This measurement column addresses frequently asked questions received by our customer support center and provides answers to those questions.
When measuring sound, there is an item called "Frequency Weighting Characteristics (A/C/Z)" in the settings.
Analysis devices such as FFT analyzers and data stations have two settings related to "frequency weighting characteristics (A/C/Z)". One is "analog filter" in the input settings, and the other is "frequency weighting" in the analysis settings.
This time, we will discuss real-time octave analysis (RTA analysis) using analysis equipment and processing in FFT analysis.
We will explain the process and how these two settings affect the analysis results.
Regarding the two settings related to "Frequency Weighting Characteristics (A/C/Z)"
This guide explains how to configure the "analog filter" and "frequency weighting" settings for the DS-3000 acoustic vibration analysis software and the CF-9200/9400 FFT analyzer.
- In the case of DS-3000 acoustic vibration analysis software
The "analog filter" setting is located in the input condition settings dialog, which appears when you click the Input/Output Settings menu ⇋ Input Settings command. You can select A/C/Z for each input channel.
The "Frequency Weighting" setting can be found in the Analysis Settings menu ⇋ Frequency Weighting, where you can select from A/C/Z, etc., for each graph. However, for Schedule/Schedule 3D graphs, you can only set it for all graphs at once.
Please note that the "analog filter" setting cannot be changed in offline analysis mode.
The recording mode does not have a "frequency weighting" setting. - For CF-9200/9400 FFT analyzers
The "Analog Filter" setting is located under the soft keys Home ⇋ Input ⇋ Input Cond, and you can select A/C/Z for each input channel.
The "Frequency Weighting" setting can be accessed via the soft key Home ⇋ Analysis ⇋ Freq Calc ⇋ Weight, where you can select A/C/Z etc. for each graph. However, for Schedule/Schedule 3D graphs, the setting can only be applied to all graphs at once.
Processing flow when performing RTA analysis
When performing RTA analysis, the input signal is directly subjected to RMS detection and logarithmic calculation.
The ALLPASS value and the output values for each band obtained by passing the signal through a bandpass filter for each octave band, followed by RMS detection and logarithmic calculation are obtained. When performing a 1/1 octave band analysis, if the frequency range is HIGH, a total of 10 band outputs are obtained: 31.5Hz, 63Hz, 125Hz, 250Hz, 500Hz, 1kHz, 2kHz, 4kHz, 8kHz, and 16kHz. With a 1/3 octave band analysis, a total of 30 band outputs are obtained: 25Hz, 31.5Hz, 40Hz, ..., 12.5kHz, 16kHz, and 20kHz.
Figure 1 shows the processing flow when connecting a microphone to the analysis device and performing RTA analysis.
-
Figure 1: Flowchart for RTA analysis
The signal from the microphone passes through an analog filter, then through an RMS detection characteristic circuit and a logarithmic calculation circuit, and is displayed as the ALLPASS value on the graph. Additionally, the results after passing through each bandpass filter, RMS detection characteristic circuit, and logarithmic calculation circuit are displayed as the output values for each band. The OVERALL value is the sum of the power values of each band.
If you set the "analog filter" to A-weighting or C-weighting, you must set the "frequency weighting" to Z (FLAT). Since the analog filter applies A-weighting or C-weighting, the ALLPASS value, each band output value, and OVERALL value are all A-weighting/C-weighting applied values.
If you set the "analog filter" to Z (FLAT) characteristics and perform analysis, and set the "frequency weighting" to A characteristics, the output values for each band will be the values with A characteristics applied. Similarly, the OVERALL value is the sum of the output values for each band with A characteristics applied, so it will also be the value with A characteristics applied. The ALLPASS value is the value obtained by performing RMS detection and logarithmic calculation without limiting the frequency bandwidth, and therefore A characteristics correction cannot be applied, so the Z (FLAT) characteristic value is displayed as is.
Figure 2 shows three examples of analysis results. Figure 2-1 shows the result when both the "analog filter" and "frequency weighting" are set to the Z(FLAT) characteristic. Figure 2-2 shows the result when the "analog filter" is set to the A characteristic, and the ALLPASS value, each band output, and OVERALL value all have the A characteristic applied. Figure 2-3 shows the result when the "frequency weighting" is set to the A characteristic, and since the ALLPASS value does not have the A characteristic applied, the ALLPASS value and the OVERALL value do not match.
-
Figure 2-1 Analysis results with analog filter: Z(FLAT) and frequency weighting: Z(FLAT)
-
Figure 2-2 Analysis results with analog filter: A and frequency weighting: Z (FLAT)
-
Figure 2-3 Analysis results with analog filter: Z(FLAT), frequency weighting: A
Processing flow when performing FFT analysis
Figure 3 shows how to connect a microphone to an analysis device, perform an FFT analysis, and obtain the power spectrum.
The process flow is shown below.
When you specify A-weighting for "Analog Filter," an A-weighting filter is applied to the time-domain waveform, and the power spectrum will also have A-weighting. In this case, "Frequency Weighting" should always be set to Z (FLAT).
If you specify Z(FLAT) for the "analog filter," you specify the desired characteristics using "frequency weighting." If the software allows you to specify different "frequency weightings" for each graph being displayed, you can simultaneously display the power spectrum of the Z(FLAT) characteristic and the A-characteristic.
The power spectra obtained when specifying A-weighting for "analog filter" and when specifying A-weighting for "frequency weighting" are almost identical, so you can use either unless there are special circumstances.
-
Figure 3. Processing flow when performing analysis.
Analysis using a sound level meter
When performing sound analysis, the AC output of a sound level meter is sometimes connected to the analysis device instead of using a microphone.
If the AC output of the sound level meter outputs a Z (FLAT) characteristic signal, the above-described analysis can be performed by setting either the "analog filter" or "frequency weighting" of the analysis device to A characteristic, etc.
When the AC output of the sound level meter outputs an A-weighted signal, the "analog filter" and "frequency weighting" of the analysis device should always be set to Z (FLAT) characteristics. Since the input signal has A-weighting applied, the resulting analysis results will also be A-weighted.
summary
In this article, we introduced two settings related to the "frequency weighting characteristics (A/C/Z)" in the analysis device, and the types of analysis results that can be obtained when these settings are applied.
In most cases when performing analysis with A-weighting, you only need to set either the "analog filter" or the "frequency weighting" to A-weighting.
However, please note that if you set "frequency weighting" in real-time octave analysis (RTA analysis), you will not be able to obtain an A-weighted ALLPASS value.
(Excerpt from the email newsletter issued on April 25, 2017)
