Revel M16 Speaker Review

[QMuse]

2) My doubts are predicated on the fact they state the crossover point is at 2khz. Assuming the slope is LR2, even then that would put the breakup (which is only about 5dB above the mean SPL) at around 12dB down from the mean at 5khz.

LR2 at 2000 Hz would actually attenuate for -17.5dB at 5000Hz.

 

[infinitesymphony]

Always fascinated by tweeters that are augmented with a front diffuser.


(pictured: Boston Acoustics Lynnfield VR tweeter)

 

[hardisj]

LR2 at 2000 Hz would actually attenuate for -17.5dB at 5000Hz.

View attachment 52977

I should have known to state "approximately" given the variables and my non-desire to do that math.
But, thanks for verifying.

 

[617]

Ripply on-axis is a consequence of benign interference effects on the waveguide that do not accurately reflect the direct sound because it exists on only a single plane - the vertical plane perpendicular to the centre of the waveguide.



We see this on the Performa series too, as well as other speakers like the Kali IN-8 coaxials (and pretty much most coaxials beyond KEF).

This is a very interesting observation. I would expect it to show on the horizontal axis as well at 0 degrees, am I mistaken?

 

[QMuse]

I should have known to state "approximately" given the variables and my non-desire to do that math.
But, thanks for verifying.

Given your experience I was expecting your approximation to be much more accurate.

 

[hardisj]

Always fascinated by tweeters that are augmented with a front diffuser.


(pictured: Boston Acoustics Lynnfield VR tweeter)

Not sure if you're aware but there are a decent bit of tweeters that have these kind of hardware...
https://hificompass.com/en/speakers/measurements/scan-speak/scanspeak-d3004/604000
http://www.seas.no/index.php?option=com_content&view=category&id=45&Itemid=239
https://erinsaudiocorner.com/driveunits/seas_27afncd_tweeter/


Numerous others.

 

[hardisj]

Given your experience I was expecting your approximation to be much more accurate.

Sometimes I want to be accurate. Sometimes "in the same county" is good enough.

 

[direstraitsfan98]

If anyone wants to pool together the $1600 for the JBL HDI 1600 I'd be willing to start off by funding 20% plus I'll cover the cost of shipping to Amir. It would be under the condition that we get some of our money back when Amir has finished his results, either by one of the members of the fund purchasing the speaker, or selling it on the second hand market. Perhaps even returning it to the dealer for a partial refund (minus the restocking fee) The money would then get distributed back to everyone who partook in the pool

I'm not the one to handle logistics of handling all that money, but I'm just putting something on the table to get started. I really want to see this speaker tested.

 

[QMuse]

Sometimes I want to be accurate. Sometimes "in the same county" is good enough.

In this case it isn't good enough. I cannot imagine any scneario in which woofer would affect FR at 5000 over the LR2 crossover at 2000.

Not to mention that we don't know if the XO is in fact LR4, which is very possible.

 

[hardisj]

In this case it isn't good enough. I cannot imagine any scneario in which woofer would affect FR at 5000 over the LR2 crossover at 2000.

Not to mention that we don't know if the XO is in fact LR4, which is very possible.

That was my point. Even if I was just using -12dB as a pointer, the bump in the response of the raw driver posted (still haven't verified if that's the exact one used) is only about +5dB, just under 5khz. So, like I said, it doesn't seem plausible that would be the culprit.

But, again, we don't even know if that was the driver that was used. The only way to *know* this kind of stuff is to take it out of the speaker and test it on a very large baffle. Which just so happens to be exactly the kind of testing I'll be doing soon.

 

[QMuse]

Btw, in this article Kevin Voecks is not stating exactly the type of XO but he is saying it is of "high order" which more strikes me to be LR4 than LR2.

And XO point is at 2100Hz, not 2000.

 

[bobbooo]

SInce when is on-axis more relevant than listening window, early reflections and sound power? It doesn't even contribute to the predicted in-room response.

Sean Olive in his preference rating paper:

The predicted in-room response (PIR) represents a weighted average of the on-axis, early-reflected and sound power measurements.

 

[hardisj]

Btw, in this article Kevin Voecks is not stating exactly the type of XO but he is saying it is of "high order" which more strikes me to be LR4 than LR2.

And XO point is at 2100Hz, not 2000.

Yes, I realized I mistyped after I posted. But I didn't correct it because it's only 100hz (per octave, if we want to be technical). *shrug*


I'm somewhat surprised they'd be using a high-order crossover considering the price. Usually manufacturers try to use as few components as possible. So, if they did use something like an LR4 instead of LR2 then kudos to them.

 

[QMuse]

Sean Olive in his preference rating paper:

See here, weighting coefs included.

 

[QMuse]

Yes, I realized I mistyped after I posted. But I didn't correct it because it's only 100hz. *shrug*


I'm somewhat surprised they'd be using a high-order crossover considering the price. Usually manufacturers try to use as few components as possible. So, if they did use something like an LR4 instead of LR2 then kudos to them.

He is saying in that article that "high order" XO is used in all Harman HiFi speakers .

 

[hardisj]

He is saying in that article that "high order" XO is used in all Harman HiFi speakers .

Cool.

 

[BurgerCheese]

Hmm, indeed. It's just that I don't see any point of not using well treated paper for woofers. What do you gain?

Paper cones starts flexing (breaking up) at a much lower frequency, often around 1 khz for a 6" woofer. So the potential gain is to avoid that. Personally I don't really notice much difference in cone material as long as the end result was properly engineered.

 

[ctrl]

I don't think you are reading these curves right. All those curves, along with vertical reflections measurement curves, have been accounted for in Early Reflections curve with appropriate weighting, and in Early Reflections curve there is no peak at 5kHz.

The maximum of the resonance changes slightly depending on the measuring angle. As a result, the resonance of the average curves, like the early reflections curve, weakens and becomes wider.
Of course, at larger measurement angles, the cone-breakup resonance also loses influence on the overall frequency response.

1) Are you certain that driver is the one used here? In other Revel models SB Acoustics drivers are used. I googled "revel m16 dayton" but didn't come up with any results that lend itself to proving this is the case.

No, I am not sure. The Dayton chassis is just an example, because the cone-breakup fits quite well to Amir's measurements of the M16.

2) My doubts are predicated on the fact they state the crossover point is at 2khz. Assuming the slope is LR2, even then that would put the breakup (which is only about 5dB above the mean SPL) at around 12dB down from the mean at 5khz. IOW, I don't see the bump in the above graphic being the culprit. I don't know. I just don't see it.

If the filter flank is optimally shaped without considering the cone-breakup the bump is 10dB.

Here's an example. Crossover frequency at 2.1kHz with LR2. Let us assume that the cone-breakup is 10dB.

Hmm, indeed. It's just that I don't see any point of not using well treated paper for woofers. What do you gain?

The metal anodized cone drivers are simply a bit stiffer than pure paper cones and therefore have less partial resonances.
But if they do break up, then really...

 

[hardisj]

If the filter flank is optimally shaped without considering the cone-breakup the bump is 10dB.

Here's an example. Crossover frequency at 2.1kHz with LR2. Let us assume that the cone-breakup is 10dB.

I think that the simulated breakup of 10dB might be a bit of an exaggeration but at some point we are just arguing semantics in how we view the level relative to the mean output. So, no need for either of us to get bogged down in a debate here. Again, what would help answer the question more definitively is if we had the raw driver data. But, we don't. I'm okay with that. Just would be a nice bit of info to know but it doesn't change the fact the issue still exists (however minute it may be).

OT, but what software is that? I'd like to download it if I can.

 

[QMuse]

The maximum of the resonance changes slightly depending on the measuring angle. As a result, the resonance of the average curves, like the early reflections curve, weakens and becomes wider.
Of course, at larger measurement angles, the cone-breakup resonance also loses influence on the overall frequency response.

View attachment 52982



No, I am not sure. The Dayton chassis is just an example, because the cone-breakup fits quite well to Amir's measurements of the M16.



If the filter flank is optimally shaped without considering the cone-breakup the bump is 10dB.

Here's an example. Crossover frequency at 2.1kHz with LR2. Let us assume that the cone-breakup is 10dB.

View attachment 52981



The metal anodized cone drivers are simply a bit stiffer than pure paper cones and therefore have less partial resonances.
But if they do break up, then really...

As long as resonances are not present in those 3 curves you can't really talk about them affecting the audiblity.

Btw, do you really think Harman engineers would pick up a woofer with cone breakup issue for this speaker?

And finally, as Voecks mentioned "high order" crossover in this article it is most likely LR4, not LR2.