Difference between curves v1.

2
User guide v1

How to find the frequency response in a microphone.

‘Difference between curves’ allows to find the difference between two measurements made in Smaartv7, allowing to find
the frequency response difference and phase from a measurement in relation to a reference.

For example, to compare a regular optimized voice microphone response to a measurement microphone, considered flat
response.

Step 1, Configuring the reference system

The reference system is used to compare two microphones. This system can be any loudspeaker or full range system
capable of reproducing from 40 Hz and up to 20,000 Hz (+/-3db). This response must be very smooth since very narrow
peaks and/or underscore cancelations may compromise the measurement result.

A good way of doing these measurements is to

place the ‘Reference’ microphone in the center
of the reference system, 30 cm or 60 cm away.

Notes: To achieve good measurement and

avoiding the room acoustic interferes with this
measurement, as well as with microphones, it is
used the 3 to 1 rule. It means the distance
between the microphones and the reference
system is at least 3 times shorter than the
distance between the microphone and any
source of reflexion (walls, glasses, the ground,
Image 1
etc.)
Reference system with “reference” microphone considered flat.

Step 2, Gain structure.

To ensure good gain structure, the measurement software RTA is used as follows:

1. Take a screen shot of the

background noise with the
reference system off.
2. Turn on the reference system
along with the noise generator.
3. Adjust the level of the noise
generator so it is louder than
the background noise for at
least 20 dB.
Image 2
Once the reference system level is Shot of acustic brackground noise and capture of the noise generator level with adjusted level.
adjusted. The ‘reference’ microphone
gain and the electric measurement channel must be adjusted until the VU meter level is away from the background noise
and has enough headroom for saturation and both signals have a similar level. It is important to turn on the generator
when adjusting it.
 For the measurement to give data about the sensitivity difference between the microphones, the gain level must be the
same among the different preamplifiers.

Image 3
Pleamplifiers from 1 to 4 dedicated to microphones, all of them with the same gain.
Signal in the microphone channels 1 (‘reference’ microphone) and 6 (electric measurement) away from the background noise and
with headrooom until saturation.
Signal in channels 1 and 6 with the same level.

Step 3, Transference
Before starting, verify any unwanted correction curve is in used in Smaart.

 Go to ‘Options/Measurement Config…’

 Under any transference function channel, go to the option ‘Configure Device’

 Release any unwanted curve assignment.

Notes: Measurement Microphones are condenser microphones, which are sensitive to humidity and temperature changes.
In order not to compromise the measurement, it is important to do this process at a stable weather.

Trace capture
Obtaining the first transference.

1. 1. Pay special attention in

obtaining a good
frequency response (35 Hz
up to 20,000Hz.)
2. Make sure there is no low-
coherence in the same
frequency region. They
must be at least higher
than 97%. The low-
coherence represents that
for that frequency region
the data obtained by
measuring is not accurate.
This might be due to a bad Image 4
Reference curve.
gain structure, the
presence of reflexions, the lack of synchrony (Delay finder), etc.
3. Capturing the reference curve.
Adjust the Smaart parameters in order to average the transference for 10 seconds with any kind of smoothing or
weighting curve. Wait about 20 seconds to obtain good average and then capture the curve.
4. Capture the measurement curve. In order to fairly compare the microphones, they must be
in the same position than the reference system. This second measurement is done by using
another transference function channel. To replicate the position in a simple way, the Delay
Finder value among the channels is matched.
With the purpose of obtaining the same phase response, the microphone is set closer and/or
moved away from the reference system until both phase responses are equal. Use an average
of 10 seconds to do the transfer. Wait 20 seconds and capture. Image 5
Matched Delay Finder
between channel 1 and 6.
Notes: Adjust the average value to 16 or 8 averages to notice the phase change.

Image 6
Microphone to be measured set in the same position used for the ‘reference’ microphone.
Frequency response and phase of the microphone to be measured in this position.The phase response shows the same delay, so we know they’re in the
same position in relation to the reference system.
Step 4, Using the ‘Difference between Curves’ v1.2
Excel file has 3 different tabs:

‘ASCII Import’ tab: Allows to import the curves previously obtained in Smaart as follows:

 Right-click on the “reference” microphone memory and click “Copy to ASCII.

 Paste the copied values in the Excel sheet.

Notes: Data like “reference” (green area) or “measurement” (blue area) can be pasted in the Excel sheet.
 Import the measurement data as well.

 Paste these information as measurement.

Graph tab: Shows the resulting graphic from the difference in the reference and measurement curves.

The next parameters can be modified in this tab:

 Offset: Allows to compensate the level in one of the curves, either in negative or positive values.
 Average and Smooth response: Allows to smooth the response either of frequency or phase in octave fractions
steps from 1/48 octaves and up to 1 octave.

The Coherence (30 Hz- 20 Hz) section shows how measurement coherence is performing:

 Average Coherence: Shows the coherence average value. This parameter must be over 97%, if not, measurement
can be compromised.
 Lowest Coherence: Shows the lowest coherence value.
 Highest Coherence: Shows the highest coherence value.

Average sensitivity difference: shows the average sensitivity difference between both microphones.

For example, the frequency response can be noticed in the voice microphone in relation to the measurement microphone.
If measured frequency response is compared to the manufacturer given response, it is possible to notice its similarity.

Image 7
Obtained frequency response in “Difference between curves” and manufacturer given response.
Creating correction curves for a microphone
Knowing the difference between 2 microphones, ‘Difference between curves’ can generate correction curves to make
both microphones frequency response to match.

Capturing Reference and Measurement curves

A curve for each microphone must be captured, taking the next considerations into account:

 To have a Reference system capable of reproducing at least from 40 Hz and up to 20,000 Hz (+/- 3dB).
 To minimize the room acoustic effect on the measurement.
 By using the RTA, adjust the gain structure until the reference system level is at least 20 dB higher than the
background noise.
 Make sure there are not correction and weighting curves in the measurement system.
 Check the frequency and coherence response.

Obtaining a curve for the

Reference microphone
Set the microphone in the middle of
the reference system 30 or 60 cm
away to visualize the “reference”
microphone frequency response.
When the graphic is satisfactory,
change the measurement averages to
10 seconds. Wait 20 seconds and
capture the curve for the ‘reference’
microphone.

Obtaining a curve for the

measurement microphone
Image 8
(microphone to be corrected) Frequency response obtained for the “reference” microphone
In order to fairly compare the
microphones, the Reference microphone position must be replicated. To easily replicate it the phase must be observed as
it follows:

 Match the Delay Finder

values for the second
microphone.
 Configure the
measurement to do 8 or 16
averages and move the
microphone until making
the phase response match.

Now that both microphones are in

the same position, the
measurement must be configured
to do averages for 10 seconds. Image 8
Wait 20 seconds and capture the “Reference” microphone capture and capture of the microphone to be corrected.
curve of the measured
microphone.
Importing the curves to Excel
1. In order to import the traces to Excel, use
the “ASCII Import” tab.
2. By right-clicking on the stored data and
selecting “Copy To ASCII”, take the trace
from the reference microphone in Smaart.
3. In the Excel file, go to the “ASCII Import” tab
Imagen 10
and paste the trace in the green section
(Reference).
4. Repeat the process for the Measurement
microphone curve and paste in the blue
section (Reference).
Image 9
Check out the plotter Importing and exporting captured measurements.
Go to the “Graph” tab and make sure:

 The phase and frequency response

are the expected ones.
 The “Average Coherence” value must
be higher than 97%, if not, curves
need to be captured again.
 The setting of the Offset values allows
to compensate microphones
sensitivity differences. If the
preamplifier gain structure was done
correctly, these values must be set in
0, this is how the correction curve will
automatically adjust the microphone
sensitivity differences. Image 11
Graph tab.
Notes: The “Smooth” values don’t affect the
correction curve. These values only affect the plotter.

Mic Correction 1/12 oct and 1/24 oct

This tab allows you to export the correction curve from the difference
between both curves. The correction curve created in this section has 110
correction points with a resolution of 1/24 octaves.

Notes: It is important to consider that correction curve only corrects the

frequency response, the phase response can’t be corrected in this process.

Image 12
Mic correction tabs.
Exporting the correction curve
To export the correction curve, it is as
simple as selecting the Excel “Save as”
option and save the file with the text
format “Text (Tab delimited) (*.txt)”.

Image 13
Save as tab delimited text. Spanish and English menu.
Importing the curve in Smaart
1. Go to “Options/Measurement Config”…

2. Under any transference function channel go to the “Configure Device” option.

3. Select “Mic Correction Curves” and then “Import”.

 4. Select the correction curve file you want.

5. Assign it to the channel you want.

Notes: The curve will be assigned to a sound card channel; this means the microphone must always be connected to the
same preamplifier.
Verification
Calculate again using the microphone with the correction curve. You should obatin a very similar to the first microphone
trace (“reference microphone”).

Image 14
Measurement of microphone 2 using the created correction curve and objective curve (“reference microphone”). You can see
both microphones are calibrated.