ARTA FR measurement problem

[donradick]

Friends,

I seem to have a new problem with ARTA FR measurements, I get deep nulls at 5K and 10K that aren't present in Holm measurements, so I know that my microphone (PE calibrated, never dropped) and my PC and amp are ok.

The pics below are the same exact speaker, same microphone position, same PC. I am not concerned about the low end response, only the nulls in the high frequency response.

Here is the ARTA FR


Here is the HOLM FR:

[DjDisturbed]

Re: ARTA FR measurement problem

Never seen that happen before and I use ARTA for measuring too.
What I do find odd about it is, the nulls are exactly an octave apart.
To me it looks to be be more around 6k and 12k ( roughly 2.26" & 1.13").
I mainly answered so if someone does know or you figure it out.
I can look back at this thread if I ever have the same problems :D .

[donradick]

Re: ARTA FR measurement problem

Thanks - I double checked my audio setup, rebooted, etc.
Windows 7 HP laptop with Blue Icicle used for microphone input.
Double-checked all inputs and outputs are 16 bit 44.1 Khz.

[charlielaub]

Re: ARTA FR measurement problem

Tell us about the geometry of the measurement setup (distances in room, to speaker, etc.), and what kind of measurement are you doing? It's helpful to know all the details explicitly. Did you happen to see my tutorial on FR measurements using ARTA?

Easy gated impulse frequency measurements using ARTA

I know that things (artifacts) like this happen with nearfield measurements when the distance from source to mic corresponds to some multiple of frequency or vice versa. Perhaps there is a reflection from some nearby surface causing interference?

-Charlie

[donradick]

Re: ARTA FR measurement problem

Tell us about the geometry of the measurement setup (distances in room, to speaker, etc.), and what kind of measurement are you doing? It's helpful to know all the details explicitly. Did you happen to see my tutorial on FR measurements using ARTA?

Easy gated impulse frequency measurements using ARTA

I know that things (artifacts) like this happen with nearfield measurements when the distance from source to mic corresponds to some multiple of frequency or vice versa. Perhaps there is a reflection from some nearby surface causing interference?

-Charlie

Charlie - you are my ARTA guru! I printed your guide about 6 months ago, and follow it religiously. Measurement was about 44 inches off the floor, and about 55 inches or so from the ceiling, and 1 meter (39 inches) on axis from the front of the woofer. I believe that I set the ARTa gating properly, just before the main reflection after the impulse. Gating time was 4.2 ms IIRC.

This problem first surfaced a couple of weeks ago when I was re-working and measuring an old Fisher speaker. At first I thought it was specific to that speaker, then remeasured my Targas, which previously Arta measured at +- 2dB from 1K to 18K - they showed the same nulls at 6K and 12K that I'd been seeing on the Fishers.

[takitaj]

Re: ARTA FR measurement problem

If you're using a mic calibration file, try it without it. Maybe it got corrupted?

TJ

[charlielaub]

Re: ARTA FR measurement problem

Don,

Try doing a "near-er" field measurement, e.g. only from 6 inches away or less, or a true nearfield measurement (as close to woofer dustcap apex as possible without letting dustcap contact mic when running sweep). You will be able to increase gating all the way to the end of the impulse and the FR will be immune from room responses, reflections, and so on. If the nulls have not disappeared, it's not an interference problem.

The 6k, 12k Hz pattern does suggest interference to me, however, so that is where my money lies at this point.

-Charlie

[donradick]

Re: ARTA FR measurement problem

Don,

Try doing a "near-er" field measurement, e.g. only from 6 inches away or less, or a true nearfield measurement (as close to woofer dustcap apex as possible without letting dustcap contact mic when running sweep). You will be able to increase gating all the way to the end of the impulse and the FR will be immune from room responses, reflections, and so on. If the nulls have not disappeared, it's not an interference problem.

The 6k, 12k Hz pattern does suggest interference to me, however, so that is where my money lies at this point.

-Charlie

OK, thanks! It'll be a couple of days before I can mesaure again - got a crazy work schedule coming up. I'll update this thread with the new results.

[ernperkins]

Re: ARTA FR measurement problem

Don -

Are you using the same gate times for Holm and ARTA? From the graphs I'd guess the Holm graph might have smoothing (1/12, 1/6, 1/3, etc.) selected instead of "Impulse Response". Just a shot in the dark.

Regards....

[donradick]

Re: ARTA FR measurement problem

some more about the physical measurement setup.
I am measuring in my study, a fairly small room with an 8 foot ceiling.
I set the speaker on the edge of my desk and the mike about 39 inches away.
After a few measurements, I realized that I needed to raise the speaker so the driver under test was basically equidistant from the floor and ceiling.

SO... Question.

Choice #1 is to raise the speaker being measured by placing it on top of a utility (solid) plywood box.

Choice #2 is to to raise the speaker being measured by placing it on top of a small (12 inch ) table.

I figured that Choice 2 would be better, less reflections and better BSC measurements, so I've been using #2 for the last few months. Maybe that's the problem.

In a few days I'll be able to get back to this, and try with both methods, as well as Charlie's nearfield measurement check.

[rpb]

Re: ARTA FR measurement problem

some more about the physical measurement setup.
I am measuring in my study, a fairly small room with an 8 foot ceiling.
I set the speaker on the edge of my desk and the mike about 39 inches away.
After a few measurements, I realized that I needed to raise the speaker so the driver under test was basically equidistant from the floor and ceiling.

SO... Question.

Choice #1 is to raise the speaker being measured by placing it on top of a utility (solid) plywood box.

Choice #2 is to to raise the speaker being measured by placing it on top of a small (12 inch ) table.

I figured that Choice 2 would be better, less reflections and better BSC measurements, so I've been using #2 for the last few months. Maybe that's the problem.

In a few days I'll be able to get back to this, and try with both methods, as well as Charlie's nearfield measurement check.

I use a speaker stand, and set the stand on a small speaker box to raise it to 48" above the floor.

Is there a second speaker that could be getting the sweep signal as well as the one you are measuring? In HOLM, you select left or right. It's possible to select both too. Did that once by accident.

Also, I'm with takitaj on checking the correction file.

[jmb]

Re: ARTA FR measurement problem

Hi Don,
Just a suggestion: if you measure the speaker on a table or desk that has a solid front, you effectively increase the baffle size and increase the necessary distance to obtain a true farfield measurement. Though it is not likely the cause of your problem, it is something to consider to obtain clean measurements.

Jay

[donradick]

Re: ARTA FR measurement problem

Hi Don,
Just a suggestion: if you measure the speaker on a table or desk that has a solid front, you effectively increase the baffle size and increase the necessary distance to obtain a true farfield measurement. Though it is not likely the cause of your problem, it is something to consider to obtain clean measurements.

Jay

Yah, that's why I switched the stand from a solid box to a small table - but I'm now going to go back and test the same speaker and mke distance with both type of stands, just in case there is a measurable difference.

[jonpike]

Re: ARTA FR measurement problem

I'd say it would be an idea to get a mic stand. But as said, probably not causing this problem...

Does ARTA have input polarity settings? Might something gotten reset?

[donradick]

Re: ARTA FR measurement problem

I'd say it would be an idea to get a mic stand. But as said, probably not causing this problem...

Does ARTA have input polarity settings? Might something gotten reset?

I have a mike stand, I've been talking about a speaker stand in this thread.
I'll be able to play with this some more Monday evening.