Author Topic: Room Analyzer - best practice?  (Read 6009 times)

0 Members and 1 Guest are viewing this topic.

gwarder

  • Newbie
  • *
  • Posts: 4
    • View Profile
Room Analyzer - best practice?
« on: February 06, 2014, 10:34:27 AM »
I am having a great deal of difficulty getting any real results. So I'd like to start by ruling out "user error" as I am fairly new to this.

My goal is to measure a church where I have recorded a choir/organ ensemble. The end result is for a Masters paper I'm writing.

To measure the room, I have placed an AKG C414 XLIIS in omni mode about 10' up and slightly behind the conductor's position (same place I set the mic array). The mic went into a Millennia mic pre, then to an Apollo 16 interface, finally into a Mac laptop. The Apollo interface also fed a monitor speaker placed in the centre of the choir. This was for the excitation signal.

On launching Electroacoustics Toolbox, I opened Device I/O and selected (and started) the Apollo 16 interface.

Next created a Dual FFT Analyzer. In the bottom drawer I set the source to the appropriate Apollo 16 input. Under the "Excitation" tab, I Selected the appropriate output channel...sine sweep...Logarithmic. I left all other settings at default.

Next I selected Control->Start All Tools. I confirmed that I could hear the repeating excitation signal in the room and that the mic was picking it up. While the excitation signal was in progress, I selected Control->Capture all Data. Then I stopped all tools.

Next I created a Room Analyzer and selected the captured measurement. There are LOTS of n/a boxes in the analysis. Nothing captured in C80, nothing for 4 areas of C50, NO T60 measurements, several bands empty in T20 and T30. Nothing for D80 and 4 blanks for D50. The T measurements that I did get are quite different from a sweep done with "FuzzMeasure" and all seem extremely short.

This is a sanctuary that has a fairly low noise floor (lights, fans and organ turned off during measurement), and a "starter pistol" excitation produces a decay that lasts roughly 4-5 seconds to the naked ear.

I'm wondering what is wrong in my setup. Mic position? Polar pattern? Something in the excitation?

Thanks!

FaberAST

  • Administrator
  • Sr. Member
  • *****
  • Posts: 311
    • View Profile
    • www.FaberAcoustical.com
Re: Room Analyzer - best practice?
« Reply #1 on: February 06, 2014, 10:57:18 AM »
gwarder,

Quote
Next created a Dual FFT Analyzer. In the bottom drawer I set the source to the appropriate Apollo 16 input. Under the "Excitation" tab, I Selected the appropriate output channel...sine sweep...Logarithmic. I left all other settings at default.
Did you select the same (excitation) output channel for the Reference input (next to the Source input selection)?

Quote
Next I selected Control->Start All Tools. I confirmed that I could hear the repeating excitation signal in the room and that the mic was picking it up. While the excitation signal was in progress, I selected Control->Capture all Data. Then I stopped all tools.
If you have any other tools in your project, I would recommend just starting and stopping the DFFT Analyzer directly, rather than all tools simultaneously. Also, for room measurements, you might want to set the averaging type to Linear and let the analysis complete (and stop on its own) before you capture the data. (The default number of averages, 10, may be overkill, but it won't hurt your measurement.)

Also, before capturing the data in the DFFT Analyzer, I would recommend that you inspect your measurement data to be sure it looks right. If you are familiar with ETC or squared impulse response curves, I would recommend that you select the Squared IR function in the Display tab and make sure you see the kind of acoustic decay you would expect in your measurement. If you don't have valid measurement data in the DFFT Analyzer, you can't expect to get meaningful results when you load it into the Room Analyzer.

The Coherence function can also be used to verify the quality of your measurement, as a function of frequency. A coherence of 1 is ideal. The smaller the coherence, the less meaningful your data is at a particular frequency.

Ben

gwarder

  • Newbie
  • *
  • Posts: 4
    • View Profile
Re: Room Analyzer - best practice?
« Reply #2 on: February 06, 2014, 12:55:56 PM »
gwarder,

Quote
Next created a Dual FFT Analyzer. In the bottom drawer I set the source to the appropriate Apollo 16 input. Under the "Excitation" tab, I Selected the appropriate output channel...sine sweep...Logarithmic. I left all other settings at default.
Did you select the same (excitation) output channel for the Reference input (next to the Source input selection)?

Yes I have. Both my excitation output and my Live Input reference are set to Output 10 on the Apollo 16.

Quote
Next I selected Control->Start All Tools. I confirmed that I could hear the repeating excitation signal in the room and that the mic was picking it up. While the excitation signal was in progress, I selected Control->Capture all Data. Then I stopped all tools.
If you have any other tools in your project, I would recommend just starting and stopping the DFFT Analyzer directly, rather than all tools simultaneously. Also, for room measurements, you might want to set the averaging type to Linear and let the analysis complete (and stop on its own) before you capture the data. (The default number of averages, 10, may be overkill, but it won't hurt your measurement.)

I'm not sure I know what you mean by "waiting for the analysis to complete (and stop on its own). Once I start the Dual FFT, it just runs. The excitation signal continues until I stop the tool. I did let it run for two minutes before stopping it. Should I wait longer? Seems like 10 repeats should happen much more quickly.

At which point do I click on "Capture Data?" While the excitation is running? After I stop the tool? I've tried both and I'm just getting an empty graph.

Quote
Also, before capturing the data in the DFFT Analyzer, I would recommend that you inspect your measurement data to be sure it looks right. If you are familiar with ETC or squared impulse response curves, I would recommend that you select the Squared IR function in the Display tab and make sure you see the kind of acoustic decay you would expect in your measurement. If you don't have valid measurement data in the DFFT Analyzer, you can't expect to get meaningful results when you load it into the Room Analyzer.

The Coherence function can also be used to verify the quality of your measurement, as a function of frequency. A coherence of 1 is ideal. The smaller the coherence, the less meaningful your data is at a particular frequency.

Ben

gwarder

  • Newbie
  • *
  • Posts: 4
    • View Profile
Re: Room Analyzer - best practice?
« Reply #3 on: February 06, 2014, 01:23:17 PM »
After a restart, I do see signal and can capture some data (the excitation still doesn't stop on its own).

However, the numbers aren't making sense so I'm wondering if there's something wrong with the sample. I am getting NOTHING for C80 (just N/A) and the only C50 measurement is @ 63 Hz.

I need C80 at 500, 1k and 2k, RT60 at 125, 250, 500 and 1000, and EDT at 500, 1000, 2000 and 4000 for the paper I'm writing.

gwarder

  • Newbie
  • *
  • Posts: 4
    • View Profile
Re: Room Analyzer - best practice?
« Reply #4 on: February 06, 2014, 01:35:48 PM »
One last point for now....I took another sample three minutes later and changed nothing on the setup. The results seem fairly different....and lots more "N/A"s on the chart.

This is a sanctuary where a starter's pistol excitation takes 3-4 seconds to decay....and in the countryside to the noise floor is pretty quiet.

FaberAST

  • Administrator
  • Sr. Member
  • *****
  • Posts: 311
    • View Profile
    • www.FaberAcoustical.com
Re: Room Analyzer - best practice?
« Reply #5 on: February 07, 2014, 05:28:43 PM »
Quote
I'm not sure I know what you mean by "waiting for the analysis to complete (and stop on its own). Once I start the Dual FFT, it just runs. The excitation signal continues until I stop the tool. I did let it run for two minutes before stopping it. Should I wait longer? Seems like 10 repeats should happen much more quickly.

At which point do I click on "Capture Data?" While the excitation is running? After I stop the tool? I've tried both and I'm just getting an empty graph.

If you set the averaging type to Linear, the measurement will stop after the specified number of averages has been reached. I would recommend capturing the data after that.

Your screenshots clearly indicate that you are not yet getting valid measurements within the DFFT Analyzer, so that issue needs to be resolved before worrying about bad results from the Room Analyzer.

Ben