• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Can we discuss the BMR Tower?

alexis

Member
Audio Company
Joined
Jun 3, 2020
Messages
97
Likes
159
@alexis Every software applies more smoothing when doing gated measurements as you get closer to gate frequency/time. With gate of 5-6ms you get something like 1/6 octave smoothing below 1kHz. If there is need for high resolution measurements (and there is at least when you are developing a loudspeaker) you need to get to at least 1/24 octave smoothing.
I applied "no smoothing" to OmniMic when I take measurement. I take it at 1 meter at the center of the tweeter with gating at 5.3ms. I'd get high accuracy at 1kHz. Both Dennis and I don't use smoothing when measuring speakers. I used the same setting for QCing all BMRs and Towers.
 

Zvu

Addicted to Fun and Learning
Joined
May 1, 2020
Messages
830
Likes
1,416
Location
Serbia
I applied "no smoothing" to OmniMic when I take measurement. I take it at 1 meter at the center of the tweeter with gating at 5.3ms. I'd get high accuracy at 1kHz. Both Dennis and I don't use smoothing when measuring speakers. I used the same setting for QCing all BMRs and Towers.

Read what you quoted, please. I never wrote that you did any smoothing. Software does it automatically whether you want/know of it or not.
 
Last edited:

Zvu

Addicted to Fun and Learning
Joined
May 1, 2020
Messages
830
Likes
1,416
Location
Serbia
I do appreciate that your comment is indeed constructive. Please don't take my posts as an unreasonable defense.

James Larson said in his review:


On AH, he also was asked about the 600 Hz bump. He replied:

This does suggest that the various graphs showing the 600 Hz bump have a greater visual impact than it's sound does to at least one listener.

All in all, this review reveals what seems to be a wonderful speaker. With that out of the way, bring on those cheap shots from the peanut gallery ;) :rolleyes::
• That it can't be driven by 5 watt amp
• That the grills look like they were mounted wrong
• That it suffers from the lack of controlled dispersion

Thanks for that quotes. People develop different levels of sensitivity on nuances like that depending on their music of choice and hearing. We have Erin's last video where he corrected Kali LP6-V2 just 1dB at 8kHz and got more "air".

Good thing in all this is - James measured it and we see what is happening. It can be easilly EQ-ed in software based on Larson's measurements and everyone who buys it can judge for himself if it is perceivable to him or not.

So good stuff after all :)
 

Dennis Murphy

Major Contributor
Technical Expert
Joined
Mar 17, 2020
Messages
1,071
Likes
4,535
Read what you quoted, please. I never wrote that you did any smoothing. Software does it automatically whether you want/know of it or not.
Point taken--you didn't claim anyone was applying smoothing. I'm still curious as to exactly what happens in the Praxis and OmniMic measurements that incorporate a transition to room mode as you descend below 800 Hz. I'm not sure smoothing is invoked, but I just don't know.
 

Zvu

Addicted to Fun and Learning
Joined
May 1, 2020
Messages
830
Likes
1,416
Location
Serbia
I am using HolmImpulse, ARTA and REW and all three apply certain amount of smoothing as you get closer to gate frequency. In REW you can choose between window functions.

Never used Omnimic but i guess that information should be somewhere in the manual.
 

alexis

Member
Audio Company
Joined
Jun 3, 2020
Messages
97
Likes
159
Read what you quoted, please. I never wrote that you did any smoothing. Software does it automatically whether you want/know of it or not.
OmniMic has a "No smoothing" setting. I just need to trust Bill knows what he is doing as he wrote the program.
 

Zvu

Addicted to Fun and Learning
Joined
May 1, 2020
Messages
830
Likes
1,416
Location
Serbia
OmniMic has a "No smoothing" setting. I just need to trust Bill knows what he is doing as he wrote the program.

In God you can trust. Everyone else you verify.

Here's how you can check what your software actually does without inclusion of beliefs, fate and religion.

Open one of your measurements with gate good down to 6ms. Look at frequency response above gate frequency.
Nof move the gate at 2ms and observe frequency response above the gate frequency.

If the shape of the curve above gate frequency changed, software aplies smoothing and is working as almost all other software available. The lower you go, the smoothing is more substantial. That's why you see so much factory loudspeakers with problems in midrange. It is hard to measure accurately down there. And that's why it is so important what Amir and Erin are doing. They provide us with a clear view in loudspeaker performance that almost all manufacturers (two or three aside) keep hidden.

Here is measurement ofKef Q100 midbass with 11ms gate and 4ms gate

11 мили секунди.png 4ms.png

Observe the differences above gate frequency (blue arrow on frequency axis). Where do you think the differences come from and why is even above 1kHz response affected, being much above 250Hz (4ms gate time) ?

HolmImpulse is software Earl Geddes used to obtain frequency response measurements and then imports that into his software to calculate directivity and sound power curve. He used it because math behind the software is spot on.
 
Last edited:

alexis

Member
Audio Company
Joined
Jun 3, 2020
Messages
97
Likes
159
In God you can trust. Everyone else you verify.

Here's how you can check what your software actually does without inclusion of beliefs, fate and religion.

Open one of your measurements with gate good down to 6ms. Look at frequency response above gate frequency.
Nof move the gate at 2ms and observe frequency response above the gate frequency.

If the shape of the curve above gate frequency changed, software aplies smoothing and is working as almost all other software available. The lower you go, the smoothing is more substantial. That's why you see so much factory loudspeakers with problems in midrange. It is hard to measure accurately down there. And that's why it is so important what Amir and Erin are doing. They provide us with a clear view in loudspeaker performance that almost all manufacturers (two or three aside) keep hidden.

Here is measurement ofKef Q100 midbass with 11ms gate and 4ms gate

View attachment 163086 View attachment 163085

Observe the differences above gate frequency (blue arrow on frequency axis). Where do you think the differences come from and why is even above 1kHz response affected, being much above 250Hz (4ms gate time) ?
OmniMic doesn't change anything above gating frequency when I change the gate timing and the result is very consistent.
 

Zvu

Addicted to Fun and Learning
Joined
May 1, 2020
Messages
830
Likes
1,416
Location
Serbia
OmniMic doesn't change anything above gating frequency when I change the gate timing and the result is very consistent.

If there is an option to show resolution as i have in Holm (20-25dB grid height in 1-2dB increments) i'd gladly examine your screenshots.

Is there an option in OmniMic to show 1dB or 2dB grid height resolution or is 5dB the best it can do ?
 

tktran303

Addicted to Fun and Learning
Forum Donor
Joined
Mar 27, 2019
Messages
683
Likes
1,178
Hello Dennis and alexis,

Mali is right, the gating affects the lowest frequency that can be shown (as shown by Zvu's blue marker), but don't forget that it also affects the resolving ability or granularity of the frequency measurements. So a 10ms gate means the frequency response resolution is 100Hz, and 4ms gate is 250Hz.

This is where the Klippel has the distinct advantage. With the transfer function measurement module and the in-situ compensation module you can get fine granularity to 20Hz. So you can see all the high-Q resonances in that region in the midrange to lower treble; 500-2KHz. Now 500Hz to 2000Hz is two octaves in musical speak, but in math speak it is a span of 1500Hz. So with a 4ms gate, for example, the granularity is 250Hz. Between 500Hz to 2000Hz, that gives you 6 data points.

That is why little blips in the important midrange around 500-2KHz can be missed in conventional gated measurement, if your indoor measurement allows a gate of around 3-5ms.

Mali can measure with a 10ms gate, but the weather's not conducive to doing that where I lived.
I really should bite the bullet and find an unused basketball court, or buy a Klippel.
 
Last edited:

Dennis Murphy

Major Contributor
Technical Expert
Joined
Mar 17, 2020
Messages
1,071
Likes
4,535
Hello Dennis and alexis,

Mali is right, the gating affects the lowest frequency that can be shown (as shown by Zvu's blue marker), but don't forget that it also affects the resolving ability or granularity of the frequency measurements. So a 10ms gate means the frequency response resolution is 100Hz, and 4ms gate is 250Hz.

This is where the Klippel has the distinct advantage. With the transfer function measurement module and the in-situ compensation module you can get fine granularity to 20Hz. So you can see all the high-Q resonances in that region in the midrange to lower treble; 500-2KHz. Now 500Hz to 2000Hz is two octaves in musical speak, but in math speak it is a span of 1500Hz. So with a 4ms gate, for example, the granularity is 250Hz. Between 500Hz to 2000Hz, that gives you 6 data points.

That is why little blips in the important midrange around 500-2KHz can be missed in conventional gated measurement, if your indoor measurement allows a gate of around 3-5ms.

Mali can measure with a 10ms gate, but the weather's not conducive to doing that where I lived.
I really should bite the bullet and find an unused basketball court, or buy a Klippel.
I'm not disagreeing with any of this, and I'll invest in a Klippel just as soon as I sell 1,000 more towers. However, I have had some experience comparing Amir's Klippel measurements with my own for the same speaker (the same physical speaker--my Pioneer BS22 mod and the New Advent Loudspeaker), and with measurements of one of my BMR's vs. those made in the NRC anechoic chamber, and while minor peaks and dips were a little sharper on the higher resolution plots, there wasn't really any new information that would help a designer. And, of course, all of this begs the issue of whether we can actually hear much of that. Plus, a nearfield of the port response would flag important issues. I did that with one of the BMR's sent to James, and I'll do it with the other when the Capital Audio Fest is over. Room 306. ( I used to hand out really good home made chocolate chip cookies until the hotel got wind of it and shut me down.)
 
Last edited:

Martin

Major Contributor
Forum Donor
Joined
Mar 23, 2018
Messages
1,895
Likes
5,536
Location
Cape Coral, FL
I'm not disagreeing with any of this, and I'll invest in a Klippel just as soon as I sell 1,000 more towers. However, I have had some experience comparing Amir's Klippel measurements with my own for the same speaker (the same physical speaker--my Pioneer BS22 mod and the New Advent Loudspeaker), and with measurements of one of my BMR's vs. those made in the NRC anechoic chamber, and while minor peaks and dips were a little sharper on the higher resolution plots, Thee wasn't really any new information that would help a designer. And, of course, all of this begs the issue of whether we can actually hear much of that. Plus, a nearfield of the port response flag important issues. I did that with one of the BMR's sent to James, and I'll do it with the other when the Capital Audio Fest is over. Room 306. ( I used to hand out really good home made chocolate chip cookies until the hotel got wind of it and shut me down.)

Will you be at the Florida Audio Expo?

Martin
 

tktran303

Addicted to Fun and Learning
Forum Donor
Joined
Mar 27, 2019
Messages
683
Likes
1,178
It's phenomenal that you managed to such a smooth and extended response out to almost 90 degrees Dennis. I bet it sounds great, even from another room.

I've sent you a PM re: an order.
 

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,340
Likes
688
Hey maybe I missed something, but how are these 90 dB? Aren't those BMR drivers like 52 dB sensitivity? OK I jest but all the ones I've seen have been very low. Frustratingly because (having once upon a time done R&D involving honeycomb panel speakers) I can believe the great off-axis response.

As for the looks, the BMRs being smaller than the tweeter is what looks peculiar. If they had some kind of wrap/cover over that area so you just saw the driver surfaces then it would be OK, as the cabinets have a nice curve and the finish seems nice in the pictures.
 

Dennis Murphy

Major Contributor
Technical Expert
Joined
Mar 17, 2020
Messages
1,071
Likes
4,535
Hey maybe I missed something, but how are these 90 dB? Aren't those BMR drivers like 52 dB sensitivity? OK I jest but all the ones I've seen have been very low. Frustratingly because (having once upon a time done R&D involving honeycomb panel speakers) I can believe the great off-axis response.

As for the looks, the BMRs being smaller than the tweeter is what looks peculiar. If they had some kind of wrap/cover over that area so you just saw the driver surfaces then it would be OK, as the cabinets have a nice curve and the finish seems nice in the pictures.
I'm not sure where the 90 dB figure is coming from. My measurements show more like 86 dB, and James measured 85.4. The BMR unit I use is rated at 86 dB. Baffle step compensation complicates things, since it's shared by the woofer and mid in this design. But I'm able to boost sensitivity with the second order acoustic crossover slopes, which provide more overlap between the woofer and midrange.
 

Dennis Murphy

Major Contributor
Technical Expert
Joined
Mar 17, 2020
Messages
1,071
Likes
4,535
Hey maybe I missed something, but how are these 90 dB? Aren't those BMR drivers like 52 dB sensitivity? OK I jest but all the ones I've seen have been very low. Frustratingly because (having once upon a time done R&D involving honeycomb panel speakers) I can believe the great off-axis response.

As for the looks, the BMRs being smaller than the tweeter is what looks peculiar. If they had some kind of wrap/cover over that area so you just saw the driver surfaces then it would be OK, as the cabinets have a nice curve and the finish seems nice in the pictures.
Forgot to address the cosmetic side of your post. Although I don't particularly care whether the mids are smaller or larger than the tweeter, we are currently working with the factory to get cosmetic rings over the BMR drivers. We'll see what the sample baffle looks like.
 

fluid

Addicted to Fun and Learning
Joined
Apr 19, 2021
Messages
691
Likes
1,196
Omnimic has different options which one is selected will determine what is displayed as so probably everybody is right from a certain point of view. All, blended, or only to circled in red below

Omnimic 1.png

This is what the manual has to say
"OmniMic can show impulse responses, and you can select how much of an impulse response you want to analyzer to look at when forming a frequency response. Click on the impulse response graph at the latest part of the impulse response you want OmniMic to include when it computes frequency response. The selected portion will appear as a red trace, the rest will remain black. The shape of the 'windowing' that tapers down the impulse response is shown in green or gray."

If you gate by using "only to" it should behave as Zvu has pointed out. If you use blended something else will happen as Dennis has shown.
 

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
About that 500Hz problem in Larson's measurements...

I can clearly see that happening in all his outdoor measurements, so it must be a methodological artefact. Most likely a reflection from something on the side - see how ridge is curved, it gets higher in freq when speaker is rotated.

Big and deep cabinets on the other hand have uneven horizontal radiation around 500Hz...


image_large2

image


Larson image-tile.jpg
 
Last edited:
Top Bottom