What is a FIR filter?

[Chris A]

I was typing all this when Edgar replied...

 

3 hours ago, Schu said:

can I give this a try for around $300<?

I think that your risk of not achieving your goals is quite high at that low price point.  I'd increase it to at $500 (used) to $1000 (new), which is probably still less than the cost of your DAC or amplifier.  You could try this, but the risk is high: https://www.minidsp.com/products/minidsp-in-a-box/minidsp-2x4-hd,  I'd actually recommend something more like this if you're going to try miniDSP: https://www.minidsp.com/products/minidsp-in-a-box/minidsp-4x10-hd. and tri-amp.

 

If you're into a good-to-excellent solution, then I'd recommend XIlica XP2040 (bi-amping) or XP-4080 (tri-amping with ability to add your subs). 

 

3 hours ago, Schu said:

I assume if I am bi-amping a three way that I am still using the factory Passive XO for the HF section?

You could do this, but I'd strongly recommend moving the tweeter to the top rear portion of the cabinet.  In fact, I'd recommend that in any case since it costs almost nothing, and you'll get an 80% solution for time alignment doing only that.

 

3 hours ago, Schu said:

Will I be able to play with driver alignment at that price point?

Using a digital crossover, I'd recommend 2-in, 6-out and tri-amp for the best performance.  If you're only willing to bi-amp, then aligning the tweeter on top of the La Scala is pretty much required in order to hear the full benefits.  As with Edgar, you need to split out the woofer from the midrange/tweeter if bi-amping.  Otherwise, you'll be dealing with woofer-induced back-EMF noise bleeding into your midrange frequencies.

 

Chris

[Chris A]

By the way, you could use the original active crossover settings in the Belle tri-amping thread as a starting point...but decrease the mid delay by about 0.5 ms (i.e., longer midrange horn in the La Scala) and increase the tweeter delay by about 0.5 ms (same thing, due to the longer midrange horn), and cross over at 400 Hz instead of 600 Hz.  This will get you fairly close.

 

I found that there isn't any real EQ required on the K-55 using the K-600 horn, and I assume that's also the case with the K-401. 

 

Chris

[Schu]

okay... let me digest this a bit. I can not deny that I am very intrigued because listening critically is really my goal.

[Coytee]

My head is about to explode. 

 

1.  The FIR stuff is still too foreign to me (for me to think I even have an inkling of an understanding)

2.  I'm willing to bank on what you (Chris) has to say....so thank you.  I'm likely to just get the non-FIR version

[Chris A]

More info on this subject:

 

[Chris A]

A little tutorial on FIR filters compared to IIR filters which is pretty good:

 

https://web.archive.org/web/20140423022411/http://www.prosoundweb.com/article/fir-ward_thinking_examining_finite_impulse_response_filtering_in_sound_rein/av

 

Chris

[mikebse2a3]

I know this is an older thread but I’ve been recently researching FIR filters and thought this was an interesting observation.

 

https://www.dsprelated.com/freebooks/filters/Linear_Phase_Really_Ideal.html

[Chris A]

Mike,

 

When I get the necessary hardware, I plan to do the FIR filtering trick on the Jubs, K-402-MEH, and/or AMT-1s with Belle bass bins (i.e., I'm pretty sure that the miniDSP 2x4 HD I have doesn't have nearly enough taps to flatten the phase outside of the crossover interference band--so I'm going to need JRiver or something like that with a dedicated PC to do the convolution engine duties).  I don't expect that the audible difference is going to be very much when I try FIR filters, as discussed in the article that you linked just above.  I think that minimum phase is probably much closer to where you want to be--as posted in the "Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers" thread.  When you're within 45 degrees of flat phase (i.e., zero relative phase across the listening band--not linear phase) and group delay below 1 ms everywhere (psychoacoustic smoothing) above the Schroeder frequency of the room--nominally 150--200 Hz, I think that you're already well within the ear's discrimination of phase or group delay distortion threshold.  This is just like being within 90 degrees of correct time alignment at the crossover interference band.  I think that people are going overboard with the linear phase requirement--more than the ear can discriminate.

Chris

[Chris A]

One more comment:

 

In order to hear the linear phase performance of loudspeakers, you need full-range directivity of the loudspeaker--not just above 800 Hz like the hybrid horn-woofer loudspeakers (Cornwall, Forte, etc.).  Think about it--the real difference in sound quality occurred in my listening room when I corrected the phase and group delay below 800 Hz.  Hearing it is highly dependent on the elimination of early reflections from around the loudspeaker in-room in order to hear that linear phase without it being obliterated by early reflections in-room.

 

If you notice, every loudspeaker in my setup has full-range directivity to below the Schroeder frequency of the room (which is measured at ~100 Hz in my room).  I actually changed out the Cornwall bass bins in favor of Belle bass bins in order to get that directivity that I don't have with direct radiating woofers from the Cornwalls, thus losing about 1/2 octave of bass extension in the surrounds in order to achieve it. 

 

Chris

[Kalifornian]

As someone who started at the cheap end (minidsp) and then a little better (dbx), and then FINALLY ending with the Xilica 4080 do yourself a favor and save up and get the xilica. By the time I got to the Xilica I had wasted the difference in cost in upgrading ..and the Xilica rocks.  I am biamping jbl 2446h on large format jbl horns with dual 18 " jbl lowers and with REW and xilica it only took me about a week of playing with it to get it figured out and the results are phenomenal, i'll never go passive again unless forced to.

[mark1101]

Richard,

 

Here's my advice to add to all this stuff.  Sorry for any common sense detected here.  

 

Stick to equipment that allows you to use Roy's settings on your Jubs.  Remember that?  It is a hard solution to beat.  And with all we read..........what can you truly trust to be better?

 

Consider replacing your K69s with 691s.............consider getting either a DC-One or Xilica XP.........which both support Roy's settings for the 691s.  That will give you the newest Klipsch drivers and newest settings with least cost..............and a guaranteed immediate success.

 

Next I would look at some new electronics to continue when appropriate.  A whole other subject for another thread.

 

Finally, if you want a FIR get a FIR X-mas tree.

[mikebse2a3]

5 hours ago, Chris A said:

Mike,

 

When I get the necessary hardware, I plan to do the FIR filtering trick on the Jubs, K-402-MEH, and/or AMT-1s with Belle bass bins (i.e., I'm pretty sure that the miniDSP 2x4 HD I have doesn't have nearly enough taps to flatten the phase outside of the crossover interference band--so I'm going to need JRiver or something like that with a dedicated PC to do the convolution engine duties).  I don't expect that the audible difference is going to be very much when I try FIR filters, as discussed in the article that you linked just above.  I think that minimum phase is probably much closer to where you want to be--as posted in the "Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers" thread.  When you're within 45 degrees of flat phase (i.e., zero relative phase across the listening band--not linear phase) and group delay below 1 ms everywhere (psychoacoustic smoothing) above the Schroeder frequency of the room--nominally 150--200 Hz, I think that you're already well within the ear's discrimination of phase or group delay distortion threshold.  This is just like being within 90 degrees of correct time alignment at the crossover interference band.  I think that people are going overboard with the linear phase requirement--more than the ear can discriminate.

Chris

 

I've been looking at the DEQX approach and would love to have one to experiment with (like I've got the time ) to see what is audible and preferable.

 

Chris if you have the time can you explain/describe (or if you already have please give me a link) how you are limiting the excursion of the TAD4002 with the method your using now.?

 

Thanks,

Mike

[Chris A]

9 hours ago, mikebse2a3 said:

Chris if you have the time can you explain/describe (or if you already have please give me a link) how you are limiting the excursion of the TAD4002 with the method your using now.?

See the plot below to understand the magnitude of the fractional order crossover approach that is achieved using the method I describe.  Reading the plot below...what is the effective order of crossovers that's achieved (in dB/octave), and the effective crossover frequency?

 

 

Note that this is not using high pass or low pass filters, but rather the natural response of the drivers/horns themselves--and attenuating PEQs. The drivers/horns themselves tell you where the crossover point should be, so no boosting of the driver/horn response is necessary around the crossover frequency.  This is an important point.  It is the same point that Tom Danley made in his guidance on this subject, and on the way that the SH-50 crossover/balancing network actually works to achieve 127 dBSPL cont., 133 dBSPL peak at 1m (with safety factor). 

 

Chris

[Chris A]

Here is the payoff:  the phase response of the entire loudspeaker:

 

 

Chris

[mikebse2a3]

2 hours ago, Chris A said:

Note that this is not using high pass or low pass filters, but rather the natural response of the drivers/horns themselves--and attenuating PEQs.

 

Thanks Chris

 

I would like to understand the relationship between the driver’s natural roll off due to excursion limitations, back chamber compression and the PEQs so that I can better understand what the compromises versus benefits are with this method. I’m especially concerned about the excursion limitations with this method.

 

Can you share the PEQs your using with the K402/TAD in conjunction with the driver’s natural rolloff to achieve these results..?

 

Thanks,

Mike

[Chris A]

I tell you what--I've given you enough information so that you can try it, too.  Knowing your technical capabilities over the 12 years that we've known each other, I have no doubt that you too can achieve similar results in your room with your Jubilees.  Of course, you may not be able to lower your K-402 quite as much as mine are currently due to your after-sale customization of the KPT-402-HF assembly and the custom bass bin fronts...but I think that you can get close enough to hear the (rather startling) difference in sound quality while listening to your best orchestral recordings--for instance. 

 

Why don't we then compare notes after you try it, you showing the same data in like format that I've generated above and whatever other data you need to solve your conundrum? I believe that only you are going to be able to generate the data that you need to convince yourself that your worst fears are either founded or unfounded (as you have done before on another subject).  I don't believe that any data that I have will show you what you want beyond what I've already posted here...and the understanding of the basic physics of the setup.  I've tried it and it works for me.  It also works for a handful of other Jub owners (and those with other loudspeaker types) whom I've helped online since May 1 dial-in their loudspeakers, as well as Danley Sound Labs--with none of them reporting any issues.  Instead, all of them report that they are delighted with the sound quality. 

 

Chris

[mikebse2a3]

Chris I’m sorry if I’ve inconvenienced you and I didn’t realize my follow up questions would be a problem and I’ve meant no disrespect for what you have been doing.

 

I wish I had more time to explore this and many other things on my mind but life has limited that for me at this time.

 

Thanks for your time Chris,

 

miketn

[Chris A]

Hope that things work out well, Mike.  This is a  hobby--not a profession--so I understand your time limitations.

 

Chris

 

P.S., if you want to discuss this subject more, perhaps a PM is the best way to go. 

[ClaudeJ1]

On 2/22/2017 at 2:01 PM, Edgar said:

You are better-off biamping between the woofer and the mid/tweet than you are biamping between the woofer/mid and the tweet. A direct connection between the woofer and its amplifier will bring a noticeable improvement.

I agree with this notion and did that very thing for a few years with my Quarter Pie setup.