• 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!

Panel vs Cones and Domes

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
I've posted this picture (or one like it) before:

Conditions:

Red - JBL LSR 308
Blue - MartinLogan reQuest
UMIK-1 at listening position, both (L/R) speakers playing a mono signal, no smoothing
(ignore the level difference)

upload_2017-7-2_16-25-16.png


What interests me is the amount of "hair" on the measurement of the JBL vs the ML. Many more and deeper dips in the JBL response.

I assume it is the influence of reflections in my less-than-anechoic room.

But I noticed this, using the ML, when measuring on and off-axis, at three positions across the top of the sofa:

Blue: ML at listening position (centered between the speakers)
Green: ML at about three feet to the right and left of the listening position.

upload_2017-7-2_16-40-5.png


So...

What conclusion can I draw from this observation?
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
I think you are right about the room reflections.

About the off-axis ML's I would think you are getting lesser level from the panels directly and more reflections.

I need to try it again, but I was measuring some microphones one vs another. I stuck up a 305 on a stand because it was handy. Later I noticed it has less grass (remembering your posts from the panels vs cones in the past). I had the LSR305 at a corner where two rooms intersect. So the wall behind the speaker bends 90 degrees away from the speaker. Down a hall on one side and into a larger room on the other. This means there are no surfaces for reflections until much further away. Basically imagine the regular inside corner of the room. Instead now think of having the speaker on the outside of that corner rather on the inside. The mics were cardioids and the speaker had no near side or rear reflecting surfaces. Reflections had to bounce back from walls in front of the speaker other side of the room. The graph was similar to your ML graph. Very little grass in the shape of it. So I don't think the grass is really from the speaker itself.

But I need to try that again, and see how it compares inside a corner vs outside.
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
About the off-axis ML's I would think you are getting lesser level from the panels directly and more reflections.

I think it is cancellation due to timing (phase) of the direct sound, as two speakers play, at different distances, when off-axis.

I'll look at the impulse response. If there are more reflections, it would show up there.

But, I don't think the grass is from the speakers themselves, either.
 
Last edited:
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
Martin Logan Far left Impulse Response, using the left as the timing reference, non-normalized:

Shows the impulse of the left and right (delayed due to greater distance).

upload_2017-7-2_17-40-21.png


And with the Central impulse response overlaid:

upload_2017-7-2_17-44-37.png
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
Okay that makes sense. Do you have any of those for both speakers where you only tested one speaker rather than stereo pairs?
 

fas42

Major Contributor
Joined
Mar 21, 2016
Messages
2,818
Likes
191
Location
Australia
My conclusion is that the ear/brain does a remarkably good job of correlating what it hears three feet away, as compared with the "ideal" listening position - otherwise, it should come across as being from another universe, in terms of what the graphs show. So, again, what was that point about having a "perfect" FR at some position in the room ... ?
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
Do you have any of those for both speakers where you only tested one speaker rather than stereo pairs?

Not for off-axis. At the time I was interested in seeing to what I was listening, which is two speakers.

I have single speaker at the listening position, though. Currently, I tend to measure left/right/both.

The DRC software measures L/R and L again... I initially assumed it looks at both together, but later realized it doesn't.

---

Oh, interesting, hadn't noticed this. Single speakers (L or R MartinLogan) each have more hair than both at once:

upload_2017-7-2_18-38-45.png
 
Last edited:
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
My conclusion is that the ear/brain does a remarkably good job of correlating what it hears three feet away, as compared with the "ideal" listening position - otherwise, it should come across as being from another universe, in terms of what the graphs show. So, again, what was that point about having a "perfect" FR at some position in the room ... ?

I see three choices available to me here for the sweet spot, a.k.a. "some position in the room":

1. Leave it alone
2. Make it worse.
3. Try to make it better.

I gravitate toward the third choice as being the more interesting option.
 
Last edited:

Sal1950

Grand Contributor
The Chicago Crusher
Forum Donor
Joined
Mar 1, 2016
Messages
14,068
Likes
16,598
Location
Central Fl
What interests me is the amount of "hair" on the measurement of the JBL vs the ML. Many more and deeper dips in the JBL response.

I assume it is the influence of reflections in my less-than-anechoic room.
Take one of the JBL's outside, measure, and see if it gets a shave?
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
Take one of the JBL's outside, measure, and see if it gets a shave?

I don't listen indoors from outside.

I have been known to listen outdoors - with an open window - to the indoor speakers, though. No measurements were taken.
 
Last edited:

Don Hills

Addicted to Fun and Learning
Joined
Mar 1, 2016
Messages
708
Likes
464
Location
Wellington, New Zealand
There are two good ways to measure speakers outdoors.
Either mount them on the top of a 30 foot pole, or bury them in the ground facing up with the baffle flush with the ground. The second method doesn't work so well with planars and designs like the Kii Three.
 

jhaider

Major Contributor
Forum Donor
Joined
Jun 5, 2016
Messages
2,822
Likes
4,514
What conclusion can I draw from this observation?

Not much, yet.

That's a great first start, but it looks like you only took your measurements at one point. That's fine if you're in the relative nearfield and gating the impulse appropriately, but if you're measuring a room response you need to either take a spatial average or a moving-mike measurement, because in the statistical frequency band results can vary dramatically with placement. Even if you are just measuring for one listening spot!

My usual process is based on a procedure from an old Geddes and Blind AES paper. For a single spot I usually measure six points: "middle of the head" and five further points in a rough ellipse around that, roughly corresponding to an arm-span.

Some examples of this approach in action may be found in the bench test sections of the following reviews:

http://hometheaterhifi.com/reviews/...rantz-av7702-surround-sound-processor-review/

http://hometheaterhifi.com/reviews/receiver-processor/receivers/denon-avr-x4100-a-v-receiver-review/

http://hometheaterhifi.com/reviews/...olution-xt-8f-floor-standing-speakers-review/
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
That's a great first start, but it looks like you only took your measurements at one point.

It is a single-point measurement, at, or just behind, the center of my head, if it were there, instead of the mic.

That's fine if you're in the relative nearfield and gating the impulse appropriately, but if you're measuring a room response you need to either take a spatial average or a moving-mike measurement, because in the statistical frequency band results can vary dramatically with placement. Even if you are just measuring for one listening spot!

Here's six spots across the couch averaged (not including the sweet spot) and the sweet spot measurement from last August:

upload_2017-7-3_22-30-44.png

My DRC uses single-spot measurement, I critically listen at a single spot, when two people are on the couch we don't hear anything terribly wrong, except a slightly skewed stereo image. Casual listening elsewhere in the room doesn't make me think anything is broken.

The 40-50Hz dip is a phase anomaly, which I attribute to my asymmetrical room. It looks severe, but doesn't seem to be bothersome. I haven't tried sub-repositioning yet.

Single speaker measurements have only a hint of that dip. Left, Right, and Both speakers (mono signal):

upload_2017-7-3_23-21-45.png

As for frequency response variations, as you move away from the "sweet spot", I investigated that somewhat in this little video. Pink noise is playing, and then the right speaker is delayed in 0.02ms increments simulating moving the microphone 6.8mm at a time. I stopped at 170mm as the effect was both regular and clearly displayed:

https://www.screencast.com/t/KenwrJj4WGA

From a practical perspective, I'm OK with my feeble methods and results.
 
Last edited:

jhaider

Major Contributor
Forum Donor
Joined
Jun 5, 2016
Messages
2,822
Likes
4,514
Here's six spots across the couch averaged (not including the sweet spot) and the sweet spot measurement from last August:

View attachment 7706

Thanks for that, but that's a little bit different than what I suggested would be useful: a spatial average of random points centered on the listening position.

Stand at your current listening position measurement, draw a circle-ish with your index finger, and put the microphone at a few points on the rim of that circle. Then average them in REW.

My DRC uses single-spot measurement,

Which one is that?

ARC, Dirac, RoomPerfect, Harman ARCOS, Audyssey, and Trinnov all rely on multiple points to derive their transfer functions. A DRC system can't "know" the room with data at just one point.
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
Which one is that?

AcourateDRC and a MiniDSP OpenDRC-DI

I tried a Dirac Live demo - where the PC does convolution. I didn't care for the results when measured.

There is the possibility of "operator error", but, overall, I just didn't like it.

I could have bought the MiniDSP with Dirac (convolution), but it was about $500 more, and the Dirac version disables the use of any other filter generators. Being experimentally inclined, I didn't bite. That was three years ago.

I started with rePhase (very manual) to generate filters, learned something, then spent $85 on the AcourateDRC.
 
Last edited:

jhaider

Major Contributor
Forum Donor
Joined
Jun 5, 2016
Messages
2,822
Likes
4,514
Thanks, I'd never heard of that one. Was your dissatisfaction with Dirac based on a calculation single-point measurement or a spatial average of measurements around the listening spot? If so, I expect you'll get more accurate correction doing that, though you may prefer the sound of the current filters.

From experience seeing large variances from small mike movements, especially while measuring speakers with very constant directivity, I just don't trust any correction based on a single-point measurement. I also don't trust any in-room measurement taken at a single point. You'll want to do a spatial average as I outlined above, or a moving mike measurement*, if you want to validate the results of the correction.

*http://www.ohl.to/audio/downloads/MMM-moving-mic-measurement.pdf
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
I may have a problem with the "spatial average" if it is an arithmetic mean of logarithmic values.

(80 + 70 + 60) / 3 = 70

Whereas, the average of 80dB, 70dB and 60dB is 75.68dB (if I did somebody's math correctly)

upload_2017-7-5_2-23-34.png
 
OP
RayDunzl

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
Looking at REW, it does seem to calculate an average correctly...

Yellow (dashed) perfectly overlays Black - Yellow being the exported REW average of 6 locations, and Black being laboriously calculated.

upload_2017-7-5_3-26-36.png
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Thanks, I'd never heard of that one. Was your dissatisfaction with Dirac based on a calculation single-point measurement or a spatial average of measurements around the listening spot? If so, I expect you'll get more accurate correction doing that, though you may prefer the sound of the current filters.

From experience seeing large variances from small mike movements, especially while measuring speakers with very constant directivity, I just don't trust any correction based on a single-point measurement. I also don't trust any in-room measurement taken at a single point. You'll want to do a spatial average as I outlined above, or a moving mike measurement*, if you want to validate the results of the correction.

*http://www.ohl.to/audio/downloads/MMM-moving-mic-measurement.pdf

Dirac does take multiple measurements. My experience was that it was effective, but one thing you had to do was get the initial central measurement location absolutely spot on.

Acourate does use a single measurement spot and works very well. In fact it is the best correction I have come across, but it does use psycho-acoustic correction. If you try to correct the high frequency variations as presented with a high resolution single measurement you will be in trouble due to the variations you mention, not so much of an issue at lower frequencies.
 
Top Bottom