rePhase, a loudspeaker phase linearization, EQ and FIR filtering tool

Right if its science for our hearing but that front also moves from time to time, regarding real world wave shapes i can do my own science and see output verse input is hopeless distorted no matter what we hear :).

Wave shape output 50Hz SQ wave from a DC - light-speed domain:
633927d1504658724-rephase-loudspeaker-phase-linearization-eq-fir-filtering-tool-2000b-png



Wave shape output 50Hz SQ wave from summed DC - 500Hz pass-band with a 2nd order LR XO point then 500Hz - light-speed pass.band:
633928d1504658724-rephase-loudspeaker-phase-linearization-eq-fir-filtering-tool-2001b-png




Wave shape output 50Hz SQ wave from a 20-20kHz domain:
633929d1504658724-rephase-loudspeaker-phase-linearization-eq-fir-filtering-tool-2002b-png



Wave shape output 50Hz SQ wave from summed 20 - 500Hz pass-band with a 2nd order LR XO point then 500Hz - 20kHz pass-band:
633930d1504658724-rephase-loudspeaker-phase-linearization-eq-fir-filtering-tool-2003b-png

Oh nooo, not the squared (head...:p) wave test agaiiin...:wave2s:
 
Oh nooo, not the squared (head...:p) wave test agaiiin...:wave2s:

In all my experience, science and listening agree.......
perfect magnitude response,
perfect phase response,
perfect impulse response,....
perfect square wave response......

get any ONE of them right, and ALL the others are right too.

It is truly glorious, how getting any one of them right,.... sounds ;)
 
It is utterly funny to see the propagation of the generalisation of this kind of idea.

I rarely see CSD graphs... some people seems to have equalized perfectly their loudspeakers system but the CSD shows that their hear more resonances than audio signal.

Think can follow what you mean but can also follow what mark100 mean and actual pages back he shared nice CSDs, don't know if its rarely but up front below two-wayer correction clean up CSD.
 

Attachments

  • 2000.png
    2000.png
    48.7 KB · Views: 390
  • 2001.png
    2001.png
    110.9 KB · Views: 112
  • 2002.png
    2002.png
    23.3 KB · Views: 109
It is utterly funny to see the propagation of the generalisation of this kind of idea.

I rarely see CSD graphs... some people seems to have equalized perfectly their loudspeakers system but the CSD shows that their hear more resonances than audio signal.

Hi, what does CSD stand for ?

Fwiw, I meant perfect responses in anechoic type environment. ..
I tune my speakers outdoors for closest substitute.

I've found that magnitude, phase, impluse, and yes ..I've even looked at square wave response...........I've found they all move together...fix one, and others get fixed...that's my real point here ....

But more importantly, sound improves. Hardest test of all imo, is off axis square wave response
 
Hi, what does CSD stand for ?
Fwiw, I meant perfect responses in anechoic type environment. ..
I tune my speakers outdoors for closest substitute.
I've found that magnitude, phase, impluse, and yes ..I've even looked at square wave response...........I've found they all move together...fix one, and others get fixed...that's my real point here ....
But more importantly, sound improves. Hardest test of all imo, is off axis square wave response

Agreed, resonances will show as decays on a CSD, and as a bumps or dips in the magnitude response.
This is just a matter of using the correct visualization for what you want to show (or hide ;)).

You can't fix resonances, the only cure is to d'ont trigger them.
When i tune my multi-sub configuration, i use pure tones and band passed white noise... a 50Hz pure tone can give a flat FR response at 0dB from 20 to 200Hz.
 
Account Closed
Joined 2001
Seeing as how many people can tolerate a lot of things that I have a different opinion about, that does not mean that much to me. Which is why I went on to pursue such a thing and make up my own mind. No regrets yet.

Most people who've audibly evaluated phase distortion haven't done it correctly anyways.

Pursuing something and making up your own mind is fine and dandy....as long as you understand all the variables involved and the comparison is valid.

Dave.
 
Most people who've audibly evaluated phase distortion haven't done it correctly anyways.

Dave, as I am a likely member of the above group (safe bet), I wonder if you or anyone might have suggestions as to what I might do differently from this test I ran awhile back:

I looked at a square wave output (~300 Hz) from one of my sealed 2-ways on a stand, with my measurement mic about a meter away at tweeter level. Needless to say, the output bore no resemblance to the input.

I then constructed a convolution with rePhase, accounting for the LR8 crossover slopes, F3 and sealed alignment, simply attempting to un-twist the electrical/mechanical phase respone. To my surprise, the measured waveform became recognizably square-like; obviously band-limited, and with spectral errors, but by god if you squinted a little, it was a square wave!

The catch, as you might guess, is that this corrected output sounded exactly the same to me. That includes the square wave, and any music I could come up with to try to reveal any differences.

I was kind of amazed by this. If I have it right, the LR8 xover alone was applying a whopping 720° shift. And the visual evidence was pretty unassailable - there was definitely something "wrong" before, and it sure looked "fixed" after.

Did I miss something here? Is there anything else I might try in order to hear this better?

Sorry if this is boring, this stuff is likely rudimentary to most of you. Still, any thoughts or suggestions are appreciated.

Jim
 
Account Closed
Joined 2001
Select three or four well-recorded pieces from your music library and get them onto a computer....if not already.

Make sure they don't have any musical peaks above about -3dbFS. If they do, scale them lower so they don't.
You need to do this because processing via an appropriate crossover will increase the peak levels a bit. (You don't want clipping to be a variable.)

Make an exact copy of each and then process each copy with the appropriate all-pass function for the crossover you're interested in. (Example: LR4 would be a 2nd-order all-pass.) You can use Goldwave or similar.

Listen to the original and processed copy without knowing which is which. It's easiest with headphones, but you could use a normal set of speakers also.

Cheers,

Dave.
 
Most people who've audibly evaluated phase distortion haven't done it correctly anyways.

Pursuing something and making up your own mind is fine and dandy....as long as you understand all the variables involved and the comparison is valid.

Dave.

Wesayso's approach to phase manipulation/correction is quite different in some ways to linearizing the phase turn from a crossover filter. And that difference is important when discussing the conclusion or if the evaluation method is valid.

I have built a speaker similar to his and used a correction filter based very heavily on his where there is Excess phase correction up to 1K. The correction aims to follow the generated minimum phase of the frequency response at the listening position. There is a very audible change in the sound of the bass when changes are made to the amount of phase correction or if it is removed completely. This is more room correction (or room speaker interaction) than pure speaker correction. I have never tested the same sort of correction on any other speaker to know if the result is the same but on my line array the difference is clearly audible.

This is the change in phase when viewed through a 6 cycle FDW. There are always arguments as to whether this is the "right" way to go but I find it hard to believe that other's could not hear what I am hearing in the same situation.

attachment.php


Jim what does the phase of your speaker look like at your listening position with and without the correction? Testing at 1m metre might not help you because the room will mess up the phase and then it becomes hard to hear the difference between two different messes :)
 

Attachments

  • In vs Out Phase.jpg
    In vs Out Phase.jpg
    91.4 KB · Views: 948
I'm afraid I haven't taken any measurements of my main speakers at the listening position. My approach to this setup thus far has been conservative driver correction (derived from quasi-anechoic testing) + baffle step comp for the satellites, and room correction + Linkwitz transform for the 2 sealed subs.

Last but not least, a "room" curve, arrived at over time via subjective listening. This ended up being a gentle tilt downward from 32 Hz to 16 Khz @ -0.5 dB/oct, for a total of -4.5 dB using 3 cascaded low-Q shelving filters. Everything is IIR using miniDSP.

I've been so pleased with the sound I'm getting, and so busy with other things (stupid day job!) that I haven't gone any further down the rabbit hole - yet. But I do understand the "2 messes passing in the night" scenario you describe - I'll keep that in mind going forward.

Jim
 
Last edited:
Measurements at the listening position will be a good place to start to understand the impact of any changes you make. Even if a measurement might not tell the whole story its still a good indicator.

I have found that spatially averaged impulse measurements like those that were mentioned by Swiss Bear to be a good base for room correction.

Your approach sounds sensible. I have about twice the slope of yours on the arrays at near 1dB/octave but that is quite speaker (directivity) dependent as on LX521 I like less tilt.