JBL M2 Reference Master Monitor Review

[ernestcarl]

It would cost about the same as buying my own Klippel.
Anyways, I figured that VituixCAD might have some magic that would help simulate directivity index. I've downloaded it to see for myself, but being horrendously impatient as well as horrifyingly stupid, I figured I'm better off asking instead.

If not for simulating such things, what is it for?

I think you'd need the individual driver measurements to simulate it reliably. The spinorama contains the measurement of the system as a singular whole.

 

[Absolute]

I think you'd need the individual driver measurements to simulate it reliably. The spinorama contains the measurement of the system as a singular whole.

I thought maybe it was possible to insert the data from M2 into the program and then change the crossovers to see the change in the directivity index/polar pattern.
Perhaps I'm reaching, or perhaps it's simply a matter of box width and woofer size and no need for any other information, I don't know.

 

[ctrl]

I thought maybe it was possible to insert the data from M2 into the program and then change the crossovers to see the change in the directivity index/polar pattern.
Perhaps I'm reaching, or perhaps it's simply a matter of box width and woofer size and no need for any other information, I don't know.

Nope, that would be too easy! As already mentioned you have to measure the drivers individually, for example as described here by @napilopez.
You do not have to make 72 measurements for each driver, it is sufficient to measure +-90° in 10° steps instead of +-180°.

The other possibility would be to create a BEM simulation (or something similar) for the loudspeaker and use it to generate the 72 measurements for each driver. This would be less accurate and would certainly cost you more time than creating the measurements.

 

[Absolute]

Nope, that would be too easy!

I was afraid it would be
Well, I guess I have drag the whole shebang outside somewhere to measure it myself.
I might do that by the end of next decade

Thanks for replying and the tips!

 

[puppet]

Question for the intelligent people here;

Is there a way to simulate how the spinorama would look if we changed the crossover to, say, LR 24 dB/oct at 700 hz?
I'm curious to see how that would affect the directivity curves.

Been experimenting a bit with crossover slopes and frequencies and thought I liked my example above better than the original.
Completely different tweeter, though, so keep that in mind.

Are you measuring your changes? Looked to me that JBL used a 4th order Butterworth slope, which would normally sum with a 3db peak. If they optimized for "flat" with the crossover, it would safe to say that you'd see a dip here when using a 4th order LR slope. Since you're using a different CD you will need to measure it's response at your intended xo point. See if it even makes it there, smoothly, with a target curve as a guide. Same with the woofer. Forget about the xo software's frequency points here as you will probably need some degree of overlap to achieve a "flat" response. Run the tweeter driver w/inverted polarity to adjust delay on it to produce a deep notch will achieve good time alignment. Change the polarity back to normal and give it a listen ... see if the tweeter is still happy.

Changing this shouldn't effect directivity one wit. That is a function of the waveguide and woofer interaction at the chosen xo point ... which JBL already has worked out.

 

[Qbd]

So this is a full range speaker, and therefore presumably it has a fairly low tuned port. Say you were going to use it for home theater and cross it over to the subs at 60-80 Hz, how much «better» (louder) would it be if a similar speaker was designed with a more appropriate port tune for the 60-20000 Hz range?

 

[Duke]

So this is a full range speaker, and therefore presumably it has a fairly low tuned port. Say you were going to use it for home theater and cross it over to the subs at 60-80 Hz, how much «better» (louder) would it be if a similar speaker was designed with a more appropriate port tune for the 60-20000 Hz range?

If a same-size-as-the-M2 loudspeaker was OPTIMIZED for one octave less low-end extension (including a different woofer or woofers with very different parameters), theoretically you could gain as much as 9 dB of efficiency. In practice you'd probably need to use two woofers and either very aggressively EQ the high frequency section or completely re-design it, probably trading off radiation pattern width for efficiency within the coverage angle.

Obviously this is not the ONLY way to increase efficiency over the M2, it's just one example, meant to illustrate how much efficiency you can theoretically gain by giving up an octave of bass extension.

This might be a more practical example, a (passive) loudspeaker with 3 dB more voltage sensitivity than the M2: PBN M2!5. This probably illustrates the upper limit of voltage sensitivity we can squeeze out of the M2's driver/waveguide combination before needing significant active EQ to boost the top end.

(Looking through the other end of the spyglass, my first example illustrates that extending the bass one octave deeper without increasing the enclosure size will cost you about 9 dB of efficiency. This is why small boxes that go deep are not very efficient. Or, if we DO keep the efficiency the same and extend the bass one octave deeper, we will need eight times as much internal volume, assuming optimized woofer parameters. Tradeoffs, always tradeoffs.)

 

[Absolute]

The assholeness of physics strikes again!

 

[Absolute]

Are you measuring your changes? Looked to me that JBL used a 4th order Butterworth slope, which would normally sum with a 3db peak.

Of course, I'm a measureholic!

The crossover is LR 36 dB/oct (acoustically, or at least extremely close to it for the tweeter and an exact match for the woofer) in the M2 with the tweeter crossed at 782 hz and the woofer at 790 hz, tweeter playing in negative polarity.

If they optimized for "flat" with the crossover, it would safe to say that you'd see a dip here when using a 4th order LR slope. Since you're using a different CD you will need to measure it's response at your intended xo point. See if it even makes it there, smoothly, with a target curve as a guide. Same with the woofer. Forget about the xo software's frequency points here as you will probably need some degree of overlap to achieve a "flat" response. Run the tweeter driver w/inverted polarity to adjust delay on it to produce a deep notch will achieve good time alignment. Change the polarity back to normal and give it a listen ... see if the tweeter is still happy.

I think they optimized for even phase/dispersion around the crossover. I know that the intended listening axis is between the drivers roughly at the JBL logo because they told me, and you can probably see that at the -10 degree measurement in Erin's data.

The driver I have is a 18sound 4015 (4 inch membrane) Beryllium tweeter and is more better suited for lower duties than the original D2430k tweeter. It's a trade-off though because that one is better at high frequencies by nature.

I can distortion measure my tweeter crossed at 700 LR 4th order and see if it's a bad idea. It certainly sounds very good at normal levels, more meaty and full than the original crossover.

Changing this shouldn't effect directivity one wit. That is a function of the waveguide and woofer interaction at the chosen xo point ... which JBL already has worked out.

I would think that the crossover point and slopes would affect dispersion, but perhaps not?

 

[puppet]

Not any appreciable amount. If you can stay, within reason, the area that JBL used .. you should be OK. I doubt you'd even notice a 900hz xo to be honest.

 

[puppet]

If a same-size-as-the-M2 loudspeaker was OPTIMIZED for one octave less low-end extension (including a different woofer or woofers with very different parameters), theoretically you could gain as much as 9 dB of efficiency. In practice you'd probably need to use two woofers and either very aggressively EQ the high frequency section or completely re-design it, probably trading off radiation pattern width for efficiency within the coverage angle.

Obviously this is not the ONLY way to increase efficiency over the M2, it's just one example, meant to illustrate how much efficiency you can theoretically gain by giving up an octave of bass extension.

This might be a more practical example, a (passive) loudspeaker with 3 dB more voltage sensitivity than the M2: PBN M2!5. This probably illustrates the upper limit of voltage sensitivity we can squeeze out of the M2's driver/waveguide combination before needing significant active EQ to boost the top end.

(Looking through the other end of the spyglass, my first example illustrates that extending the bass one octave deeper without increasing the enclosure size will cost you about 9 dB of efficiency. This is why small boxes that go deep are not very efficient. Or, if we DO keep the efficiency the same and extend the bass one octave deeper, we will need eight times as much internal volume, assuming optimized woofer parameters. Tradeoffs, always tradeoffs.)

I'd add that the cleanliness of the mid range would suffer as well. A 15" driver can do a couple of things very well, just not at the same time. Deep bass and a pleasant mid range ... I wouldn't expect that. In an application like this, the mid range has to rule out first.

 

[Ron Texas]

Efficient or not, the M2 can take enormous amounts of power and play stupid loud.

 

[Absolute]

Efficient or not, the M2 can take enormous amounts of power and play stupid loud.

Yes, I've had the mother of all "holy shit" moments when the M2 preset in the Crown cdi 4/1200 failed to load properly and the Crown sent full power of white noise to the woofer in one speaker.
It took me a second or two to shut it down, but my ears was ringing like crazy for hours and my heart was beating heavily for an hour or two before I settled down. I've never been so scared in my entire life, I thought the building was coming down.

These things will kill your ears long before you run into any headroom-issues. Apart from the low-end in a large room, of course.

 

[Duke]

I'd add that the cleanliness of the mid range would suffer as well. A 15" driver can do a couple of things very well, just not at the same time. Deep bass and a pleasant mid range ... I wouldn't expect that. In an application like this, the mid range has to rule out first.

I agree with you about prioritizing midrange quality.

Note that as long as the cone's behavior is good within its passband, a large prosound woofer can be competitive with high-quality small midwoofers in motor-strength-to-moving-mass ratio (using Bl^2/Re to calculate motor strength), and is generally superior in applications where a relatively high direct-to-reverberant sound ratio is desired. Of course a big cone starts to misbehave at a lower frequency than a comparable small cone, which in turn calls for a lower crossover frequency, so JBL went to great lengths to develop a horn-and-compression-driver combination which could go low enough to take over from the woofer at 800 Hz, and high enough to not need a supertweeter, while retaining excellent pattern uniformity over its passband.

Imo the primary penalty the M2 imposes on users is its sheer size (and price); I do not think it imposes a qualitative penalty in exchange for its SPL and bandwidth capabilities.

 

[Newman]

Seems like you don't really understand what Klippel near-field scanner actually does, nor understand the limits of anechoic chambers.

I think I do.

What you don't seem to understand is that the answer to inexperienced-amateur-Erin-with-a-big-expensive-new-toy getting a different result to Harman's professionals with their anechoic chamber, doesn't have to be "which tool has the higher technical resolution capability, oh well that's the one that's right and more accurate".

Perhaps an example?

Erin's result:-

Harman's result:-

Erin's result with a different EQ:-

Whoops. Based on that, I reckon the speaker EQ was not the same for the two tests, or at least, it is a more likely explanation than jumping to the conclusion that the Harman lab cannot measure FR within 3 dB in the mids and highs.

So, perhaps a gentle enquiry might have been a better strategy for you than the cocky put-down approach that you chose?

 

[richard12511]

Efficient or not, the M2 can take enormous amounts of power and play stupid loud.

And sound incredible doing it. I know a couple people with M2s who've told me they can get up to 130dB with multiple subs.

Comparing these to the 708p, the M2 has a more constant directivity, and over a much wider range, while playing much louder. I think you get your money's worth. I've been listening to the 708p over the past few weeks, and I think they sound great. 708p is (imo) quite underrated by the Olive score, and I'm thinking the same is true for the M2. M2 is probably one of the best speakers that exist, especially for reference level listening.

 

[richard12511]

I think I do.

What you don't seem to understand is that the answer to inexperienced-amateur-Erin-with-a-big-expensive-new-toy getting a different result to Harman's professionals with their anechoic chamber, doesn't have to be "which tool has the higher technical resolution capability, oh well that's the one that's right and more accurate".

Perhaps an example?

Erin's result:View attachment 151796

Harman's result:View attachment 151797

Erin's result with a different EQ:View attachment 151798

Whoops. Based on that, I reckon the speaker EQ was not the same for the two tests, or at least, it is a more likely explanation than jumping to the conclusion that the Harman lab cannot measure FR within 3 dB in the mids and highs.

So, perhaps a gentle enquiry might have been a better strategy for you than the cocky put-down approach that you chose?

I'm inclined to think these differences are due to:

1. Sample variance. The Harman published spins are no doubt of a near perfect reference sample. That dip/peak may be attenuated in better samples.

2. NFS vs Anechoic chamber. The NFS is higher resolution and more accurate than Harman's anechoic chamber. Anechoic chamber spinorama is just 70 measurements. The NFS takes 1000+ measurements, and can show problems that the lower resolution system just can't see.

3. Smoothing. There is no doubt in my mind that Harman smooths these results before releasing them. We've seen many times now that the NFS (Amir and Erin) reveals problems that just don't exist in the Harman spins.

The NFS didn't exist when Harman made those measurements, but it's fantastic that it does now, and is giving us even more precise measurements. I understand Harman themselves have now purchased one.

 

[Newman]

IMO the incredible correlation between the Harman spins and the Erin chart with different EQ, pushes that explanation to the top of the list (that different EQ is used).

Plus, when Neumann test their KH80 using both methods, the small differences don't support the idea that NFS is so much more resolving that 3 dB differences (in the mids and highs) for the M2 are explained.

(above 10k is due to an uncalibrated mic on the NFS).

 

[napilopez]

IMO the incredible correlation between the Harman spins and the Erin chart with different EQ, pushes that explanation to the top of the list (that different EQ is used).

Plus, when Neumann test their KH80 using both methods, the small differences don't support the idea that NFS is so much more resolving that 3 dB differences (in the mids and highs) for the M2 are explained.
View attachment 151837(above 10k is due to an uncalibrated mic on the NFS).

I'm not convinced the 'different' EQ explanation works, unless you mean to say that Harman is using the wrong EQ in at least some of the speakers it is shipping. Earlier I posted Sound and Recording's anechoic measurement of the M2, which tracks Erin's measurement more closely than Harman's in the upper mids and highs. So their speakers would have needed the wrong EQ as well.

So the explanation is not necessarily that the owner or Erin speaker is using a different EQ, but perhaps rather that the production units vary enough from the 'golden sample' -- or early devlopment unit -- used in the official spinorama that the EQ isn't perfect.

The fact that Erin's measurements can be 'fixed' with EQ doesn't mean much regarding the accuracy of the measurement. In my experience it is almost always the case when third party measurements differ from manufacturer measurements that the directivity remains the same, therefore any FR deviations are naturally EQable. So yes, something could be wrong on the end of the measurement, but it could also be a problem of manufacturer consistency -- and whether the EQ is fully valid for the speaker the user actually gets.

I recently saw this happen with the LS50 Meta and LS50WII; KEF's measurements appear to be from the development phase, or are otherwise not fully representative of the end product (they are shelved down by about 2dB). The end product as measured by 4 or 5 different sources is identical in directivity, but different in overall tonality.

In the case of the KH80, keep in mind Neumann's anechoic measurement is an average of 8 samples with ground plane used for the bass. But yes I don't expect the NFS is dramatically more resolving than an anechoic chamber above the bass.

 

[Newman]

But yes I don't expect the NFS is dramatically more resolving than an anechoic chamber above the bass.

We agree.

Assuming it's speaker sample consistency, then it becomes pretty pointless to choose between models based on one measuring closer-to-ideal than the other, beyond a certain point. And, based on the M2 differences we are observing, that 'certain point' is pretty rough, really.