right compared to the rest of the music. but in my experience, and tell me what you think about this .


Like , I’ll measure with Dirac, set my targets , I’ll be super happy , the next day on way to work when it’s about 50deg hotter in my car, the bass just seems whimsy compared to last night, almost as if something changed.

The sub is almost purely minimum phase behavior, so turning it up with a bass knob to restore just enough so it’s sounds right again dosent make any ill effects on the sub. It just restores the bass for the conditions.

I’ve had excellent luck doing that. , it’s super easy to hear when the blend is right, I know you can hear it come together within a dB or two, it’s super obvious when the level is back, the ambiance comes back, the midbass is enhanced almost when the Harmonics between sub and midbass are just right and play together.


I still think maybe this is all that is needed. I had the same issues and this solved it for me , and gave me control over every different recording type and current road noise conditions.

Am I off base?
 

·
Registered
Joined
·
733 Posts
Newbie question: could you specify the sub crossover slope to whatever you want?

If you could, you should use 48db slopes so using a bass knob won't affect cohesion so much since you're using FIR filters.


Sent from my iPhone using Tapatalk
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #577 (Edited)
Newbie question: could you specify the sub crossover slope to whatever you want?

If you could, you should use 48db slopes so using a bass knob won't affect cohesion so much since you're using FIR filters.


Sent from my iPhone using Tapatalk


Not noob ,


But yeah you could , if one wanted


Or do like a 40hz 12db

The thing is , Dirac will “correct” any slope you put on it to match your target
Unless your target follows that shape .


If it’s in a post Dirac situation where you lower or add steeper slop than by all means



But , I don’t think it’s always necessary

I’ve experienced where the sub just simply looses power based on conditions, and that is in step with pro audio in concerts , FOH ppl are always adjusting the bass real time as heat from venue rises and conditions just change
And face it out in car conditions change quite a bit.

With the old way of tuning sub you more than likely didn’t have near the eq on the sub, so it could be some frequency are changing while others are not.
And wasn’t noticed as much.
It really is hard to say,


But to answer , yeah you could definitely do that.

I would just be sure it’s just simply not something related to power loss.


But I honestly don’t know the exact reason why this happens, it just does.
I’m kinda relieved to know it’s not just my setup that does this.
But at the same time, it’s a little annoying but easily addressed .

But your right in your assumptions, there’s a few ways it could be addressed.

For me, I just use sub control and don’t do anything with crossovers and it comes back just fine, the blending isn’t a issue at all , it’s just a little level, and it’s not much, I usually only have to turn it up a tiny bit...I’ve noticed in the morning when it’s cool, I can go back to where it was when I measured. It fluctuates a little that I’m certain

Unless I’m listening to a 1975 recording of “the cars” or something, than I crank it up a lot a lot, but that’s a different issue all together we already talked about.
So I don’t want to confuse anything.


I fixed this problem the first day I ever used DL ever in my life, lol
I was like, nope! I got to have my bass right.
At the time it seemed like a easy solution, but I do remember it threw me on one hell of a spin trying to figure out why, I just gave up and started tuning and measuring with the sub barley on and used sub control and haven’t looked back. It just worked
 

·
Registered
Joined
·
4,370 Posts
Discussion Starter #578 (Edited)
If someone is measuring the bass response they like with REW, (like Ian posted) and it still sounds weak ,
I might think maybe it’s this:;

I am just guessing here , but do you think maybe because DL is ignoring certain modal peaks it’s making the bass sound weak?

For instance if you have a room mode at 200 and it’s reverbant reinforcement because of room size, so you now have a little more energy there than you old flat (boring) rta with old (boring) dsp (lol sorry)

If that bump shouldn’t be taken out wouldn’t it make the bass sound low because to you the midbass sounds high? Ya know, the way your interpreting how loud the bass is compared to other frequencies. Maybe your thinking that because you now have extract mid-bass energy.

If your not using PEQ and forcing down the modal range (which adds or takes away energy at the wrong time in that range) the way Dirac does things would show your after to be flat, but the energy left alone in that range so the peak should still have some of that peak in your measurements, they should reflect that, not by tons and tons but if you measured with REW after like he did , that midbass peak that’s reverbant should still be somewhat still there.


And that also leads me to ponder this. Maybe DL is doing something in that so adding a little volume just fixes things. Because it’s a car. The bass in a car is different and so is the energy transfer. I can swear just turning the sub up a tiny bit after seems to fix it 100% for me.

I’m just thinking out loud, would love your thoughts on that


Edit: if you want to know which peak is reverbant grab a tone Gen, you’ll hear it .



Have you ever seen the videos where the guy puts sand on a sheet of metal and send vibrations through the metal, the sand goes to like 2 spots, than a oactave abovtthat the sand moves to 4 vibrating piles of sand instead of 2, than a octave above that it goes to 8 evenly spaced piles of vibrating sand.

Your reverbant room modes do the same thing. Some spots go high pressure evenly spaced throughout the car, the higher up in frequency more and more nodes appear..

No point to that......just a visual. That video is cool. It’s a guy that makes speakers out of cardboard boxes and flat things with tactile transducers
 

·
Registered
Joined
·
733 Posts
You know enough about this DSP to summarize this entire thread.


Sent from my iPhone using Tapatalk
 

·
Premium Member
Joined
·
3,563 Posts
Totally agree,

And if he can do overlays with smaart man that could help us,

I’ve only been able to compare single measurements of individual speakers and a measurement of two (or more) speakers in REW with the same mic location and same windowing methods etc.

It’s clear to me it’s not trying so much to “linearize” or make phase flat, as much as it’s trying to make the phase the same for bolth left and right , the entire transfer function. I’ve seen the phase still have some GD properties and some wraps from XOs , but left and Right were so so much alike vs with Dirac off.


With smaart he should really be able to see what’s going on every step of the way,


John please do! I’ll put my learning hat on and let you take it away.
Yeah at the frequencies he is at the wavelength is large.
So individual sub, and individual mid-bass phase plot(s) should tell if there is something happening to cause concern.

If one unplugs the midbass and sound gets louder in the cross over region, then that is a qualitative clue. But the quantitative plot is ideal.
 
561 - 580 of 928 Posts
Top