Welcome, Guest
Username: Password: Remember me
This is the optional Forum header for the Suggestion Box.

TOPIC:

Re: Linkwitz transform 10 years 5 months ago #4215

  • devteam
  • devteam's Avatar
  • Away
  • Administrator
  • Administrator
  • Posts: 9049
  • Thank you received: 1376
Interesting stuff going on here! Thanks to Charlie for sharing his great work and I think that a lot of miniDSP users will be interested in hearing more of your progress.

Should we start a dedicated thread to your good work Charlie? Maybe easier for users to find the information and we'll gladly put it as a sticky being a topic that I'm sure our community will enjoy.

Happy holidays to all,

DevTeam
miniDSP, building a DSP community one board at a time.

Please Log in or Create an account to join the conversation.

Re: Linkwitz transform 10 years 1 month ago #5183

  • DjSinae
  • DjSinae's Avatar
  • Visitor
  • Visitor
Anybody implemented this in minidsp

S. Lipshitz and J. Vanderkooy linear phase crossover

Please Log in or Create an account to join the conversation.

Last edit: by DjSinae.

Re: Linkwitz transform 10 years 1 month ago #5227

  • dreite
  • dreite's Avatar
  • Offline
  • Platinum Member
  • Platinum Member
  • Posts: 1551
  • Thank you received: 702
It's stretching quite a bit for the authors/designers to call those linear-phase crossovers. :)

If the question is, can you create a linear-phase crossover with the miniDSP? The answer is "yes", but only in the case of a true first-order acoustic crossover. Anything else doesn't satisfy the linear-phase requirement of constant group-delay and a perfectly flat phase response with the outputs summed together.

This question is off-topic for this thread anyways.

Cheers,

Dave.

Please Log in or Create an account to join the conversation.

Linkwitz transform 7 years 4 days ago #19457

  • mr-marlen
  • mr-marlen's Avatar
  • Offline
  • Senior Member
  • Senior Member
  • Posts: 50
  • Thank you received: 7
Hello, Dreite!
I think You are a little wrong about linear phase crossover. You can make it also with 2nd order.
You can see such filter based on real speaker on picture in this post. In comparison to standart LR12 (etc filters) we have peaks near crossover regions on vertical axis instead of dips. But maybe it is better, who knows :). Also we need speakers with more bandwidth and lower distortions.
Don't ask how it sound. I still didn't listen to this.

cloud.mail.ru/public/2gij7rLmvhFZ/lin%20phase.jpg
Attachments:

Please Log in or Create an account to join the conversation.

Last edit: by mr-marlen.

Linkwitz transform 7 years 3 days ago #19469

  • dreite
  • dreite's Avatar
  • Offline
  • Platinum Member
  • Platinum Member
  • Posts: 1551
  • Thank you received: 702
Well, it's a topic of debate.....depending upon the conditions we define for the requirement.

A person could make a good case that linear-phase operation is only possible with headphones....and not even all of those. :)

Your example doesn't show linear-phase response. Multiple phase wraps which appear to be time-of-flight or some other pure delay built into your measurement. If you back that out it appears your phase response might be fairly flat.

Dave.

Please Log in or Create an account to join the conversation.

Last edit: by dreite.

Linkwitz transform 7 years 3 days ago #19477

  • mr-marlen
  • mr-marlen's Avatar
  • Offline
  • Senior Member
  • Senior Member
  • Posts: 50
  • Thank you received: 7

Well, it's a topic of debate.....depending upon the conditions we define for the requirement.

A person could make a good case that linear-phase operation is only possible with headphones....and not even all of those. :)

Your example doesn't show linear-phase response. Multiple phase wraps which appear to be time-of-flight or some other pure delay built into your measurement. If you back that out it appears your phase response might be fairly flat.

Dave.

There is Impulse response in my measurement, so this wraps only due to distance delay. I added here a picture of 500Hz square (filter on tweeter was a little different). And ofc. lower 200hz it will be closed box with its delay. But what about subwoofer with Linkwitz correction and such filters as mine on upper side? There will flat Group Delay for very low freq. without much delay 40-1000 ms if we compare this type of IIR filters to FIR filters with high slopes.
And another interesting trick - WMTMW. Side drivers M - M; W - W; will have dips below axis on vertical plot. Such thing will compensate amplitude rise and give us almost flat vertical plot.
Attachments:

Please Log in or Create an account to join the conversation.

Last edit: by mr-marlen.

Linkwitz transform 7 years 3 days ago #19486

  • dreite
  • dreite's Avatar
  • Offline
  • Platinum Member
  • Platinum Member
  • Posts: 1551
  • Thank you received: 702
Yeah, I'm with you and will stipulate that if you measure a system designed with this objective, in free-field conditions, and with the microphone in a precise location you can achieve a flat phase response. But it's going to break down off-axis and when other variables.....like a reverberant listening room are introduced. :)

I view this as primarily an academic exercise. In my opinion there are other, much more important issues, that need attention in speaker system design before worrying about "linear-phase" reproduction capability.

Regardless, this is still way off topic for this thread in this area of the miniDSP forum. There are hundreds of various threads regarding this topic on various audio forums. :)

Dave.

Please Log in or Create an account to join the conversation.

Linkwitz transform 6 years 11 months ago #20003

  • swest
  • swest's Avatar
  • Offline
  • New Member
  • New Member
  • Posts: 1
  • Thank you received: 0
Greetings,

First off, I will confess to having not read this entire thread. I have searched it, and not gotten any hits for the keywords of my question.

Some background, and my question, follows:

I currently use a Marchand Electronics BASSIS (a variable, analog, Linkwitz Transform device) to re-Q, and boost, my subwoofer array, which is 4 sealed Dayton Titanic MK IIIs (F3 = ~46hz, Q =~.707...)

This works quite well. Unfortunately, mine is a single-channel BASSIS, and my new Marantz processor will accommodate 2 subwoofers (and Audyssey will EQ them separately)... So, for best results, I am going to need another BASSIS, or a 2-channel BASSIS to replace my single channel version.

Q: Can the miniDSP replace the BASSIS? Both in terms of the functionality, as well as the 'quality' of the resulting Frequency Response curve.

Thanks.

- s.west

p.s. I haven't checked the forum rules, regarding the posting of external URLs, but to be on the safe side, I didn't post the link to the BASSIS. However, it can be easily found via a google search, if need be.

Please Log in or Create an account to join the conversation.

Linkwitz transform 6 years 11 months ago #20006

  • dreite
  • dreite's Avatar
  • Offline
  • Platinum Member
  • Platinum Member
  • Posts: 1551
  • Thank you received: 702
Just about any DSP-based equalizer on the market nowadays can perform the function of the Bassis......and a whole bunch more. :)
You won't have the continuously adjustable settings, but any electrical response can be duplicated with a miniDSP unit.

However, if Audyssey is in use for EQ, why do you need a BASSIS in the first place?

Dave.

Please Log in or Create an account to join the conversation.

Linkwitz transform 4 years 2 months ago #36139

  • JJ2106
  • JJ2106's Avatar
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 106
  • Thank you received: 8
Sorry again to dig this (very) old thread.
For those who want to dispense with the advanced biquad functionality, I just wanted to mention that the strict equivalent of a Linkwitz Transform can be achieved with two standard filters: a peaking filter and a lo-shelf filter.
These two can be applied in whichever order, but assuming you apply the peaking filter first, here are the parameters:
- peaking: Fpk=Fb, Gain(v/v) = Qtarget/Qb, Qpk=sqrt(Qb*Qtarget)
- lo-shelf: Fls =sqrt (Fb*Ftarget), Midpoint gain (v/v=Fb/Ftarget, Qls=Qtarget (DCgain=2*Midpointgain)
Hope this may be useful, albeit to one member only!
J-J
The following user(s) said Thank You: Juoigâ

Please Log in or Create an account to join the conversation.

Linkwitz transform 1 year 9 months ago #50001

  • keselobojo
  • keselobojo's Avatar
  • Offline
  • New Member
  • New Member
  • Posts: 6
  • Thank you received: 0
I guess everything is solved now?

Please Log in or Create an account to join the conversation.

Moderators: devteam