Linear Phase EQ or Minimum Phase EQ single driver speaker

[Hasan Aydin]

When I want to equalize a Single Driver Speaker at the listening position (The Room is not treated perfectly but its symmetrical and very large). Should I use a Linear Phase EQ or Minimum Phase EQ?
Especially when the speaker itself has some resonances (peaks and dips) at higher frequencies.

 

[DVDdoug]

You can't hear phase shifts introduced by EQ, unless there is a "reference". i.e. if you phase-shift one stereo channel you might" hear something" when the waves mix in the air. If both channels match you won't hear the phase shift.

Phase shifts (such as an all-pass filter) can cause clipping in the digital domain but of course EQ can cause clipping anyway if you're "not careful", even if there is no phase shift.

...In most 2 or 3-Way speakers the polarity of one driver is flipped 180 degrees because there is a phase lag in the low-pass and a lead in the high-pass and that puts them back in-phase at the crossover frequency (where both drivers are putting-out sound). Nobody notices it. (The phase shifts caused by the distance between drivers and different distance to the ear can create interference issues in the crossover range.)

 

[MRC01]

As I learned the practice of room correction, room resonances are minimum phase, so use minimum phase EQ to correct them.

 

[Hasan Aydin]

Thank you! I read in Bob Katz's book that he uses Linear Phase Room Correction, is that because he has speakers with multiple drivers or is that just a matter of taste?
Because there is definitely a sound difference between these EQs (at least for headphones), but what would be technically correct?

I've already looked at this:

but I'm still not sure, because from my measurements the phase looks like a minimal phase response (at the listening position), but the step response looks like a speaker with multiple drivers and crossovers according to this video. Although it is definitely a single driver speaker.

 

[Hasan Aydin]

When I want to equalize a Single Driver Speaker at the listening position (The Room is not treated perfectly but its symmetrical and very large). Should I use a Linear Phase EQ or Minimum Phase EQ?
Especially when the speaker itself has some resonances (peaks and dips) at higher frequencies.

To answer my own question: I have found that it depends on...If I want to correct the ANECHOIC response of a single speaker then I have to use a minimum phase EQ, but in a normal room at the listening position the phase is no longer minimum, it is mixed. In theory, an EQ that allows me to adjust the phase independently would be perfect in this case.

Feel free to correct me if I am wrong.

 

[KSTR]

There is a simple rule; minimum-phase problems require minimum-phase corrections, otherwise you will get pre-ringing artifacts.

Single speaker drivers are (very close to) minimum-phase systems hence any correction filters must be minimum-phase, too, as we want the total result to be minimum-phase. Notable exception: back-loaded horns and other exotics.

Multi-way speakers are (typically) neither minimum-phase nor linear-phase and therefore their correction method is also neither one, rather it is a mixed-phase correction. The excess phase (phase on top of the natural minimum phase for a given frequency response) must be corrected with its time inverse to yield a minimum-phase system, which then is further corrected with minimum-phase means to the frequency response target.

In-room response is adding another aspect but contrary to common belief most room problems are minimum-phase.

 

[ernestcarl]

the step response looks like a speaker with multiple drivers and crossovers according to this video. Although it is definitely a single driver speaker.

Can you show us how it looks in the very nearfield (e.g. 0.5m) and at the MLP?

MDAT file would be even better.

 

[Hasan Aydin]

Can you show us how it looks in the very nearfield (e.g. 0.5m) and at the MLP?

MDAT file would be even better.

Sure!

WeTransfer - Send Large Files & Share Photos Online - Up to 2GB Free

WeTransfer is the simplest way to send your files around the world

we.tl

 

[maty]

Multi-way speakers are (typically) neither minimum-phase nor linear-phase and therefore their correction method is also neither one, rather it is a mixed-phase correction. The excess phase (phase on top of the natural minimum phase for a given frequency response) must be corrected with its time inverse to yield a minimum-phase system, which then is further corrected with minimum-phase means to the frequency response target.

In-room response is adding another aspect but contrary to common belief most room problems are minimum-phase.

And coaxials like small 5.25" from KEF?

I agree with minimum-phase. Better sound in my second audio system (KEF Q100 very modified) and very usually recordings with physical instruments and voices without autotune.

 

[ernestcarl]

Sure!

WeTransfer - Send Large Files & Share Photos Online - Up to 2GB Free

WeTransfer is the simplest way to send your files around the world

we.tl

I'm not seeing much excess phase to "fix"

*applied "estimate IR delay" in REW

Your single driver measurement is not all that different from my own Fostex 6301

 

[ernestcarl]

he step response looks like a speaker with multiple drivers and crossovers according to this video. Although it is definitely a single driver speaker.

The step looks exactly like a speaker with a single driver. Of course, no single driver speaker is "perfect".

 

[voodooless]

The significant advantage of FIR over IIR is that you have the choice of min vs lin phase, and depending on the number of taps, you have much more fine-grained control over the EQ frequency response. With IIR, you'll need to add more and more filters. With FIR, you just have a frequency resolution dependent on the number of taps, and within that, you can have as many corrections as you want.

Now the disadvantage is that you need a massive amount of filter taps to do proper bass corrections because you need high frequency-resolution. So generally a combination of FIR and IIR is an efficient way to do corrections. Tools like Dirac also effectively use this approach, limiting the number of taps to 1024, or so, at 48k sampling.

 

[ernestcarl]

when the speaker itself has some resonances (peaks and dips) at higher frequencies.

I personally would not necessarily want to cut down the apparent hump/peak in the higher frequencies:


The curves used for the RMS average was taken nearfield at 10 degree increments.

The high FR of many single driver speakers sound rather dull compared to multiways with dedicated HF drivers:

*with EQ where room is very dry


And here is yours:

 

[Hasan Aydin]

The step looks exactly like a speaker with a single driver. Of course, no single driver speaker is "perfect".

Thank you! Maybe I need to learn to interpret my measurements better

Your single driver measurement is not all that different from my own Fostex 6301

you got me

 

[Hasan Aydin]

I personally would not necessarily want to cut down the apparent hump/peak in the higher frequencies:
View attachment 284179 View attachment 284180

The curves used for the RMS average was taken nearfield at 10 degree increments.

View attachment 284181

The high FR of many single driver speakers sound rather dull compared to multiways with dedicated HF drivers:
View attachment 284182
*with EQ where room is very dry


And here is yours:
View attachment 284183

Wow, I will try it out!

Is there a reason why you used a FDW with 6 Cycles instead of smoothing (e.g. psychoacoustic or var)?

 

[ernestcarl]

The significant advantage of FIR over IIR is that you have the choice of min vs lin phase, and depending on the number of taps, you have much more fine-grained control over the EQ frequency response. With IIR, you'll need to add more and more filters. With FIR, you just have a frequency resolution dependent on the number of taps, and within that, you can have as many corrections as you want.

Now the disadvantage is that you need a massive amount of filter taps to do proper bass corrections because you need high frequency-resolution. So generally a combination of FIR and IIR is an efficient way to do corrections. Tools like Dirac also effectively use this approach, limiting the number of taps to 1024, or so, at 48k sampling.

I have to admit that I use/abuse mixed phase FIR EQ on my Fostex 6301 (possibly ineptly) in order to get it to play nicely with the rest of my speakers in a multichannel environment -- although, I do not really fill-in the dips as much as the example I showed before. Hmmmn... I do wonder if it's completely pointless to bother with the small phase manipulations I do in the HF (done to reduce apparent cancellation/lobing when summed with front mains):



However, when playing with the fronts in MCH "upmixing" mode:

Wow, I will try it out!

Is there a reason why you used a FDW with 6 Cycles instead of smoothing (e.g. psychoacoustic or var)?

It discards later incoming energy that you might not necessarily want to include/contaminating your FR curves when measuring at a distance -- i.e. you sort of get a semblance of the "early FR" as well as it smooths and get rids of extra phase turns like what's seen of the bass managed center channel graphic above (15 cycles apparently was not enough windowing).

As to what exactly is the best windowing and/or smoothing settings... it depends. Personally, I just play around with the different settings -- and, sometimes, I don't apply any at all.