Genelec 1032A Measurements (including spinorama)

[hardisj]

Also anything past 100° gets erratic really quickly, I'm not sure what's up with that and if it's a limitation of the ground plane method or if I'm doing something wrong (although I'm not sure what I could be doing wrong).

I see the issue. Look at your IR. Your 0 - 100 degree IRs all start at ~0ms with the window appropriately set to 15ms. All your >=110° start -50ms and have the window placed in the wrong spot. Not a big deal, though. Easy to fix.

Also, make sure to note that not every measurement will be reflection free for the same amount of time. For instance, let's say you have a wall 20 meters to the right of the speaker. At 0 degrees, 15 ms may be adequate because you're at least that far away. But as you close in on 90 degrees to the right of the speaker, you're 18 meters from the side wall now because your microphone is 2 meters to the right of the speaker. You'll have to adjust each window individually to make sure you take care of that issue.

Here are your results and my notes with corrections:

0 Degrees
FR:

IR:

Notice in your 0 degrees those ripples? That is an indication that a reflection or something else has creeped in to your measurement. Looking at your IR, you can see your windows are not set up properly.

Move the left side in like so:

now you have this:

Which fixes those LF ripples (errors). Original = red. New (fixed) = blue.

Do this for ALL your measurements.







110° Measurements:

Similar situation. The windows need to be corrected.

started with this:

Adjust the IR Window to begin at the initial impulse and end at the first reflection:

You can see that if I keep the window time of 15ms, there is still a strong reflection at about -40ms

Shifted the IR window but shorted it as well to 11ms

and now you have the new response curve that is much more smooth (thanks to the window being placed at the right spot):

Green = original
Blue = corrected

That's it. Pretty simple. Just tedious.

Hopefully that helps. I kind of rushed through this to show you what's up but should give you enough of a clue to correct the others.

Note, also you'll be better served to adjust the starting time but I didn't do that here because it might confuse you in this first cut.

 

[TimVG]

I see the issue. Look at your IR. Your 0 - 100 degree IRs all start at ~0ms with the window appropriately set to 15ms. All your >=110° start -50ms and have the window placed in the wrong spot. Not a big deal, though. Easy to fix.

Also, make sure to note that not every measurement will be reflection free for the same amount of time. For instance, let's say you have a wall 20 meters to the right of the speaker. At 0 degrees, 15 ms may be adequate because you're at least that far away. But as you close in on 90 degrees to the right of the speaker, you're 18 meters from the side wall now because your microphone is 2 meters to the right of the speaker. You'll have to adjust each window individually to make sure you take care of that issue.

Here are your results and my notes with corrections:

0 Degrees
FR:

View attachment 93096

IR:
View attachment 93098


Notice in your 0 degrees those ripples? That is an indication that a reflection or something else has creeped in to your measurement. Looking at your IR, you can see your windows are not set up properly.

Move the left side in like so:

View attachment 93100

now you have this:

View attachment 93101

Which fixes those LF ripples (errors). Original = red. New (fixed) = blue.

View attachment 93102



Do this for ALL your measurements.







110° Measurements:

Similar situation. The windows need to be corrected.

started with this:

View attachment 93104

View attachment 93105




Adjust the IR Window to begin at the initial impulse and end at the first reflection:

You can see that if I keep the window time of 15ms, there is still a strong reflection at about -40ms

View attachment 93107


Shifted the IR window but shorted it as well to 11ms
View attachment 93108

and now you have the new response curve that is much more smooth (thanks to the window being placed at the right spot):

Green = original
Blue = corrected

View attachment 93109







That's it. Pretty simple. Just tedious.

Hopefully that helps. I kind of rushed through this to show you what's up but should give you enough of a clue to correct the others.

Note, also you'll be better served to adjust the starting time but I didn't do that here because it might confuse you in this first cut.

Thanks Erin for going over this! I'll be sure to check out the individual measurements. I'll see if I can do round two tomorrow to figure out some of the other issues as well.

Huge amount of respect for you, napilopez and amir (and probably others) who pump out reviews on a consistent basis, it's a whole lot of work.

 

[napilopez]

Per what @hardisj pointed out above and the discussion in the other thread, I went and fixed the timing for the messy rear hemisphere data, which was because REW picked the wrong reference timing for the impulse response. With that fixed, we get a much more reasonable looking spin in the bass region (1/24 smoothing).

The low-mid hump remains but a more realistic spin overall.


Here are the ER horizontal component curves, as well as the Horizontal ERDI:

And the verticals:

Also, I hadn't shown it before, but the normalized polar plots look much better too. Horizontal:

Vertical:

Not normalized looks much the same:

Vertical:

 

[TimVG]

Per what @hardisj pointed out above and the discussion in the other thread, I went and fixed the timing for the messy rear hemisphere data, which was because REW picked the wrong reference timing for the impulse response. With that fixed, we get a much more reasonable looking spin in the bass region (1/24 smoothing)

View attachment 93124

The low-mid hump remains but a more realistic spin overall.


Here are the ER horizontal component curves, as well as the Horizontal ERDI:

View attachment 93136And the verticals:
View attachment 93135
Also, I hadn't shown it before, but the normalized polar plots look much better too. Horizontal:

View attachment 93128

Vertical:
View attachment 93137


Not normalized looks much the same:

View attachment 93129

Vertical:
View attachment 93138

Wow, thanks for going over all of this!

 

[TimVG]

Lifting the speaker a little short of 6ft off the floor (on a column though) gave the following results. Slightly better overall. Less high frequency energy and the dip at 750 got filled in. The 1,5kHz peak remains so it must be there, as it isn't in other speakers (next post). The rising response at 200Hz also remains.

Zooming in

 

[TimVG]

Testing out the (small and easy to handle) Behringer 1030 with different mic setups using the GP method yields no mentionable differences between 100Hz and 2kHz

Compared to free standing however we do see a difference of about 1dB in the upper bass and a bit more at 200Hz.

In the treble range keeping the mic flat (with a small between the floor and the capsule) seems to give the correct amount of HF energy compared to angling it down on a stand

 

[TimVG]

Apart from an artifact at 4kHz I'm starting to get decent correlation between Genelec's 1032C measurement and my own.

 

[TimVG]

Unfortunately this method brings about other issues (especially off axis) around that 4kHz point. Seems like applying some mild EQ to the original spinorama made by @napilopez below 1kHz should give us the best approximating of what is actually going on.

Looks like I'll have to build myself a measurement platform.

 

[TimVG]

Phase 2

 

[andreasmaaan]

@TimVG, just to clarify, there's no need to elevate the speaker as much as 6' off the floor (although there's certainly no problem with doing that, as long as the speaker is angled towards the mic). But I'd say 1' or 2' should be adequate.

 

[TimVG]

Well this room has a height of 13ft and I already have the stand ready..

 

[andreasmaaan]

Well this room has a height of 13ft and I already have the stand ready..

View attachment 93315

 

[hardisj]

Looks good. Yea, as long as you are sufficiently away from the ceiling and floor. The first reflection is all that matters; whether that comes from the floor or ceiling, doesn't matter. I'm sure you already know this but I just want to make certain.

But, yep... that's great. That extra couple ms will be quite useful. I would even encourage you to use these for your mid-to-high frequency response and use the ground plane for your low-to-mid. Merge the two. It'll be more of a pain in the butt, BUT the data will be more accurate and will be less prone to questioning. Even VCAD has the ability to merge the two measurements together (you'd substitute your GP measurement for VCAD's "nearfield" measurement, in this case).


Also, make sure your center of rotation is about the front baffle. I know some will use the center of the speaker but it needs to be the baffle (per the CEA-2034 standard, iirc). That complicates things as it means shifting the weight of the speaker on the stand off-center. But, I would encourage you to do that.

 

[TimVG]

Also, make sure your center of rotation is about the front baffle. I know some will use the center of the speaker but it needs to be the baffle (per the CEA-2034 standard, iirc). That complicates things as it means shifting the weight of the speaker on the stand off-center. But, I would encourage you to do that.

That's something I considered but alas I can't do it with the stuff available to me at the moment but it's something I am planning on making.

Looks good. Yea, as long as you are sufficiently away from the ceiling and floor. The first reflection is all that matters; whether that comes from the floor or ceiling, doesn't matter. I'm sure you already know this but I just want to make certain.

That's what I assumed.

I would even encourage you to use these for your mid-to-high frequency response and use the ground plane for your low-to-mid. Merge the two. It'll be more of a pain in the butt, BUT the data will be more accurate and will be less prone to questioning. Even VCAD has the ability to merge the two measurements together (you'd substitute your GP measurement for VCAD's "nearfield" measurement, in this case).

I assume you mean GP outside? I'm getting a weird hump in the upper bass it seems using the GP method indoors. Or, the speaker simply isn't flat there.

 

[hardisj]

I assume you mean GP outside? I'm getting a weird hump in the upper bass it seems using the GP method indoors. Or, the speaker simply isn't flat there.

Yes. If possible, that would be idea. As long as you are >40 feet away from a nearby building/surface you should be able to get 20-600Hz with high accuracy and merge that to your farfield measurements so that you have high resolution in the midrange. That'll show high-Q resonances that typical quasi-anechoic measurements won't show.

Or, you could cut things short, do a single GP outdoors and overlay that with your indoor measurements gated to 6ms. It won't be perfect, but it'll be sufficient for a DIY'r until you are able to get time/patience to do the full suite of GP SPIN measurements.

 

[Biblob]

Also, make sure your center of rotation is about the front baffle. I know some will use the center of the speaker but it needs to be the baffle (per the CEA-2034 standard, iirc). That complicates things as it means shifting the weight of the speaker on the stand off-center. But, I would encourage you to do that.

You have any tips for making this section on a stand? I have a cut a IKEA lazy susan, so I could twist it, but it's not very stable on a stand. I would worry the speaker could tumble of it.

 

[TimVG]

Well 0 to 90 horizontal is looking better for sure.

Vs before

 

[hardisj]

You have any tips for making this section on a stand? I have a cut a IKEA lazy susan, so I could twist it, but it's not very stable on a stand. I would worry the speaker could tumble of it.

I built the lazy susan out of some 2x2 ft sections of ¾” plywood. I built a 4 foot tall stand out of plywood. I placed the lazy susan offset from the center and then placed the stand on top of the lazy susan so that the baffle was aligned to the center of rotation of the lazy susan. The stand is held in place by a large concrete paver placed at the bottom so that it won’t rock or fall over. It works. And is quite simple.

This is a photo of that setup:

The problem is the verticals of tower speakers. Once you lay the speaker on its side with the measurement axis being at the top, then I have to use scrap wood to build a support leg with a roller wheel on the bottom to hold the other end up and still be able to spin freely.

BUT that sucked. So, I went and bought this scaffolding which puts the speaker up about 6 feet off the ground. I took my lazy susan and added some extensions to it so that I could fix the center of rotation wherever I wanted it to be; clamping it to the scaffolding legs. The scaffolding has wheels, which I unlock and then the lazy susan is what keeps the center of rotation in check. The concrete paver holds it in place. This works EXTREMELY well. Best method I've come up with to date. SUPER easy to rotate the tower as well.

Obviously, the scaffolding can't be used in the garage. So, I kind of just pick and choose the method I want to use depending on weather and what I am measuring.

I made this video just to give a demonstration. Skip toward the end because you don't care about all my yapping.

 

[TimVG]

here's the updated .mdat file in case @napilopez wants to have another go.

https://we.tl/t-7PYMQwOJV4

 

[Biblob]

I built the lazy susan out of some 4x4 ft sections of ¾” plywood. I built a 4 foot tall stand out of plywood. I placed the lazy susan offset from the center and then placed the stand on top of the lazy susan so that the baffle was aligned to the center of rotation of the lazy susan. The stand is held in place by a large concrete paver placed at the bottom so that it won’t rock or fall over. It works. And is quite simple.

This is a photo of that setup:
View attachment 93344




The problem is the verticals of tower speakers. Once you lay the speaker on its side with the measurement axis being at the top, then I have to use scrap wood to build a support leg with a roller wheel on the bottom to hold the other end up and still be able to spin freely.

BUT that sucked. So, I went and bought this scaffolding which puts the speaker up about 6 feet off the ground. I took my lazy susan and added some extensions to it so that I could fix the center of rotation wherever I wanted it to be; clamping it to the scaffolding legs. The scaffolding has wheels, which I unlock and then the lazy susan is what keeps the center of rotation in check. The concrete paver holds it in place. This works EXTREMELY well. Best method I've come up with to date. SUPER easy to rotate the tower as well.

Obviously, the scaffolding can't be used in the garage. So, I kind of just pick and choose the method I want to use depending on weather and what I am measuring.

View attachment 93343

View attachment 93342

I made this video just to give a demonstration. Skip toward the end because you don't care about all my yapping.

Thanks, I might steal some of these ideas to test for my own
The video is not accesible for me though, it's 'private'.