• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Revel W990 Review (in-wall speaker)

Sancus

Major Contributor
Forum Donor
Joined
Nov 30, 2018
Messages
1,735
Likes
4,164
Location
Canada
It's awfully easy to complain that an enclosure should be built when you aren't the one who actually has to do that for every in wall speaker review...
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
35,320
Likes
134,794
Location
Seattle Area
It's awfully easy to complain that an enclosure should be built when you aren't the one who actually has to do that for every in wall speaker review...
:)

I actually started by building such enclosures. It produced horrible results due to leaks in the back and diffraction effects -- neither one of which is in play in a real installation.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
35,320
Likes
134,794
Location
Seattle Area
On what page do we find the data for the 3.5 inch deep stud space the speaker mounts in?
Where did you get that from? Maybe it is going in a 6 inch stud wall. Or a false front wall with 2 feet behind it. How about the total volume or leakage into the rest of the back space? You think of that?

I think that the in-wall speaker should be tested in a enclosure similar in properties and volume to where it will actually be used.
All of our regular speaker measurements are anechoic. Surely that is not how they get used.

Since there are no walls that are representative -- just as no room is representative -- the only thing we can do is to pick anechoic infinite baffle as the reference.
 

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
509
Likes
287
Where did you get that from? Maybe it is going in a 6 inch stud wall. Or a false front wall with 2 feet behind it. How about the total volume or leakage into the rest of the back space? You think of that?


All of our regular speaker measurements are anechoic. Surely that is not how they get used.

Since there are no walls that are representative -- just as no room is representative -- the only thing we can do is to pick anechoic infinite baffle as the reference.

Hello,

Sure.

The wall could be built with 5 1/2” studs or a furred out 2 foot deep space. As always the exception is not the rule. The typical wall is framed with 3 1/2” studs. Leaving the typical infinite enclosure that speaker is installed in as a 15” width between the studs and 3 1/2” deep and sort of sealed front and back with drywall or something similar. The stud space may leak sum, that is what we have to work with and perhaps should be included in the software model.

If the only thing to do is select anechoic infinite baffle then do just that. Your instrumentation is set up with time windowing to remove the reflections, there you have anechoic. If your software is modeling an infinite baffle on the backside of that speaker build one. Take a half sheet of drywall, screw on a couple of 3 ½” studs and close up the back side with a smaller piece of drywall and your mockup is much closer to the software model that you are using.

If you want anechoic infinite baffle there you go. What you did may be good enough. I won’t bother you about it again.

Thanks DT
 

YSC

Major Contributor
Joined
Dec 31, 2019
Messages
1,474
Likes
1,017
Hello,

Sure.

The wall could be built with 5 1/2” studs or a furred out 2 foot deep space. As always the exception is not the rule. The typical wall is framed with 3 1/2” studs. Leaving the typical infinite enclosure that speaker is installed in as a 15” width between the studs and 3 1/2” deep and sort of sealed front and back with drywall or something similar. The stud space may leak sum, that is what we have to work with and perhaps should be included in the software model.

If the only thing to do is select anechoic infinite baffle then do just that. Your instrumentation is set up with time windowing to remove the reflections, there you have anechoic. If your software is modeling an infinite baffle on the backside of that speaker build one. Take a half sheet of drywall, screw on a couple of 3 ½” studs and close up the back side with a smaller piece of drywall and your mockup is much closer to the software model that you are using.

If you want anechoic infinite baffle there you go. What you did may be good enough. I won’t bother you about it again.

Thanks DT
I think that can be too much a request to Amirm, he's just a hobbyist doing good to the community for his geekiness. since from what I saw the NFS isn't something light and easy to move around as a PC, and it's placed inside a room in his house, it's unreasonable to ask him to take bricks to build a wall to compromise room space just for the testing of in wall speakers, which isn't easy to demolish when one day he felt that's it, no more testing (that day will eventually come sadly).

I would say what he did is good enough for the purpose, and personally I would not compare these with complete box speakers in Spin data alone due to measurement inequality, but use to compare in the same category, which I believe when someone opt for in wall speakers, there's more than simply wanting best performance. In this regard both use the same unfair baffle should be fair enough
 

nathan

Active Member
Joined
May 24, 2020
Messages
166
Likes
122
I think you may be on to something. In several of the Revel manuals for their in wall speakers it is impied they tested in a traditional North America stud bay, which is 3.5”deep, between 16” and 24” wide, and about 96” tall. And they go so far as to note that if there is blocking, how to adjust the location of the speaker to a particular ratio between top and bottom of the bay, to control waves in the stud bay in a desirable fashion.

So then could one assume that they designed these speakers to work optimally with a two to three cubic foot enclosure? Would changing the shape but preserving that volume be a fair approximation?

Or is this really just an indirect indictment of Revels strange choice not to include sealed enclosures even though several competitors do, who seem to get more predictable results perhaps because of that?

Hello,

Sure.

The wall could be built with 5 1/2” studs or a furred out 2 foot deep space. As always the exception is not the rule. The typical wall is framed with 3 1/2” studs. Leaving the typical infinite enclosure that speaker is installed in as a 15” width between the studs and 3 1/2” deep and sort of sealed front and back with drywall or something similar. The stud space may leak sum, that is what we have to work with and perhaps should be included in the software model.

If the only thing to do is select anechoic infinite baffle then do just that. Your instrumentation is set up with time windowing to remove the reflections, there you have anechoic. If your software is modeling an infinite baffle on the backside of that speaker build one. Take a half sheet of drywall, screw on a couple of 3 ½” studs and close up the back side with a smaller piece of drywall and your mockup is much closer to the software model that you are using.

If you want anechoic infinite baffle there you go. What you did may be good enough. I won’t bother you about it again.

Thanks DT
 

Trouble Maker

Senior Member
Joined
Jan 6, 2020
Messages
483
Likes
446
Location
Columbus, Ohio, US
Or is this really just an indirect indictment of Revels strange choice not to include sealed enclosures even though several competitors do, who seem to get more predictable results perhaps because of that?

I've been thinking, dreaming, about a full surround system with in walls for a while. I keep dithering back and forth between a sealed box as part of the designed package being a must have. Your note spurs another thought. Other than a few* rare, and usually very expensive to purchase and more difficult and costly to install, sealed boxes are going to be quite small. So there's the tradeoff, which always happens, is either sealed box that can retrofit with a really small volume, or variable volume of an open back retrofit style.

*I've seen a few subs and in wall speakers that basically take the whole stud space for their sealed enclosure. But those will need to be installed and then drywall installed over them. This will clearly be a more difficult and costly install than a retrofit. And from my memory they are usually prohibitively expensive.
 

nathan

Active Member
Joined
May 24, 2020
Messages
166
Likes
122
Triad, JBL, Procella, and several other brands more common for consumers like maybe KEF, Focal, etc, seem to get good performance with sealed in wall speakers with surprisingly modest sized enclosures. Up to a point.

For example, the Triad lines until the Gold level can fit in a normal US stud bay with an opening essentially the size of the speaker grill.

But yes at the high output level especially subwoofer type gear, it becomes more of a construction product. For sure.

I've been thinking, dreaming, about a full surround system with in walls for a while. I keep dithering back and forth between a sealed box as part of the designed package being a must have. Your note spurs another thought. Other than a few* rare, and usually very expensive to purchase and more difficult and costly to install, sealed boxes are going to be quite small. So there's the tradeoff, which always happens, is either sealed box that can retrofit with a really small volume, or variable volume of an open back retrofit style.

*I've seen a few subs and in wall speakers that basically take the whole stud space for their sealed enclosure. But those will need to be installed and then drywall installed over them. This will clearly be a more difficult and costly install than a retrofit. And from my memory they are usually prohibitively expensive.
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
1,266
Likes
2,498
Location
US East
On what page do we find the data for the 3.5 inch deep stud space the speaker mounts in? I believe that you will have better luck modeling with the woofer TS/P and the dimensions of the stud cavity in the wall.

If I was to hazard a guess that is where the speaker manufacturer started.

This information is not in the Kippel patent.
What does it got to do with the Klippel NFS? So if the test is done in a 2π anechoic chamber, your "stud issue" goes away?
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
35,320
Likes
134,794
Location
Seattle Area
The wall could be built with 5 1/2” studs or a furred out 2 foot deep space. As always the exception is not the rule. The typical wall is framed with 3 1/2” studs. Leaving the typical infinite enclosure that speaker is installed in as a 15” width between the studs and 3 1/2” deep and sort of sealed front and back with drywall or so
Speaker doesn't care about the size of the studs. it cares about the volume behind it which you can't possibly control. This will always be a variable that you need to deal with using equalization as I explained.

The back volume also impacts bass. In that region, my measurements correlate very well with Harmans:

index.php


Spin%2B-%2BRevel%2BW990.png


Agreement up to 1 kHz is excellent with the same valleys and hills. Clearly then Harman's test conditions were similar to mine and not what you are suggesting.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
35,320
Likes
134,794
Location
Seattle Area
Your instrumentation is set up with time windowing to remove the reflections, there you have anechoic.
You are wrong. You don't need a Klippel scanner if all you want to do is gate the measurements. Klippel only uses that for high frequencies as indicated in the graph I already showed you:

index.php


You see that amber vertical line with the text "Reflection-free Frequency?" Everything after that is gated out. Below that, actual radiation from the reflections is computed (as shown in dashed blue) and subtracted from the total energy to arrive at anechoic response.

Spend some time learning the technology instead of complaining and making wrong assumptions.
 

DualTriode

Addicted to Fun and Learning
Joined
Oct 24, 2019
Messages
509
Likes
287
Hello,

Once again the exception is not the rule.

So far I am about 95% correct. Where would the fun be if there was nothing new to learn?

I said that the software calculated anechoic results; part is from time windowing and now you tell us part is with other tools. The results are still an anechoic approximation, just as I said, just a bit off in the details.

I was right on target about the size and shape of the stud space where the speaker is installed.

The infinite baffle assumption is nowhere near the target. The approximation to an enclosure where the speaker will be installed is a 2 maybe 3 cubic foot space with some degree of leakage. The installed Qtc, Resonant Frequency and F3 frequency will be completely different than the open baffle conditions where it was tested.

Take a look over at https://www.linkwitzlab.com/models.htm for the models of Di-Pole speaker response; peaks and nulls made up of flower petals.

“The frequency response has a characteristic 6 dB/octave roll-off towards lower frequencies and sharp nulls at higher frequencies. The off-axis response, which is shown for 30, 45 and 60 degrees, is lower in amplitude at low frequencies by 1, 3, 6 dB respectively and follows a figure-eight or cos(a) pattern with nulls at +/-90 degrees. The cos(a) pattern is maintained for D/l < 0.1, down to the lowest frequency. The pattern widens with increasing frequency. There are also deep notches on-axis every time that D is a multiple of a wavelength and the polar pattern looks like made up of flower petals.”

Fritz Linkwitz (see the link)

The peaking in your FR measurements between 150hZ and 450hZ is typical Open Baffle response.

I have measured many speakers, open baffle some closed and some with intentional leaks, each having different perhaps signature shaped frequency response plots.

Now you are measuring some sealed in-wall speakers and some open back speakers expecting the results to be comparable.

Perhaps the results are good enough.

Thanks DT



The big picture:

My conclusions are not so much different than yours.

“Conclusions
Measuring speakers this way is challenging. Assuming we trust our measurements, the frequency response plots are not great. I expect far better from Revel. Maybe they know something about how these get used resulting in different tuning. Sadly because I don't have a wall to mount them in, I can't tell you anything about the subjective results either. Given what we have, and the high cost, I would not be buying these”.
 

MdeVelde

Member
Joined
Aug 30, 2021
Messages
19
Likes
8
The reference axis ("on-axis") for these speakers should be 30 degrees horizontal.
Since these are wall mounted speakers there is no toe-in possible.

Btw, I feel this should be standard practice with regular free standing speakers too to find the optimal horizontal axis.
Some speakers are designed for full toe-in while other speakers are designed to point straight forward and should be measured under 30 degrees horizontal.

For what it's worth I'm all for designing speakers that are not meant to be toed-in. You get a better power response and more natural reflections in the room that way which results in better sound quality.
 

respice finem

Addicted to Fun and Learning
Joined
Feb 1, 2021
Messages
993
Likes
1,371
Some (not many) manufacturers issue detailed placement instructions, which should be mandatory IMHO.
Most rooms will "play" more than the speakers, especially if the speakers are placed "at random".
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
35,320
Likes
134,794
Location
Seattle Area
Some (not many) manufacturers issue detailed placement instructions, which should be mandatory IMHO.
I am not sure there is much science behind any of those recommendations. Half the time they are boilerplate writing anyway. They want to show us reality, give us measurements and why they recommend what they recommend. Not that some designer in his ad-hoc testing thought it sounded better one way or the other.
 

nathan

Active Member
Joined
May 24, 2020
Messages
166
Likes
122
The reference axis ("on-axis") for these speakers should be 30 degrees horizontal.
Since these are wall mounted speakers there is no toe-in possible.

Btw, I feel this should be standard practice with regular free standing speakers too to find the optimal horizontal axis.
Some speakers are designed for full toe-in while other speakers are designed to point straight forward and should be measured under 30 degrees horizontal.

For what it's worth I'm all for designing speakers that are not meant to be toed-in. You get a better power response and more natural reflections in the room that way which results in better sound quality.

Hmmm... I guess the raw on and off axis measurements will let you ferret out this information if you want. Personally I'm not a fan of speakers designed specifically to be used NOT on axis, since my center speaker is always on axis, and I use the same kind of speaker for all positions in my listening room.

The CEA 2034 standard is pretty clear about standardizing things, which seems to be the basis (but not the full extent) of testing on ASR. Thats good and bad. Good in that it lets you compare speakers in a similar specifically defined room and placement. Bad in that it doesn't actually represent many specific real rooms etc so there is still an X factor. (For example: Got a lot of side wall absorption panels? You're breaking CEA2034 which expects normal domestic USA drywall.) But CEA2034 tends to be a better indicator of how a particular speaker will sound in a regular domestic living room than just about anything else. And if one wants to ignore it because it's not how one uses speakers in ones own room, luckily ASR also publishes the data behind the data.

---

Anyway, I disagree with the notion that in wall speakers can't be toed in. Lots of us home theater and surround sound buffs build baffle walls and have each speaker towed in similar to how the center speaker is...ie, on axis with the primary listening position. There are a plethora of examples but to choose one almost random example from the YouTubes, look at the HT Gurus channel, where everything he designs and builds is done that way. Looks at Anthony Grimani, Dennis Erskine, Nyall at Acoustic Frontiers, this is the way they use in wall speakers, as well. These are some of the industry leaders and their design choices get aped all over the place by other designers and by DIY folks.
 

nathan

Active Member
Joined
May 24, 2020
Messages
166
Likes
122
Some (not many) manufacturers issue detailed placement instructions, which should be mandatory IMHO.
Most rooms will "play" more than the speakers, especially if the speakers are placed "at random".

I can see the argument that for subjective listening one could or should try to follow the manufacturer guidelines. Maybe not for all the subjective listening but for at least a segment.

For testing and measuring, it sort of doesn't matter in that the tests have so much data that any room placement can be modeled from the data, and anyway, testing with additional variables like placement just means one isn't doing science at all anymore (too many variables changing at once).
 

respice finem

Addicted to Fun and Learning
Joined
Feb 1, 2021
Messages
993
Likes
1,371
For testing and measuring (speakers) there are very precise conditions (as for the setup used by ASR) - what I mean is getting the most of the speakers potential in my listening room, and this response can be measured, too, however in a different way (and will always be more or less different than "anechoic" or "calculated anechoic" because of the room response playing in).
Reading the above link helps to get my point, which isn't about testing at all.
With wrong or unprecise positioning of speakers in any listening room, the SQ will more or less suffer.
 

nathan

Active Member
Joined
May 24, 2020
Messages
166
Likes
122
For testing and measuring (speakers) there are very precise conditions (as for the setup used by ASR) - what I mean is getting the most of the speakers potential in my listening room, and this response can be measured, too, however in a different way (and will always be more or less different than "anechoic" because of the room response).
With wrong or unprecise positioning of speakers in any listening room, the SQ will more or less suffer.

Luckily because ASR measures thousands of positions, you can derive that information, if you are interested. That is exactly what CEA2034 is:

Take the anechoic measurements of a speaker, and transform them into "in room" measurements. The room is precisely defined in terms of size, shape, resonance, reflections, speaker position relative to listener, boundaries, etc. So you don't actually need a physical room to measure in. That is how reliable the translation of detailed anechoic measurements is.

It sounds like what you want is to see the calculated in room response in something other than the default room used for CEA 2034. I don't blame you, and in theory its "just" a math problem once one defines the alternate room & placement that one cares about.

Or maybe I misunderstand what you are describing or seeking???
 
Top Bottom