Home Theater Forum and Systems banner

1 - 12 of 12 Posts

·
Registered
Joined
·
66 Posts
Discussion Starter · #1 · (Edited)
Here is what I am doing:

1. I start by making a sound card calibration file, but instead of using a sound card loopback connection I am feeding the frequency sweep into a speaker with a mic positioned two feet away. This takes into account the characteristics of the speaker, the mic, the room and the sound card.

2. Using the calibration file created above, I am making a measurement of the mic in the same exact location and setup. The filters/EQ settings in the calibration file, when applied to the measurement, "flatten out" the measurement readings.

3. I then apply one-third octave smoothing to the resulting measurement.

In the accompanying curve you can see that this works well above 40 Hz. Below 40 Hz things get a bit bumpy and there is no consistency from one measurement to the next regarding the frequency or amplitude of these bumps. Any ideas why this might be? I have tried swapping mics, preamps, speakers and audio interfaces with no luck, and have played around with different equalizer settings, etc.



Unsmoothed:

 

·
Registered
Joined
·
678 Posts
You are close already, but you can use the post padding thread here to get to five posts, and sometime after that (~1 hr) you should be able to post images.
 

·
Premium Member
Joined
·
6,711 Posts
Once your post count is up and the hourly update has happened you can just go back end edit then save your posts (no need to make any changes) and the images will show up.

The bumpiness below 40Hz could be several things, but if the speaker has little output down there the measurement will be more noise than signal.
 

·
Super Moderator
Joined
·
9,279 Posts

I believe you can upload pictures to your post using the “Go Advanced” option and then “Upload and Manage Files and Attachments.” But here are your pictures:

In the accompanying curve you can see that this works well above 40 Hz. Below 40 Hz things get a bit bumpy and there is no consistency from one measurement to the next regarding the frequency or amplitude of these bumps. Any ideas why this might be? I have tried swapping mics, preamps, speakers and audio interfaces with no luck, and have played around with different equalizer settings, etc.



Unsmoothed:


1. I start by making a sound card calibration file, but instead of using a sound card loopback connection I am feeding the frequency sweep into a speaker with a mic positioned two feet away. This takes into account the characteristics of the speaker, the mic, the room and the sound card.

2. Using the calibration file created above, I am making a measurement of the mic in the same exact location and setup. The filters/EQ settings in the calibration file, when applied to the measurement, "flatten out" the measurement readings.

3. I then apply one-third octave smoothing to the resulting measurement.
I’m at a loss as to what you’re trying to accomplish here. Can you explain?


Regards,
Wayne
 

·
Registered
Joined
·
66 Posts
Discussion Starter · #6 ·
The bumpiness below 40Hz could be several things, but if the speaker has little output down there the measurement will be more noise than signal.
I am using speakers with 6" woofers, which is probably not very big for this purpose.

I’m at a loss as to what you’re trying to accomplish here. Can you explain?
If I substitute a different mic for the calibration mic, I will have the frequency response of that mic relative to the calibration mic. A calibrated Behringer mic is on order from Cross Spectrum.
 

·
Premium Member
Joined
·
15,054 Posts
What speakers are you using? 6" drivers wont go much below 60Hz.
 

·
Premium Member
Joined
·
15,054 Posts
The Reveal 6s are a nearfield monitor, (good speakers by the way) they are not going to fill an average room well. The Tannoys go down to 63Hz so I would not expect anything meaningful at 40Hz or below to be produced.
Again the A6s are also nearfield monitors and even though tre website says they go down to 45Hz I think thats being fairly optimistic.
 

·
Registered
Joined
·
678 Posts
... 1. I start by making a sound card calibration file, but instead of using a sound card loopback connection I am feeding the frequency sweep into a speaker with a mic positioned two feet away. This takes into account the characteristics of the speaker, the mic, the room and the sound card.

2. Using the calibration file created above, I am making a measurement of the mic in the same exact location and setup. The filters/EQ settings in the calibration file, when applied to the measurement, "flatten out" the measurement readings.

...

In the accompanying curve you can see that this works well above 40 Hz. Below 40 Hz things get a bit bumpy and there is no consistency from one measurement to the next regarding the frequency or amplitude of these bumps. Any ideas why this might be? ...
Thanks to Wayne's prodding and his posting of your images, I have re-read the description of your experiment. It is an interesting hypothesis, that like a loopback test of a soundcard calibration, building a calibration file from the speaker response in room should produce a flat response from the same speaker. I think the problem you are seeing is the time effect, especially from standing waves at the low end. The peak response from the sub may not appear at time zero, even if it is time-aligned with the mains. I'm not sure exactly how the soundcard calibration file is built, but regardless of the details of the algorithm, it is a flat picture of the response curve, specifying level and phase at each frequency. The sound from the sub is a time varying phenomenon.

Like Wayne, I'm at a loss as to what you are trying to accomplish, if not to test a curious hypothesis. More useful would be to build the soundcard calibration in the normal fashion. Then measurements of the receiver/speaker in the room will give you overall frequency response, and how the response decay varies in time.

... If I substitute a different mic for the calibration mic, I will have the frequency response of that mic relative to the calibration mic. A calibrated Behringer mic is on order from Cross Spectrum.
If you want to compare the performance of the two mics, it would be better to take measurements with a normal soundcard cal file, export the results, and subtract them in, say, Excel.

Bill
 

·
Registered
Joined
·
66 Posts
Discussion Starter · #11 ·
The peak response from the sub may not appear at time zero, even if it is time-aligned with the mains.
There is no subwoofer. It is all main speaker.

Then measurements of the receiver/speaker in the room will give you overall frequency response, and how the response decay varies in time.
That's not what I'm interested in. I want to measure mics, not receivers or speakers.

If you want to compare the performance of the two mics, it would be better to take measurements with a normal soundcard cal file, export the results, and subtract them in, say, Excel.
I want to analyze the performance of a single mic at a time. In order to do that it is necessary to calibrate the system, i.e. the reference mic, the speaker, the room, the sound card and the power amp.
 

·
Registered
Joined
·
66 Posts
Discussion Starter · #12 ·
OK, I got out and literally dusted off a speaker with an 8" woofer which used to have a problem (but the problem has now disappeared). Yes indeed, the bottom is now much smoother than with either of the speakers I tried with 6" woofers. JohnM was right, thanks:

 
1 - 12 of 12 Posts
Top