Spatial averaging of measurements for room EQ

[Daverz]

Up to now I've been using single point measurements at the listening position for my room EQ experiments, but now I'd like to try averaging of measurments taken from several points at the listening position.

I found this document (PDF)

http://forums.melaudia.net/attachment.php?aid=22240

He recommends taking measurements at the vertices and center of a parallelepiped (or rectangular cuboid, if you prefer).

https://upload.wikimedia.org/wikipe...Cuboid_simple.svg/480px-Cuboid_simple.svg.png

Above: measurement points for very blockheaded listener.

That's 9 measurements per speaker!

I'll try it if I have the time, after I figure out the easiest way to place the mic at the 9 points (perhaps it doesn't need to be a very precise selection of points.)

But does anyone have a simpler procedure that works for them?

 

[RayDunzl]

When you're done...

Please show a comparison of the single-point correction vs the multipoint described above.

 

[Blumlein 88]

This 9 point measure is what Dirac suggests you use. It really doesn't take all that much time. Dirac doesn't require extreme precision. Apparently using time delay etc. it figures out where those points are vs the first one. They don't quite use a a cube vertex pattern. Your center one. Then a row of 4 that are higher at the back and a row of 4 lower in the front.

What I've done is use a short microphone stand. You do the center. Adjust height, do the rear ones, adjust height and do the lower ones.

Here is a Roon pattern. A microphone stand with an arm would work fine for this.

 

[Daverz]

Here is a Roon pattern. A microphone stand with an arm would work fine for this.

View attachment 29791

Thanks, that last one looks the most relevant for my listening chair. Interesting that they cover the whole back of the chair rather than just a head shape.

One more question: mic up toward the ceiling (90 degrees) or pointed at the plane of the speakers (0 degrees)? Getting the mic close to the back of the chair is difficult with the 0 degree orientation.

When you're done...

Please show a comparison of the single-point correction vs the multipoint described above.

Will do.

 

[Rja4000]

One more question: mic up toward the ceiling (90 degrees) or pointed at the plane of the speakers (0 degrees)?

That's documented.
And microphone calibration is given accordingly.

I always do pointed to the speakers

 

[Daverz]

That's documented.

Sorry for being obtuse, but what documentation are we referring to? Thanks.

And microphone calibration is given accordingly.
I always do pointed to the speakers

As do I, usually, though this does not allow getting the mic close to the back of the chair. It's a recliner, so I usually push the back of the chair down so it's out of the way. However, this does not account for reflections from the leather (it's an Ekornes Stressless chair).

 

[Kal Rubinson]

I always do pointed to the speakers

Which speakers? Do you re-aim it for each speaker with each microphone position?

 

[Blumlein 88]

You should be using an omni. Suggested to point up. Some mics have both straight ahead and 90 degree calibration curves. Either should work.

 

[Daverz]

You should be using an omni. Suggested to point up. Some mics have both straight ahead and 90 degree calibration curves. Either should work.

Thanks. I have two omni measurement mics: a Umik-1 USB mic (both 0 and 90 degree calibration files) and an Audix TR-40 analog mic (no calibration files).

Which speakers? Do you re-aim it for each speaker with each microphone position?

If you meant the questions for me, Kal, speakers are a stereo pair of Vandersteen cloth Quatros. I point the mic at the centerpoint between the speakers at ear height.

 

[Cosmik]

I recommend a crude 256-point cuboid for preliminary trials, then a 4096 point spheroid for the intermediate tests. Eventually it has to be a minimum 8192-point cloud modelled on the listener's phrenological expansion. Or maybe none of the above means anything. Can you prove me wrong on any of my recommendations? Or anybody else's?

(post-pub post )

 

[nightfishing]

I use a laser measurer on my stand pointing at the listening position to line up the mic positions. I found it much easier and more accurate then trying to take measurements in the listening position. especially if you have cushions

 

[Daverz]

I recommend a crude 256-point cuboid for preliminary trials, then a 4096 point spheroid for the intermediate tests. Eventually it has to be a minimum 8192-point cloud modelled on the listener's phrenological expansion. Or maybe none of the above means anything. Can you prove me wrong on any of my recommendations? Or anybody else's?

(post-pub post )

I take exception, sir, to a jest that assumes I lack the dimensionality to access all 256 vertices of the 8-cube. 4096 points of a spheroid are too trivial for my consideration.

 

[Blumlein 88]

I recommend a crude 256-point cuboid for preliminary trials, then a 4096 point spheroid for the intermediate tests. Eventually it has to be a minimum 8192-point cloud modelled on the listener's phrenological expansion. Or maybe none of the above means anything. Can you prove me wrong on any of my recommendations? Or anybody else's?

(post-pub post )

Supply us with something to prove wrong. Empty conjecture doesn't need addressing.

 

[amirm]

You don't need any calibration file. Or worry about the direction of the mic. Measure first. Make corrections in low frequencies. Then listen to see if it sounds better.

 

[LumbermanSVO]

I recommend a crude 256-point cuboid for preliminary trials, then a 4096 point spheroid for the intermediate tests. Eventually it has to be a minimum 8192-point cloud modelled on the listener's phrenological expansion. Or maybe none of the above means anything. Can you prove me wrong on any of my recommendations? Or anybody else's?

(post-pub post )

I use APL Workshop to do my spacial averages, and it doesn't take long to do get 150-250 sweeps, maybe a minute or so.

 

[Daverz]

Is APL Workshop something to do with the APL programming language?

I suppose the next project is a robot arm to automate the mic placement.

 

[Kal Rubinson]

If you meant the questions for me, Kal, speakers are a stereo pair of Vandersteen cloth Quatros. I point the mic at the centerpoint between the speakers at ear height.

I did mean that. So, you are not on-axis to either one and I wonder about the regularity/symmetry of its off-axis response. I prefer having the mic pointing vertically but, of course, that is just as much subject to the same objection but, in my case, the speakers are 5 main speakers plus subs. There is no single centerpoint between them.

 

[LumbermanSVO]

Is APL Workshop something to do with the APL programming language?

I suppose the next project is a robot arm to automate the mic placement.

No, it's the company that makes the software.

 

[Hipper]

Dirac uses an algorithm when it measures the nine points. That I assume will not be the same as averaging nine points and using EQ.

I use EQ (a Behringer DEQ2496) but measure with REW and manually apply the filters. I have a single seat listening chair and my practice has been, recently, to measure up to 500Hz using a signal from both speakers with the mic pointing horizontally centred between the two speakers, with it positioned in line with my ears but in the centre of my head (previously I measured up to 500Hz each speaker separately, pointing directly to the speaker and positioned at the relevant ear).

I did some tests to compare the measurements of two different mics, and pointing vertically and horizontally, and with me positioned in different parts of the room and even in the chair. These different set ups make a little difference to the frequency responses but the key point I got from this was:

1. Use a consistent set up for measuring.
2. Use measurements as a guide, not the be all and end all of everything. In other words, don't get obsessed with them!
3. As amirm's post 14 said, 'Measure first. Make corrections in low frequencies. Then listen to see if it sounds better.'

I did this and am enjoying my music so much that I haven't got round to dealing with the higher frequencies (I should add that I also have lots of room treatment and have made a fair effort to get positioning as best I could).

 

[Krunok]

Up to now I've been using single point measurements at the listening position for my room EQ experiments, but now I'd like to try averaging of measurments taken from several points at the listening position.

I found this document (PDF)

http://forums.melaudia.net/attachment.php?aid=22240

He recommends taking measurements at the vertices and center of a parallelepiped (or rectangular cuboid, if you prefer).

https://upload.wikimedia.org/wikipe...Cuboid_simple.svg/480px-Cuboid_simple.svg.png

Above: measurement points for very blockheaded listener.

That's 9 measurements per speaker!

I'll try it if I have the time, after I figure out the easiest way to place the mic at the 9 points (perhaps it doesn't need to be a very precise selection of points.)

But does anyone have a simpler procedure that works for them?

I'm using MMM (moving microphone method) described here. In REW you do it as RTA measurement with pink noise.

Here is an example how it looks with my left speaker:

This was taken from the point at LP where app my left ear is. Red line is a sine sweep and green line is RTA pink noise. As you can see they are practically identical.

This is RTA where I moved mike in a horizontal circle with a diameter of 20cm (so around the area where my head would be):

This is RTA where I moved mike in a horizontal circle with diameter of 1m (so around the area where my sofa is):

This is a single point sweep taken app 5cm behind the first sweep (shown on first pic). As you can see it differs from first sweep quite a bit.

This is sofa spatial sweep vs left ear sweep, for comparison:

With average of 9 sweeps I was getting pretty close measurement to MMM RTA but I consider MMM RTA to better represent spatial average as it has 80+ samples instead of 9. It is also easier and faster for me to make it.