How to Measure Magnepan LRS

[raindance]

Interestingly, when I've measured my .7's with REW, it reports significantly lower distortion from the main listening position when the panels are vertical versus tilted back. So I modified a pair of feet from 1.6's to fit the .7's and now they stand vertical and are elevated a few inches higher than standard off the floor. I found 5' from the front wall to work best in my 23' x 14' x 8' room, which fortunately has the space

I'm a fan of Magnepan, but have no interest in discrediting any measurements. I'm just interested in how it all works. If they measure poorly, I'll still have them - nothing in the same price range has ever come close to sounding "right" to my ears in this room.

 

[wje]

Interestingly, when I've measured my .7's with REW, it reports significantly lower distortion from the main listening position when the panels are vertical versus tilted back. So I modified a pair of feet from 1.6's to fit the .7's and now they stand vertical and are elevated a few inches higher than standard off the floor. I found 5' from the front wall to work best in my 23' x 14' x 8' room, which fortunately has the space

Also, that's why many have opted for 3rd party stands to get their Magnepans in the vertical position. Many have commented that Magenpan should offer better stands, but if they did, it would certainly increase their prices. The MG .7 features an optional upgraded set of stands with two posts and an oval acrylic bass, but is approximately $350 to purchase. But, it does life the MG .7 up a few inches and keeps the speakers vertical.

I happen to have the Mye stands with my MG 1.7i. I wish I had the larger room as you've described you have. I can keep mine positioned 24-28" from the wall, but don't have the real estate in order to go a bit further out from the wall.

 

[raindance]

@wje You know what's funny, is that the tilted back stands help the speakers sell when demoed to standing clients; it alleviates some of the beaminess- how's that for a word?

 

[raindance]

Also, that's why many have opted for 3rd party stands to get their Magnepans in the vertical position. Many have commented that Magenpan should offer better stands, but if they did, it would certainly increase their prices. The MG .7 features an optional upgraded set of stands with two posts and an oval acrylic bass, but is approximately $350 to purchase. But, it does life the MG .7 up a few inches and keeps the speakers vertical.

I happen to have the Mye stands with my MG 1.7i. I wish I had the larger room as you've described you have. I can keep mine positioned 24-28" from the wall, but don't have the real estate in order to go a bit further out from the wall.

View attachment 81201

My room is a bit messy, it's a bonus room / gym / office, but here it is.

 

[Deleted member 2944]

Post deleted , let's not lose our minds and be rude towards one another please

My mind is perfectly fine, and I don't appreciate the inference that it isn't.
A poster here seems incapable of comprehending a basic premise and responds with a snide Seinfeld inspired smart alec response.
If you're going to moderate, do it correctly please.

Dave.

 

[NTK]

Here is an AES convention paper coauthored by Huancai Lu, Don Keele and Sean Wu on a study of the spatial resolution of the method used by the Klippel NFS. This paper should shed some light on the resolving power of this technique on directivity patterns.
https://www.xlrtechs.com/dbkeele.com/PDF/Keele (2006-10 AES Preprint) - Full Sphere Measurements Using HELS Method.pdf

At the higher frequencies (where the directivity pattern can get much more complicated, which is the problem when measuring panel speakers and line arrays), the NFS can use time gated measurements and do without field separation. This eliminates the restrictions of measuring in the "source free" zone and allows measurements at closer distances (but these measurements are only applicable when sound field separation is not used).

 

[Soundstage]

This video may be of help...prehaps.

Is there a trade-off to make between the soundstage improvement at a specific location vs In Room Frequency response?

 

[richard12511]

My mind is perfectly fine, and I don't appreciate the inference that it isn't.
A poster here seems incapable of comprehending a basic premise and responds with a snide Seinfeld inspired smart alec response.
If you're going to moderate, do it correctly please.

Dave.

Getting back to the discussion, I must admit that I'm also having trouble understanding your reasonings for doubting these upcoming measurements. Your point is that due to the distance of the mic, and the size of the panel, there will be comb filtering embedded in each and every measurement. Is that correct? If so, I agree with you. Where I think our opinions differ is in how much of a problem that will be. Certainly it is somewhat of a problem, but it's a problem with every speaker, and it's precisely the problem that the NFS is terrific at solving. It's what it's meant to do.

As for how it solves that problem? Any one measurement on its own is not very useful, due in part to comb filtering, as you mentioned. But, by taking many measurements at different positions, it can start using those measurements in conjunction with one another to tease out the comb filtering effects from the actual measurements of the panel. Each measurement is like an equation (3x + 2y = 19), that's impossible to solve, but with enough equations, you can tease out the values of the variables(x and y). I think it's fair to say that this speaker will probably show more comb filtering and other "fake" response errors than most speakers, and therefore be a harder problem to solve. The beauty of this problem type is that to solve harder and harder versions, all you need is more equations(measurements). The actual strategy or algorithm for solving it remains the same. And, from what I understand, this is easy to program with the NFS. Amir just needs to tell it to take more measurements.

@amirm, if you can, please do a lot of measurements for this speaker, even if just for the purpose of assuaging doubts.

I don't own any panel speakers, but this is one of the reviews I'm looking forward to most. Will be great to finally have truly great measurements of these. I don't expect them to measure great, but as we've seen already, great measurements don't necessarily mean great sound(SVS Ultra), and bad measurements don't necessarily mean bad sound(Revel M55XC). Bad measurements may ensure that the majority won't like the sound, but any one individual may still love it, as individual preference is king(imo).

 

[Deleted member 2944]

I understand there are myriad single-point measurements here.
But I'm certainly willing to believe the processing of those many measurements will (tease out..to use your words) create a meaningful response aggregate to evaluate. I never said otherwise. If you thought I did, you should read my posts again.

The issue that irritated me is Seinfeld's mailman friend being unable to read my posts closely enough to grasp my premise regarding SINGLE (repeat SINGLE) measurements, and then creating a silly straw man argument to come back at with me with. Pitiful.

Dave.

 

[mlee]

For what its worth my line source behaves as a line source in that it falls of at 3dB per doubling distance. Measurements don't become reasonably repeatable until around the 3 meter mark. Before that (closer to the source) I get wildly differing on axis measurements gated in room measurements.

These measurements will be interesting in that the source is probably not behaving like a point or line source, but something in between.

 

[oldsysop]

I have already noticed several comments in this thread, from people who seem to be Maggie fanboys who don't want to hear potential, possible bad news from the measurements that Amir is about to conduct, making preemptive strikes to invalidate the measurement before it even starts.

+1

 

[Deleted member 2944]

A "small" speaker like the MMG/LRS doesn't exhibit the 3db/distance roll-off that a taller (5 foot plus) line-source/array would. In that sense, this speaker is more closely related to a point-source. (In most typical listening arrangements the listener will be in the farfield.)

Regardless, the NFS measurements in this case will certainly be NEAR-field and any associated nearfield/farfield transition that would be exhibited in normal usage will not be captured. I think that's understood......but not really relevant to the testing scheme here.

Dave.

 

[echopraxia]

The issue that irritated me is Seinfeld's mailman friend being unable to read my posts

Pitiful.

a snide Seinfeld inspired smart alec response.

If you're going to moderate, do it correctly please.

In case you are not already self-aware of this: Your tone here is not a very great way to rebuild trust with the mods of your ability to act civilly. It appears that after being warned by @Thomas savage, you are just doubling down with even more (completely unnecessary) inflammatory language.

Reading the post you seem so upset by, I don't see anything particularly rude or hostile there. The language seems quite polite, actually. In contrast, perhaps you are not aware that your responses appear fairly hostile and inflammatory in tone. If anything, the mods are being quite generous to you here.

The issue that irritated me is Seinfeld's mailman friend being unable to read my posts closely enough to grasp my premise regarding SINGLE (repeat SINGLE) measurements, and then creating a silly straw man argument to come back at with me with. Pitiful.

If your point is that you do understand that the NFS results should not suffer from comb filtering artifacts (assuming sufficient measurement points input into the sound field modeling algorithm), then what do you aim to achieve by bringing up comb filtering at all? Or do you have a specific concern about the mathematics used to reverse-infer the sound field from these measurements?

I think the responses you are calling 'strawmans' are simple because you wrote this:

If your microphone is only one foot away you're going to see comb filtering embedded into the measurement because of the large discrepancy in microphone distance. The acoustic path lengths from the top of the panel to the microphone and from the center of the panel to the microphone are different. I don't know how the NFS can "compute" that 'out' of the measurement.

That sounds quite a lot like you were claiming that you 'don't know how the NFS can compute comb filtering out of the measurements'. So I don't think it's a strawman when people here kindly attempt to explain to you how it is possible to do so (especially given that this topic seems to come up over and over and over again).

 

[Vini darko]

Is there a trade-off to make between the soundstage improvement at a specific location vs In Room Frequency response?

Hi , I honsetly dont know having never owned any di-poles. Also I don't have the space in my tiny room to use Ron's method. I'm stuck with shuffling around speakers a inch or two. Sorry can't be more help.

 

[Deleted member 2944]

Why is my position here so darn difficult to grasp? I've REPEATEDLY said a SINGLE-point measurement in the nearfield of a speaker like this will exhibit comb-filtering effects. The physics of that are undeniable.

I well understand the NFS take MULTIPLE single-point measurements to accomplish its goal!!! I'm just not clear on how realistic the assembled result will be. Forgive me for being just the slightest bit skeptical. I hope that's okay with you guys.
There seem to be numerous hypocritical posters here who are pre-judging my "pre-judging" and have already labeled me a Magnepan "fanboy." C'mon!

Dave.

 

[Thomas savage]

My mind is perfectly fine, and I don't appreciate the inference that it isn't.
A poster here seems incapable of comprehending a basic premise and responds with a snide Seinfeld inspired smart alec response.
If you're going to moderate, do it correctly please.

Dave.

Ok , your will receive a sanction for abusive language, another out burst like that one and you will be permanently banned...

there's no acceptable reason to use language like towards another member .

The the original moderation message was a general warning addressing your post ( informally as I did not quote it ) and others behaviours. You're post was deleted so members would of not seen you being addressed specifically. That was as lenient as I could be towards you , unfortunately you have met that leniency and goodwill with this rude and insulting post.

 

[NTK]

Why is my position here so darn difficult to grasp? I've REPEATEDLY said a SINGLE-point measurement in the nearfield of a speaker like this will exhibit comb-filtering effects. The physics of that are undeniable.

I well understand the NFS take MULTIPLE single-point measurements to accomplish its goal!!! I'm just not clear on how realistic the assembled result will be. Forgive me for being just the slightest bit skeptical. I hope that's okay with you guys.
There seem to be numerous hypocritical posters here who are pre-judging my "pre-judging" and have already labeled me a Magnepan "fanboy." C'mon!
Dave.

The measurement inputs to the NFS are the complex pressure amplitudes, which include both magnitude and phase. If only magnitude are used, we won't be able to account for the interference effects when performing summation. But when complex numbers are used in the calculations, they are fully accounted for.

 

[Ville]

The measurement inputs to the NFS are the complex pressure amplitudes, which include both magnitude and phase. If only magnitude are used, we won't be able to account for the interference effects when performing summation. But when complex numbers are used in the calculations, they are fully accounted for.

I would also like to remind about the nature of the series expansion as a solution. Not unlike a Fourier transform/series we are familiar with, also the particular series expansion used in NFS to explain and reconstruct the measured sound field has the ability to describe the sound field perfectly. The higher the order of the series, the more complex sound field can be described. Of course, like with Fourier transform or AD/DA conversion, if you will, proper sampling is required. From the results it is clear that NFS obeys both spatial and temporal sampling requirements mr. Nyqvist would approve, I think.

This analogy with representation of 1D sound signal is sound. Where it gets mind-blowing is the basis used. A Fourier series (of a 1D signal) is sines and cosines with different frequencies. 3D data of 2D sound (with amplitude and phase, is that 5D in total?) is here modeled using spherical harmonics each multiplied by a radial basis function. Each of these components takes part in explaining every measurement point as a superposition (sum). @NTK shows a couple of the first spherical harmonics in Part 1.pdf page 4 in his magnificent documentation:
https://www.audiosciencereview.com/...ner-the-mathematics-and-everything-else.9970/
It is very hard, if not impossible, to imagine or visualize them with the radial part but the point is, that is analogous to the simple series of cosines and sines.

 

[Ville]

The measurement inputs to the NFS are the complex pressure amplitudes, which include both magnitude and phase.

Sorry, can't resist...
I just realized how much fun it is to explain complex math in english:
Complex pressure here is not referring to gas issues nor somehow complicated measurement but --- wait for it --- imaginary numbers!
Now, imaginary number is not a real number... [duh!]
I think you all got the point, just google them. I had my fun, thank you!

 

[tjf]

If anyone here is concerned that the LRS will be shown to be lacking "measurement-wise", I'd suggest worrying more about the distortion results we'll see below 200-300 hz....even at 86db....I'm betting it won't be pretty...