ProSoundWeb Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1] 2 3   Go Down

Author Topic: No FIR  (Read 6625 times)

Robert Lunceford

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 585
No FIR
« on: April 19, 2018, 05:27:02 PM »

Danley Sound Labs and JTR Pro Audio are both manufacturers who are held in high regard at this forum for the quality of sound of their speakers. Both of these manufacturers design and engineer speakers with passive crossovers.
It seems like the new craze is to process speakers with FIR filters.
What is the difference in performance between a speaker with a traditional passive crossover such as a Danley and a speaker with FIR filters incorporated into their design such as JBL and Fulcrum?
Are there differences in the implementation of the FIR filters between manufacturers (some better "engineered" than others)?
Does Danley use FIR filters in their own line of power amps to improve the performance of their speakers?
« Last Edit: April 19, 2018, 05:37:48 PM by Robert Lunceford »
Logged

Art Welter

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2203
  • Santa Fe, New Mexico
Re: No FIR
« Reply #1 on: April 19, 2018, 06:41:42 PM »

1)Danley Sound Labs and JTR Pro Audio are both manufacturers who are held in high regard at this forum for the quality of sound of their speakers. Both of these manufacturers design and engineer speakers with passive crossovers.
2)It seems like the new craze is to process speakers with FIR filters.
3)What is the difference in performance between a speaker with a traditional passive crossover such as a Danley and a speaker with FIR filters incorporated into their design such as JBL and Fulcrum?
4)Are there differences in the implementation of the FIR filters between manufacturers (some better "engineered" than others)?
5)Does Danley use FIR filters in their own line of power amps to improve the performance of their speakers?
Robert,

1)Many manufacturers use both passive crossovers and FIR filters.
2)Not very "new" anymore, but increased processing power has let it "filter down" to inexpensive price points. "Gunness Focusing" made it's way into a lot of EAW/LOUD products over the years.
3)As you can see in the 2009 measurements below of a Mackie HD1521 and a DSL SH100, the HD1521 has both flatter phase and frequency response.
Speakers with flat phase response "play nice" with each other, making a mix of mains, center fill and delay mostly a matter of time alignment.
Flatter (smoother) response allows more gain before feedback and less ear fatigue.
I was rather amazed at the time that the "cheap" HD1521 sounded so good with rather ordinary speaker components.
4) Yes, some engineers are much better than others. Implementation of filters needs to account for a lot of "moving parts", and long term design experience is helpful in determining which aspects of a design are most important. Every speaker design has it's own set of priorities (and compromises), and the filter sets are "correcting" different aspects of each design.
5) I don't know, but would guess not, unless they have recently changed.

Art
Logged

Ivan Beaver

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 9534
  • Atlanta GA
Re: No FIR
« Reply #2 on: April 19, 2018, 07:58:20 PM »

Everything comes with a price.

One of the "prices" with FIR is latency (signal delay)

Does it matter?  How long depends on how low the filters are used.

Part of the problem with trying to use FIR filters on cabinets that have physical offsets, is that the settings only work on axis (or where they were derived from)

As you move off axis to where the distance from the drivers is different than on axis, things start to change, different settings are required for different positions.

No, Danley does not currently use FIR filters in any of the processing.
Logged
A complex question is easily answered by a simple-easy to understand WRONG answer!

Ivan Beaver
Danley Sound Labs

PHYSICS- NOT FADS!

David Sturzenbecher

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1966
  • So. Dak.
    • Sturz Audio
Re: No FIR
« Reply #3 on: April 19, 2018, 08:43:19 PM »


Part of the problem with trying to use FIR filters on cabinets that have physical offsets, is that the settings only work on axis (or where they were derived from)


With FIRs filters you are able to realistically achieve very high crossover slopes.  In speakers with drivers physically offset you are able to greatly minimize frequencies of overlap. For this type of speaker, the benefits of FIR can be had in more then just one area.
Logged
Audio Systems Design Engineer
Daktronics, Inc.
CTS-D, CTS-I
AES Full Member

Michael Thompson

  • Full Member
  • ***
  • Offline Offline
  • Posts: 169
    • Factor Audio
Re: No FIR
« Reply #4 on: April 20, 2018, 12:48:48 AM »

Danley Sound Labs and JTR Pro Audio are both manufacturers who are held in high regard at this forum for the quality of sound of their speakers. Both of these manufacturers design and engineer speakers with passive crossovers.
It seems like the new craze is to process speakers with FIR filters.
What is the difference in performance between a speaker with a traditional passive crossover such as a Danley and a speaker with FIR filters incorporated into their design such as JBL and Fulcrum?
Are there differences in the implementation of the FIR filters between manufacturers (some better "engineered" than others)?
Does Danley use FIR filters in their own line of power amps to improve the performance of their speakers?

FIR is just another tool in the bag of speaker designers.  Choosing the right tool (passive, active, FIR) to accompany a particular design is the answer in my opinion.
Logged

Luke Geis

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2356
    • Owner of Endever Music Production's
Re: No FIR
« Reply #5 on: April 20, 2018, 01:28:38 AM »

I would argue that in two identical speakers in which one uses FIR and the other a conventional passive crossover, the one with FIR would measure better in terms of phase response. Which one actually sounded better would be a matter of opinion. In like for like settings, the theory would be that the one with FIR filters would sound more " correct " at any given measurement location. Better is subjective at this point.

The idea of FIR is that with phase being a multitude better than a conventional unit ( of identical construction ), it would also then be able to sound more the way it was intended ( within physics ) at any given location. This means that DSP dedicated to EQ is going to be more effective at sorting out problems directly related to the speakers components. Since DSP goes hand in hand with FIR filtering " IN THE BOX ", things like eq that is pre / post X-Over can be implemented to do processing on a component basis. In short, a speaker with the DSP for FIR, should and likely will sound better than an identical passive unit.
Logged
I don't understand how you can't hear yourself

Kevin McDonough

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 316
Re: No FIR
« Reply #6 on: April 20, 2018, 05:20:09 AM »

hey

yeah setting up speaker crossovers can be a very complicated process.   On the surface you just look at the frequency range that each speaker in your set up plays in, and then divide up the frequencies accordingly, setting you crossover points to what best suits each driver.

However when you start getting into the detail of it, there are many more factors to consider.  Speakers have different dispersion patters at different frequencies, so a big consideration is choosing crossover points where these mach up as well as possible too, so there are no big jumps in coverage as you change from one driver to another. Harmonic distortion plays a part, and even if a speaker can play a frequency, doesn't mean it sounds great, so choosing what frequencies within the speakers range suit the shape of the horn path, or it's piston range, can be considerations. Lots of other things also play a part.

And then phase plays a big part; at the crossover point there will be some overlap of the frequencies, where the two different drivers are playing the same frequencies for a small section. At this point you want everything to be in phase as much as possible to create a seamless transition, any crossover points or eq added in using traditional filters also comes with an accompanying phase change which makes this process more complicated.

Balancing all these factors up, making big, complicated filter networks that have loads of components and that use these eq and phase changes as an advantage and brings everything into line together takes time, skill and experimentation. It can be done, but it's far from trivial.  I don't know if Ivan or anyone is able to post some crossover pics without revealing too many trade secrets lol, but they're an art as much as a skill and electronic circuit.

FIR filters free you from a lot of this work. They allow you to separate the EQ and crossover points from phase so you can add them in without phase change, and they allow you to use much steeper slopes than would be possible otherwise, so there is little to no interaction between the adjacent drivers and you've much more freedom with the other factors of choosing your crossover point. With processing power being relatively cheap these days, companies are happy to to go this route and it makes it much easier to get good results than from traditional filters.

k
Logged

Peter Morris

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1467
Re: No FIR
« Reply #7 on: April 20, 2018, 10:21:42 AM »

Danley Sound Labs and JTR Pro Audio are both manufacturers who are held in high regard at this forum for the quality of sound of their speakers. Both of these manufacturers design and engineer speakers with passive crossovers.
It seems like the new craze is to process speakers with FIR filters.
What is the difference in performance between a speaker with a traditional passive crossover such as a Danley and a speaker with FIR filters incorporated into their design such as JBL and Fulcrum?
Are there differences in the implementation of the FIR filters between manufacturers (some better "engineered" than others)?
Does Danley use FIR filters in their own line of power amps to improve the performance of their speakers?

To be able to separate voices and instruments the same as you can when you listen without sound reinforcement, to sound real, you need to replicate exactly what you’re your ears were hearing or as close as possible to the real sound. That is a system with a flat amplitude response, a flat phase, low distortion, appropriate directivity and no time domain issues.
 
FIR filters allow you flatten the amplitude response and flatten the phase as much as you want, its just a matter of how much latency is acceptable.

By manipulating the FIR tap coefficients you can adjust phase and amplitude separately, you can produce very high crossover slopes (there can be ringing issues) or overlapping crossovers to manipulate directivity etc. i.e you have more power to correct things and you can get closer to the ideal system.


What many people don’t realise is that analogue/passive filters cause delay, but its frequency dependant. You can use a passive filter and then use a FIR filter to flatten its phase, but you more less end up at the same point in terms of latency as if you had just used a FIR crossover.

When you listen to a system and you get everything correct it will sound more real and you will have better separation of voices and instruments.

Here is an example of a FIR design/system with acceptable latency for small FOH applications, its one of my designs (its in the DIY section of Soundforums.net). This is the promo shot, in practice its set so the LF slops gently up. Its phase response is maintained on and off both the vertical and horizontal axis.

« Last Edit: April 20, 2018, 10:27:42 AM by Peter Morris »
Logged

Mark Wilkinson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1104
Re: No FIR
« Reply #8 on: April 20, 2018, 11:36:56 AM »


Lot of great comments above...hard to disagree with anything said...(not that my disagreement means anything :)

I think Danley's and JTR's use of passive crossovers reflects more on the businesses they are in. 
Danley appears to have primarily chosen the install market, and large scale at that ! 
If you were active, where would you want the DSP ?.... up in the sky inside monster boxes with weather and bird poop ?, or down in the warm protected control room ??
Plus, let's say you ditch the passive in-speaker x-overs, for DSP in the control room...
...there are a boatload of drivers in some of those boxes...what happens when some joe blow does exactly that....blows em up from messing with the DSP ?
And then, there's also the fact Danley has a bona fide genius at building passive xovers !

I get the sense that JTR's primary market is home theatre, where customers range from audiophools who 'wouldn't dream of impairing sonic purity with EQ' lol, to guys doing complete loudspeaker control with Dirac room correction or PC convolution...
...so passive suits a widely disparate market.

Just some market speculation....

Anyway, IMO, the biggest advantages for FIR when it comes to sound quality are:
* the ability to embed practically unlimited EQs
* linear phase crossovers
* the ability to use steeper crossovers than with passive circuitry
* the ability to independently alter magnitude and phase
* and as available with any DSP,  precise time alignment (as compared to passive)

The only real disadvantage IMO, is latency.
 
I've kinda come up with a rule of thumb for effective FIR that goes: latency = 125% of the period of the lowest frequency you want to adjust.
So for example, if you wanted to use FIR down to 500Hz, which has a period of 2ms, it will probably take about 2.5ms of FIR time.
This rule of thumb works for up to about 48dB oct slopes. Less slope gets away with less FIR time, greater slope of course needs more.

I mention this rule of thumb because it might help to evaluate what FIR can actually be doing, given published latency specs.....
Maybe someone will knock down my formula, happy to learn a better one, or that I'm full of it !

The ability to embed unlimited EQs is where I think sound quality can be improved on any box. 
I've learned that on a driver-by-driver basis, response variations that occur both on and off axis, are called minimum-phase, and can be corrected with traditional EQs (IIR).  These IIR eq's correct both magnitude and phase at the same time together, and absolutely lead to better sound quality IME.
A good traditional non-FIR processor can make these, but how many EQs are in such.
Again, this correction is on each driver, or sections of the same driver, before any xover.  Once put together via crossover, minimum-phase no longer holds sway.

Others have already spoken on the value of flat phase and steep x-overs, so I'll leave those alone...other than to say both have been giving outstanding results, with Peter's DIY 90/60, with my own designs, with reworks of kf650z, and with trying active FIR on a JTR 3tx.

edit... I'd like to add a +10 to Kevin's post...
It's hard to overstatement how much easier it is to align and tune, both driver to driver, and box to box, with FIR....
....or more particularly ... with flat phase that comes from the use of linear phase crossovers, and flattening out of band phase (along with the usual out of band EQ flattening)
Whether tying drivers together, or aligning subs to mains (if u can stand the latency)...plat phase ends makes it easy.
And when phase is flat, you can slide x-over up and down a bit, to easily find the best on vs off-axis tradeoffs. 

And at least to my ears, it truly does sound better  :)









« Last Edit: April 20, 2018, 01:18:22 PM by Mark Wilkinson »
Logged

Ivan Beaver

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 9534
  • Atlanta GA
Re: No FIR
« Reply #9 on: April 20, 2018, 01:14:34 PM »

  I don't know if Ivan or anyone is able to post some crossover pics without revealing too many trade secrets lol, but they're an art as much as a skill and electronic circuit.


Here is a photo of a 3 way SH50 (on the left) and a 2way 8" coax (on the right)
Logged
A complex question is easily answered by a simple-easy to understand WRONG answer!

Ivan Beaver
Danley Sound Labs

PHYSICS- NOT FADS!

ProSoundWeb Community

Re: No FIR
« Reply #9 on: April 20, 2018, 01:14:34 PM »


Pages: [1] 2 3   Go Up
 



Site Hosted By Ashdown Technologies, Inc.

Page created in 0.038 seconds with 24 queries.