Fir filter thoughts?

[Edgar]

1 minute ago, Tarheel TJ said:

   Thanks for your insight into linear phase filters, it is a fascinating subject.  Could this problem be solved by also delaying the video signal?  It wouldn't work for video games, obviously, but it should work for movies/video, right?

 

Yes. But delaying video signals is difficult due to the data quantities involved.

 

Quote

What hardware are people using to incorporate FIR filters into their hi-fi system?

 

I do all of my signal processing on a Windows PC that I program myself. I can't speak for others.

[gnarly]

20 minutes ago, Edgar said:

 

True, they cancel each other out mathematically, but as you pointed-out, getting them to cancel acoustically is a huge problem in itself, even with identical drivers. The ringing in one driver is identical to the complementary ringing in the other, but of opposite polarity. On-axis, the two cancel each other out. Off-axis, they do not. That energy still goes into the room, different portions of it at different angles relative to the main axis of the loudspeaker. It is uncontrolled energy that corrupts the signal.

 

 

Agreed, other than as to significance.

 

I've made extensive on and off axis measurements outdoors on numerous 3-way tops using the described process, and found off-axis holds up very well.

Evidently, there is very little cancellation that needs to occur.  Off-axis measurements are simply quite good; wouldn't they show that uncontrolled energy?

 

I probably need to stress the degree of care taken with the pre-xover, driver-by-driver min phase, out-of-band flattening work. It is essential ime/imo.

Without it,  i would expect response problems in the exact nature you describe.  It's the same problem with linearizing phase on top of an existing speaker with xovers already in place. 

 

I should restress how much steep xovers have helped pragmatic driver-by-driver work, to get both on and off-axis to hold together.

 

When an acoustically complementary design is achieved, i have to view pre-ringing from linear phase xovers, even steep ones, as a total non issue.

[gnarly]

44 minutes ago, Tarheel TJ said:

   Thanks for your insight into linear phase filters, it is a fascinating subject.  Could this problem be solved by also delaying the video signal?  It wouldn't work for video games, obviously, but it should work for movies/video, right?

    What hardware are people using to incorporate FIR filters into their hi-fi system?

I know nothing about video delaying.

 

But sure have used a lot of different audio FIR platforms.

Right now, using Q-Sys with Core 110f and 500i processors. 

Before that, 4 OpenDRC-DI's. Before that, Jriver using multichannel convolution via Config File. Before that, an OpenDRC-DA-8.

So a long time journey with FIR 😀

 

Jriver is a great way to wade in imo.  

[Edgar]

47 minutes ago, gnarly said:

I've made extensive on and off axis measurements outdoors on numerous 3-way tops using the described process, and found off-axis holds up very well.

Evidently, there is very little cancellation that needs to occur.  Off-axis measurements are simply quite good; wouldn't they show that uncontrolled energy?

 

It won't show up at all with steady-state signals, only on transients.

 

Quote

I should restress how much steep xovers have helped pragmatic driver-by-driver work, to get both on and off-axis to hold together.

When an acoustically complementary design is achieved, i have to view pre-ringing from linear phase xovers, even steep ones, as a total non issue.

 

I'm not about to argue with you or anyone else about it. Steep crossover slopes solve some big problems (undesirable out-of-band response from drivers, low frequency excursion limits, etc.) while introducing some little ones (time domain ringing). As with all things, their use represents a tradeoff.

[gnarly]

31 minutes ago, Edgar said:

 

It won't show up at all with steady-state signals, only on transients.

 

 

I'm not about to argue with you or anyone else about it. Steep crossover slopes solve some big problems (undesirable out-of-band response from drivers, low frequency excursion limits, etc.) while introducing some little ones (time domain ringing). As with all things, their use represents a tradeoff.

 

Cool.  No arguing from here either.

 

My testing has been with the usual dual-channel pink, or pink periodic.  Or sine sweeps ala REW.

What transients would you suggest testing with?

Maybe wavelets?  I've just starting using them..

[Edgar]

46 minutes ago, gnarly said:

What transients would you suggest testing with?

 

Frankly, as a start I'd suggest careful listening with lots of percussion.

[gnarly]

Yes!  😄

 

I have done so much of that.  So much....     

Both indoors and out, trusting the sound of outdoors waay over indoors.

Percussive transients, which i take equals full range response,  are such a great listening test.

It's one of the ways I think flat phase smokes the typical phase wrapped signal.  All the voltages of the frequencies contributing to the transient's response stack together for a higher voltage peak when in phase..  Easy to see on a scope, easy to hear.

 

So far, traditional dual-channel measurements that have both flat mag and phase, have no doubt for me, provided the best transient (and tonal) response.

(For tonal response, this is ignoring house curves, and track-by-track tonal biases.)

 

But i'm always in search of the better measurement mousetrap.....that leads to better sound. 

If you have a measurement technique that can pinpoint transient acoustic ringing problems, I'm all ears !

 

Edited April 22, 2021 by gnarly

[Chris A]

On 4/22/2021 at 3:44 PM, gnarly said:

It's one of the ways I think flat phase smokes the typical phase wrapped signal.  All the voltages of the frequencies contributing to the transient's response stack together for a higher voltage peak when in phase..  Easy to see on a scope, easy to hear.

 

So the question is: how flat does the phase need to be?  For SPL response, audio system standards quote ±2 dB flatness (without "house curves"--which is another can of worms).  My experience is ±90 degrees (per Danley).  How about your experience?

Chris

[Edgar]

On 4/22/2021 at 3:44 PM, gnarly said:

It's one of the ways I think flat phase smokes the typical phase wrapped signal. 

 

I just noticed this part of the message for the first time. Exactly what do you mean by "phase wrapped signal"? If you mean a phase curve that passes through -180° and re-emerges at +180°, possibly multiple times, that has nothing to do with the linearity of the phase. It just means that there is more time delay in the response.

[Chris A]

I do believe that there is a listening difference between linear phase and flat phase (subjective listening), but my experience isn't something that I'd be willing to lay a hat on.  It's another question--and potentially another "agreed requirement", i.e.,

 

"How much tilt to the phase curve before it becomes audible?"

 

Chris

[Edgar]

58 minutes ago, Chris A said:

I do believe that there is a listening difference between linear phase and flat phase (subjective listening), but my experience isn't something that I'd be willing to lay a hat on.  It's another question--and potentially another "agreed requirement", i.e.,

 

"How much tilt to the phase curve before it becomes audible?"

 

Oh, come on. Move your head 5/8" farther away from the speaker and the phase will wrap once. Move it 6" and the phase will wrap ten times. 

[Chris A]

So you're saying that only group delay is audible--the rate of change of phase...?

 

Chris

[Edgar]

7 minutes ago, Chris A said:

So you're saying that only group delay is audible--the rate of change of phase...?

 

What I'm describing is actually phase delay -- phase shift due to time delay; -Phi/omega. Group delay is -d(Phi)/d(omega), and describes the time delay of the envelope of a modulated signal.

 

There is debate as to whether either of them is audible.

[Chris A]

Okay...we can agree to disagree on this point of audibility (both points, in fact)--as a function of frequency.

 

Chris

[Edgar]

Just now, Chris A said:

Okay...we can agree to disagree on this point of audibility (both points, in fact). 

 

I'm pretty sure that phase delay is inaudible. I'm not so sure about group delay -- for example, Linkwitz-Riley crossovers create group delay that varies significantly with frequency. Having compared their sound with that of perfect reconstruction crossovers with similar parameters, I can say that they definitely sound different. Whether that difference is due to the varying group delay, well, I just don't know.

 

- Greg

[Chris A]

See:  https://community.klipsch.com/index.php?/topic/182419-subconscious-auditory-effects-of-quasi-linear-phase-loudspeakers/&tab=comments#comment-2366185

 

Chris

[Edgar]

4 minutes ago, Chris A said:

See:  https://community.klipsch.com/index.php?/topic/182419-subconscious-auditory-effects-of-quasi-linear-phase-loudspeakers/&tab=comments#comment-2366185

 

"The results of these tests showed clearly that the addition of the excess phase equalizer had an effect on the perceived sound, albeit a very subtle one."

 

That matches my own observations precisely. In my case I compared Linkwitz-Riley (in-phase at all frequencies, not perfect-reconstruction, produce both group and phase delay) with my own matched-delay subtractive crossovers (in-phase at all frequencies, perfect-reconstruction, and produce only phase delay). The sonic differences were as Greenfield and Hawksford said -- subtle. But they were present.

 

- Greg

[Chris A]

I found it (phase flattening and group delay flattening) not so subtle with full-range loudspeaker directivity (down the room's transition frequency) and nearfield control of early reflections.  It almost knocked me down when I heard it.  The problem is, it's difficult to describe.

 

Chris

[Edgar]

1 minute ago, Chris A said:

I found it (phase flattening and group delay flattening) not so subtle with full-range loudspeaker directivity (down the room's transition frequency) and nearfield control of early reflections.  It almost knocked me down when I heard it.  The problem is, it's difficult to describe.

 

I need to get to Arlington someday ... 😇

 

- Greg

[Chris A]

Invitation is open--and we're all vaccinated. 

 

Chris