How would you design the Ultimate Magnepan?

[amirm]

If you guys like, I'll shut up, if discussing this a little is too painful or otherwise forbidden in Professional Scientific Audio Circles.

Phase is such an obtuse topic that it definitely merits discussion, if not here, then in another thread.

 

[RayDunzl]

He is asking why it is that the phase delay in his measurements become zero after some frequency and stays that way. Look at his measurements.

What I'm really asking, is to see some other system phase measurements. Mine (for the panel) are 'flat' because we only use the finest speakers and measurement procedures with the lame stuff we have laying around here at the Neverland East Audio Soup Kitchen.

 

[RayDunzl]

Did you do any IR calculations in REW or any other setting that might affect timing?

No, it's the default IR window Tukey .025/125/0/500 and not using a Filtered Impulse.

 

[amirm]

The thread is "How would you design the Ultimate Magnepan?" which is a planar type speaker and mine are sort-of planars, and phase interests me (and JJ from a post at AVS and the Rythmik guy (whose name escapes me at the moment) who said "Impossible")

Oh, did you say JJ also mentioned it is impossible? If so, I feel better .

 

[RayDunzl]

Here's a Primus P-363, stereo, no EQ.

I'll do this measurement again someday, since I may not have taken the same amount of care to precisely distance the speakers and microphone. I just stood them up in front of the reQuests (where they were invisible in the dim light) and fooled my Audio Buddy. They did sound pretty good.

Well, what's different?
DAC?
No.
EQ?
No.
Uh... Cables?
No.
Uh... Uh...
Ok, (lights on).
No way!

Just to show it isn't the software or something weird about the system... Sloppy measurement on the reQuests can look similar to this.

 

[RayDunzl]

Oh, did you say JJ also mentioned it is impossible? If so, I feel better .

No, as I remember he was enthusiastic about phase correction, may have acquired a box to play with himself.

Mr Rythmik Subwoofers ( Chih-Shun 'Brian' Ding) said impossible...

I'll try to find the thread there.

---

Here's one - http://www.avsforum.com/forum/91-au...plain-time-alignment-detail.html#post26612121

Here's the other... Ok, he didn't say "impossible". He said "too good to be true". - http://www.avsforum.com/forum/113-s...-audio-subwoofer-thread-110.html#post23800755

And in a private message JJ showed a little interest in acquiring the miniDSP, don't know whether he did or not.

 

[amirm]

Here's a Primus P-363, stereo, no EQ.

Oh, what do you mean by stereo? You are driving two speakers at once? If so, don't do that! It makes interpretation of the graph very complex to impossible. For low frequency analysis, that is a useful technique but not for frequencies above transition.

 

[AJ Soundfield]

Mine (for the panel) are 'flat'

Ok, so you're asking why the phase if flat for panel from Post #15?
Been told by JJ and Brian Ding that's impossible?

 

[RayDunzl]

The threads are added to post 86 above.

 

[RayDunzl]

Oh, what do you mean by stereo? You are driving two speakers at once? If so, don't do that! It makes interpretation of the graph very complex to impossible.

I don't seem to have any problem interpreting them.

You are just evading the evidence that I have near-perfect phase and you don't.

 

[amirm]

I don't seem to have any problem interpreting them.

So where they in stereo?

You are just evading the evidence that I have near-perfect phase and you don't.

I am ebay shopping for perfect phase as I type this! I will be the king of perfect phase soon. The King of audio will have all the perfect phase he wants. It is just a matter of will.

 

[AJ Soundfield]

I don't seem to have any problem interpreting them.

Well I do.
They don't make sense to me. Your AVS links have different measurements and quotes.
So once more, post #15, mic at LP, both ML speakers active, "unwrapped phase" in REW no DSP manipulation of phase?
What do you see?

 

[Fitzcaraldo215]

Ah, the Baffler himself is baffled. For some reason, I understand Ray better than I do you. But, I am sure Ray will explain.

 

[RayDunzl]

So where they in stereo?

Where? Here. Oh. Were? Yes. It takes a little while to get the microphone well centered (like within a quarter inch or less) to avoid combing and double impulse, but that is how I measure (normally) since that is how I listen (normally).

I am ebay shopping for perfect phase as I type this! I will be the king of perfect phase soon. The King of audio will have all the perfect phase he wants. It is just a matter of will.

Good luck. (check my usually subtle sneer at left).

 

[RayDunzl]

Your AVS links have different measurements and quotes.
So once more, post #15, mic at LP, both ML speakers active, "unwrapped phase" in REW no DSP manipulation of phase?
What do you see?

The AVS links are old old.

Post #15 graph:

Mic at LP? Yes
Both speakers active? Yes
Unwrapped Phase in REW? Yes
DSP manipulation? Yes, but as shown in this previously posted (post #66) graph showing an example of one channel of the AcourateDRC DSP'd output of the preamp (below), there is minimal phase change for the high frequencies, and it only tries to straighten out the lows a little.

This graph is not the speaker, this is the DSP'd signal applied to one speaker - the output of the preamp. The central line is the phase adjustment Acourate selected. The FR is something of an inverse of the in-room response.

Post 64:

Both speakers with same test signal measured at listening position with no DSP. The phase response I measure for the panel frequencies (above 180Hz crossover) are pleasingly (to my eye) flatish. No tricks on this graph, test signal -> DAC -> amp -> speakers -> carefully centered mic.

 

[TitaniumTroy]

Don, if you had to put in order of effectiveness the various Magnepan tweaks. What would top your list, vertical biamping, active crossovers, DSP, Razoring, etc...

Any thoughts on the 3.6s vs 3.7i's with their 1st order crossovers and quasi ribbon bass panel instead of the wire like mine. I would like to upgrade to the 20.x series speaker but my room is only 12'.5 wide by 26' long.

 

[AJ Soundfield]

It takes a little while to get the microphone well centered (like within a quarter inch or less) to avoid combing and double impulse, but that is how I measure (normally) since that is how I listen (normally).
The phase response I measure for the panel frequencies (above 180Hz crossover) are pleasingly (to my eye) flatish.

Ok lets see if we can wrap this up quickly before Fitz says something even dumber.
Back to what JJ said, in the "Time alignment" thread, he's referring to (inter) driver phase integration at the source, not inter channel. He was a bit ambiguous with the "surprisingly low", but to my knowledge, 700hz on up and you're good to go, especially with as you noted percussive attacks. I'll have to dig through my library for the AES references, but suffice to say, your XO being 180hzz and the panel running full bandwidth above that is going to automatically take care any multi-driver phase integration issues >700Hz.
Note I'm not saying inter-ch isn't important but relates more to spatial effects than the percussive ones you referred to, which is a source issue. Btw, some great quotes from JJ there and his linked ppt is even better. If you can understand what he's saying (sorry Fitz), you'll see it aligns quite well with Dr Toole, Greisinger, etc.
It's like a blueprint on how to make speakers, being gifted for free.
How Ethan sat beside him and Amir worked with him....and still didn't pick up anything is beyond me....
As to Brians comments, well as you say those were old measurements...and still make no sense to me.
Ok, so to complete the thread jack, I'm guessing you took issue with Amirs poo-poohing of panels, but as I stated, those results are based on specific models with issues beyond the large radiating diaphragm and dipole>monopole bass integration.
Lastly what we measure with a single pressure mic sample at some point in the soundfield, is quite different from what we hear perceptually with binaural hearing. There is some relevant data there, sure, but limited. Things are a bit more complex that that, I'm still not sure what your "mono" measurement is showing, but for a panel with 180hz XO, a 1/6th octave resolution FR from each channel independently > 500hz or so is mostly what one would need, imho.

cheers,

AJ

 

[Purité Audio]

So your synopsis regarding panels would be for the hard of thinking, good or bad?
keith.

 

[AJ Soundfield]

So your synopsis regarding panels would be for the hard of thinking, good or bad?
keith.

Good if you like them, bad if you don't.
They tie back to the comment JJ made about around 30% of folks preferring a more diffuse type sound, others a more direct. (Don't ask me where he got those numbers from).
Being dipolar the panel will inject more energy/reflections into the indirect portion of the soundfield in the room (which ties to why I was asking Dr Olive for data on the "scrim" used by Harman in their blind tests, not the direct incidence but multi-angles and total losses at the LP). There is that difference with typical monopole/box speakers, although as I pointed out previously, the narrow baffle/small driver trend with monopolar boxes leads to quite a bit of rearward radiation, even if most are blissfully unaware. The rearward frequency spectra will be quite different.
But there is also some diffuseness to the direct/onset sound from the panel, because at 3' or so in height, sounds will be arriving at your ears slightly differently from the top/bottom than middle where your ear height might be. This will tend to have spatial effects, a sort of enlarging of the image, which would be most noticeable with things that are real life more point source, like voices. They will tend to sound a bit "bigger". Panel fans might adapt to this or simply not find it objectionable, but a rapid switch blind in mono is most certainly going to make this noticeable! With things like a big orchestral piece, it may sound wonderful. YMMV.
Whether one prefers this a a whole 'nother matter. Again, the specific models tested by Harman had issues well beyond this...

cheers,

AJ

 

[Purité Audio]

Educational AJ as always, are panels something that you have considered experimenting with?
I was thinking of that Gilbert Briggs quote when he first heard Quad's ESLs.
Keith.