DiyMobileAudio.com Car Stereo Forum banner
41 - 60 of 68 Posts

·
Registered
Joined
·
568 Posts
Maybe by realizing that the spreadsheet gives curves for the acoustical response and not electrical responses?
Why don’t you just use Bessel filters to achieve the acoustic response?
Of course I realize the spreadsheet curves are for the acoustical response. But the speaker curve text files are based on the LR mathematics, not Bessel.....

If I use Electrical Bessel XO and tune to a LR curve, why bother with that, the acoustical will be a LR response. Wouldn't I need Bessel curve files to achieve a Bessel acoustical response?
 
  • Like
Reactions: oabeieo

·
Registered
Joined
·
5,505 Posts
I watched the video, and I understand it on a conceptual level, but wouldn't know how to create speaker level curves in REW for a 3-way+Sub. I'm familiar with Jazzi Excel, but it has only BW's & LR's for XO types. I'd like to try BW between Mid-Range to Mid-bass, and try Bessel between Mid-Range and tweets, but how does one ensure to have -6db at the crossover?

I'm high-jacking this thread with this question, sorry about that to the OP. @oabeieo would you consider starting a new thread ?
yeah I can do that….. I’ll start working on it tonite
I’m really bad about doing screenshots from REW and rephase as it’s very time consuming. I have hundreds of thousands of measurements and sims I can post…. I’ll try and find some good examples and try and make sense of it with everyone….

filters are fun…. (That seems like a good thread title 🤪🤪)
 

·
Registered
Joined
·
568 Posts
I added/edited this statement to my post, not sure if you seen it....


"If I use Electrical Bessel XO and tune to a LR curve, why bother with that, the acoustical will be a LR response. Wouldn't I need Bessel curve files to achieve a Bessel acoustical response? "
 

·
Registered
Joined
·
568 Posts
yeah I can do that….. I’ll start working on it tonite
I’m really bad about doing screenshots from REW and rephase as it’s very time consuming. I have hundreds of thousands of measurements and sims I can post…. I’ll try and find some good examples and try and make sense of it with everyone….

filters are fun…. (That seems like a good thread title 🤪🤪)

Cool Cool Cool !! Looking forward to this !!
 

·
Registered
Joined
·
409 Posts
I added/edited this statement to my post, not sure if you seen it....


"If I use Electrical Bessel XO and tune to a LR curve, why bother with that, the acoustical will be a LR response. Wouldn't I need Bessel curve files to achieve a Bessel acoustical response? "
No, haven’t seen that.
And no, you want the acoustical LR24dB response for its summation and acoustical phase properties, while benefitting from the electrical (and sometimes acoustical) advantages of a Butterworth or Bessel filter.
Just try it.
 

·
Registered
Joined
·
5,505 Posts
No, haven’t seen that.
And no, you want the acoustical LR24dB response for its summation and acoustical phase properties, while benefitting from the electrical (and sometimes acoustical) advantages of a Butterworth or Bessel filter.
Just try it.
sind Sie deutsch?
Ich schätze deine Intelligenz, mein Name ist Andy.... PN an mich jederzeit
 

·
Registered
Joined
·
568 Posts
do a band pass crossover in LR and look at it…. It looks like a hump….. do it in BW and it looks more flat top…. That’s maximally flat

The hump shape causes some problems in summing because the entirety of the speaker is summing dependent except the very middle

it just so happens to be the very middle that has a phase twist from it being a band pass!

so no part of the response ever has time to work itself out… it’s constantly changing and summing
I'm looking to achieve ^^this^^. Looking for the 'maximally flat' response...

By mixing different XO types and tuning down to a LR curve, well, will result in a LR hump..... Right? Am I missing something...
 

·
Registered
Joined
·
5,505 Posts
I'm looking to achieve ^^this^^. Looking for the 'maximally flat' response...

By mixing different XO types and tuning down to a LR curve, well, will result in a LR hump..... Right? Am I missing something...
there’s a lot more to it than just that…

ideally you want everything to behave all pass as far a crossover, and you’re looking for a very fast step response with no ringing. Pre or post

but there’s always compromises… having something measure good and sound good are two different things

a flat crossover and flat summation is also two different critters…. But I think I get what your sayin
 

·
Registered
Joined
·
409 Posts
By mixing different XO types and tuning down to a LR curve, well, will result in a LR hump..... Right? Am I missing something...
Yes, LR24db will sum flat, not to a bump. Acoustical phase is different from electrical phase.

You still aren't getting the concept of "what you measure is not what you put in into the DSP electrically".
What everyone agrees on (including Frank Miketta, the developer of the Mosconi DSPs btw.) is that you want a 24dB Linkwitz-Riley acoustical alignment of the crossover slopes.
What you use in the DSP to achieve this is totally independant from that, i.e. you don't have to set a 24dB LR electrical crossover to achieve an acoustical LR24dB alignment.
You might want to use Bessel filters especially for tweeters and midrange lowpass for it's electrical benefits like better impulse response, or a Butterworth filter, because it sounds better to you or whatever reason there might be.
Electrical 24dB LR filters only work like desired, when the frequency response in the area you want to put that filter to is totally flat. In case of bandpassed drivers like a midbass or midrange having two LR24dB filters electrically might shift the phase enough that you don't have the perfect phase summation electrically that everyone is talking about as every crossover changes phase naturally. If the bandwith of the driver that you put a high- and lowpass on is near enough you have interaction of the phase rotation of the low- and the highpass, but this is also true for every other crossover filter.

So use whatever electrical filter you want to achieve the acoustical response you want to have. And don't let yourself hold back from changing filter types, just because someone says "use LR24dB", because other filters might sound better to you, which you don't know if you don't try them. ;)
 

·
Registered
Joined
·
3,500 Posts
@cathul is right gang….

Listen , crossovers will do there function no matter what as long as acoustically there


see that’s where your wrong….. even order butts sum to 0db …. (Stop hitting yourself, download rephase and run the sim yourself and see) EVEN ORDER BUTTS SUM TO zero decibels

listen closely, it doesn’t matter if the crossover puts everything at +/-90 , the phase shift by itself isn’t a big deal!!! It’s when two or more things are summing when even a small phase shift can wreak havoc

it’s only in low frequencies where odd butts can be problematic… where the phase is sensitive. A butt 12 on a tweet to mid works fine. And sums fine given the proper implementation… the trade offs (extra emphasis) trade offs can be worth it

Overall , the crossover as a whole system matters, a crossover is only textbook when it’s a two way complimentary… as soon as you band pass one the phase changes dramatically throughout.. it can cause serious summing and phase issues midband as well

I’m not going to smack myself in the face, I’m not a dumdum (jk) but man,please run some sims yourself…… watch the gladen video that was just released.. he goes into some of it

so….. it’s only “rubbish” if you don’t fully understand….. because I was also in the only LR4 gang for a long time after reading countless erroneous pages online…. LR4 is a good crossover, but it’s not the only choice and definitely absolutely not the only way….

you also just made yourself look like a rook
Step it up and do your homework before you mouth off like that please
last time I checked 1 was an odd number and a 6db crossover is 1st order

unless I’m missing something??

but I disagree with your contention that 270 degrees of phase rotation is ‘back in phase’ it’s not… 360 degrees of rotation is back in phase 🤷🏼‍♂️ That’s why 24db lr are considered the go to, because they sum perfectly
 

·
Registered
Joined
·
5,505 Posts
last time I checked 1 was an odd number and a 6db crossover is 1st order

unless I’m missing something??

but I disagree with your contention that 270 degrees of phase rotation is ‘back in phase’ it’s not… 360 degrees of rotation is back in phase 🤷🏼‍♂️ That’s why 24db lr are considered the go to, because they sum perfectly
sorry odd …. I was not thinking….

so, a BW will sum perfectly based off its mathematical fourmulation for summing, the odd butts sum to 0db and even butts sum to 3db

so let’s look at a odd butt (a true quadratic) because even butts are multiples of 180 just like LRs

the odd butt is 90 out +/- meaning the hi pass is 90 deg and the low pass is -90 or 270

first off based on that the fact that they run their magnitude 90°…. so what!!!! Seriously what difference does it make if it’s at zero or 90°?

so long as all of the frequencies are equally offset it’s basically the same thing is hearing the difference between all of your speakers in your car with 0 ms of delay entered into the time alignment, and ALL Of the speakers in the car with 1 ms of delay entered into the time alignment

it delays all frequencies by 1 ms, but there is literally no audible difference… so whatgives

The LR4 is one cycle (360°) of whatever frequencie it’s -180 (-6db) point. So the high pass is one cycle of group delay from the low pass , and the sum is an allpass 2nd order Q1.2 (or abouts)

so what…… that doesn’t make it better… it makes it easier to implement and easier to have good summation and most importantly, easier to implement in a network. So better , no!!! Not better!!! Just another good option in the toolbox

now let’s say you have a highly reflective environment, and a LR and you are a 0deg and you have a reflection that is -180….. guess what

If your quadratic, and you have a reflection that’s -180 guess what…..

But it’s more then just reflections, it’s the matter of speakers summing together where some are mounted on top of the dash and the bottom of the floor or door…. And the transfer function each location exhibits… where sometimes a 90° offset would sure make things work together a lot better…..

So, again, I see your stuck on some fad trend idea that has been floating around this place for way too long. It’s simply not true that LRs are better…. Just a very good tool amongst many….

So which you probably won’t ever understand, these things make it vital to have the ability to cascade filters to have a properly ordered crossover network.

this is the single reason why the mosconi dsps are superior as they allow cascaded filters

so that way when you use a quadratic between your mid range in midbass
You can add the quadratic high pass as a cascade onto your tweeter that may is LRs , then it will subsequently inherit the phase shift of the mid range and will still sum properly and won’t be 90 out (that is all theoretically because in the real world of measurements sometimes you don’t need to do that depending on the behavior of the room and the speakers that are to be interacting with each other

regardless cascading filters is really crucial to have a linear phase response using minimum phase filters

an out of band all pass can effectively do the same thing as well….. all tools in the toolbox friend
 

·
Registered
Joined
·
3,500 Posts
sorry odd …. I was not thinking….

so, a BW will sum perfectly based off its mathematical fourmulation for summing, the odd butts sum to 0db and even butts sum to 3db

so let’s look at a odd butt (a true quadratic) because even butts are multiples of 180 just like LRs

the odd butt is 90 out +/- meaning the hi pass is 90 deg and the low pass is -90 or 270

first off based on that the fact that they run their magnitude 90°…. so what!!!! Seriously what difference does it make if it’s at zero or 90°?

so long as all of the frequencies are equally offset it’s basically the same thing is hearing the difference between all of your speakers in your car with 0 ms of delay entered into the time alignment, and ALL Of the speakers in the car with 1 ms of delay entered into the time alignment

it delays all frequencies by 1 ms, but there is literally no audible difference… so whatgives

The LR4 is one cycle (360°) of whatever frequencie it’s -180 (-6db) point. So the high pass is one cycle of group delay from the low pass , and the sum is an allpass 2nd order Q1.2 (or abouts)

so what…… that doesn’t make it better… it makes it easier to implement and easier to have good summation and most importantly, easier to implement in a network. So better , no!!! Not better!!! Just another good option in the toolbox

now let’s say you have a highly reflective environment, and a LR and you are a 0deg and you have a reflection that is -180….. guess what

If your quadratic, and you have a reflection that’s -180 guess what…..

But it’s more then just reflections, it’s the matter of speakers summing together where some are mounted on top of the dash and the bottom of the floor or door…. And the transfer function each location exhibits… where sometimes a 90° offset would sure make things work together a lot better…..

So, again, I see your stuck on some fad trend idea that has been floating around this place for way too long. It’s simply not true that LRs are better…. Just a very good tool amongst many….
I didn’t say lr are better, I said they are a good target for new guys… slight difference… use whatever I need to get to the measured phase and response

and 90 degrees out isn’t in phase… which you stated… hardly a stretch to see how that works out

had you said they will sum correctly if the reflections cause one side to be 90 degrees out, but again these are conditions you didn’t mention…

If I have reflections that make energy out of phase I don’t ever put a crossover there, again that’s not rocket science

So you’d end up 90 degrees out of phase if I did what you’re proposing and placed crossovers correctly in the first place 🤷🏼‍♂️
 

·
Registered
Joined
·
5,505 Posts
I didn’t say lr are better, I said they are a good target for new guys… slight difference… use whatever I need to get to the measured phase and response

and 90 degrees out isn’t in phase… which you stated… hardly a stretch to see how that works out

had you said they will sum correctly if the reflections cause one side to be 90 degrees out, but again these are conditions you didn’t mention…

If I have reflections that make energy out of phase I don’t ever put a crossover there, again that’s not rocket science

So you’d end up 90 degrees out of phase if I did what you’re proposing and placed crossovers correctly in the first place 🤷🏼‍♂️
i didn’t see you say anything about for noobs , but yeah I’ll pass u on that as you might have said it earlier and I didn’t see it

no point in throwing your arms up icon, I think your still misunderstand… and that’s ok it’s kinda hard to grasp unless you hear it, see it sim’ed. etc etc

I’m just tryin to help u man, I’m sure I’m not 100% right , but at least 90% ish would be my guess as it’s hard to explain this stuff and the words can be taken out of context so easily if one party isn’t on same page…. I’m tryin tho…..

So , hear me please ….. please please hear me on this ….. phase rotation is 100% irrelevant by itself, it’s only when it’s summing with something that it matters…. Otherwise it’s just pure delay which has no sonic difference by itself….

if something is 90 degrees it’s only “out of phase” if it’s playing against something that isn’t 90deg…. Then it’s a matter of choosing what’s “out”. The degrees can almost be an imaginary line… almost in some respect to certain things…. But it has to be what it is because it’s logmerithic not linear in nature
 

·
Registered
Joined
·
3,500 Posts
i didn’t see you say anything about for noobs , but yeah I’ll pass u on that as you might have said it earlier and I didn’t see it

no point in throwing your arms up icon, I think your still misunderstand… and that’s ok it’s kinda hard to grasp unless you hear it, see it sim’ed. etc etc

I’m just tryin to help u man, I’m sure I’m not 100% right , but at least 90% ish would be my guess as it’s hard to explain this stuff and the words can be taken out of context so easily if one party isn’t on same page…. I’m tryin tho…..

So , hear me please ….. please please hear me on this ….. phase rotation is 100% irrelevant by itself, it’s only when it’s summing with something that it matters…. Otherwise it’s just pure delay which has no sonic difference by itself….

if something is 90 degrees it’s only “out of phase” if it’s playing against something that isn’t 90deg…. Then it’s a matter of choosing what’s “out”. The degrees can almost be an imaginary line… almost in some respect to certain things…. But it has to be what it is because it’s logmerithic not linear in nature
Ok, your right I’m nuts, thanks for the help, I have a great understanding of summation and how it works… you on the other hand have great understanding of how to confuse the crap out of people with waffling and babble…

I don’t need to guess or workout what is out, I measure with smaart and can work it out just fine…

so what your saying is that two butterworth 12db crossovers can be in phase of one side of the crossover is 180degrees out of phase due to something external without swapping the phase of one driver or another… so in other words if you pick a bad place to put a crossover you can correct it by using a non summing crossover, I get that and it’s not the case, but why not just move the crossover to a place where you don’t have reflections and two lots of phase information (direct and reflected that creates the 90 degree phase shift effectively) and so you can never make a Crossover correctly there anyway
 

·
Registered
Joined
·
5,505 Posts
Ok, your right I’m nuts, thanks for the help, I have a great understanding of summation and how it works… you on the other hand have great understanding of how to confuse the crap out of people with waffling and babble…

I don’t need to guess or workout what is out, I measure with smaart and can work it out just fine…
everytime I argue with my wife and she starts name calling, that’s when I know I’m right.

waffles are good btwwith maple syrup and babbleing is when someone doesn’t understand the language….. that’s why the most popular launguage learning is named babble …. I’m speaking clearly misspelled words lol

i guess all I can say is I’m proud of you that you have a good handle on everything. Fantastic, good for you.
 

·
Registered
Joined
·
5,505 Posts
Ok, your right I’m nuts, thanks for the help, I have a great understanding of summation and how it works… you on the other hand have great understanding of how to confuse the crap out of people with waffling and babble…

I don’t need to guess or workout what is out, I measure with smaart and can work it out just fine…

so what your saying is that two butterworth 12db crossovers can be in phase of one side of the crossover is 180degrees out of phase due to something external without swapping the phase of one driver or another… so in other words if you pick a bad place to put a crossover you can correct it by using a non summing crossover, I get that and it’s not the case, but why not just move the crossover to a place where you don’t have reflections and two lots of phase information (direct and reflected that creates the 90 degree phase shift effectively) and so you can never make a Crossover correctly there anyway
so also the 2nd part of your statement

if you have two butt12s , yes you reverse the polarity of the low pass…. Done…. Regardless of reflections

But what I am saying is if you have a complementary butt 6,18,36 and you have a reflection and it comes back and try’s to sum with your direct sound…. And that reflection is -180 of the direct sound, it won’t be a cancellation , rather a reinforcement….
 

·
Registered
Joined
·
5,505 Posts
Even order butts the high pass comes out first…. But the BW actually comes out ahead of 0°

that does not mean it’s not casual, it’s just the way that they represent themselves. could you call it 270 ? absolutely.

it can’t see ahead of time in the future it’s just the nature of that type of crossover
 

·
Registered
Joined
·
409 Posts
so what your saying is that two butterworth 12db crossovers can be in phase of one side of the crossover is 180degrees out of phase due to something external without swapping the phase of one driver or another… so in other words if you pick a bad place to put a crossover you can correct it by using a non summing crossover, I get that and it’s not the case, but why not just move the crossover to a place where you don’t have reflections and two lots of phase information (direct and reflected that creates the 90 degree phase shift effectively) and so you can never make a Crossover correctly there anyway
In one of my many measuring sessions i experienced something like this.

Using 12db Butterworth electrical filter on both high- and lowpass, but had an acoustical LR24dB alignment.
I didn't need to swap polarity on any of the drivers. They just summed up. That's the beauty of acoustical alignment. As long as you have that acoustical LR 24dB alignment everything falls into place without thinking about electrical filters. At least that's my experience.
 
41 - 60 of 68 Posts
Top