| Application
note # 11 |
Sample
Champion
|
Room Acoustics - Acoustical Parameters Measurement
This
application note describes how to use the Sample Champion Room
Acoustics plugin for computing RT20, RT30, RT custom, EDT, D50,
C50, C custom and TS parameters (see the ISO 3382 normative).
Figure
1 - Room Acoustics plugin window
Room Response Measurement
First
of all a Room Impulse Response has to be measured by using the
main program, Sample Champion. This kind of measurement is done
by placing the sound source (fed by the MLS signal) and the
microphone (at listening position) inside the room under test.
The
sound field should be diffuse.
The ISO 3382 normative gives some guidelines for this kind of
measurement (performed by using MLS signals):
- The
sound source should be as close to omni-directional as possible
- The
sound pressure level might be low because the SNR can be increased
by averaging several measurement cycles
- The
microphone should be omni-directional
- Microphone
position should be far from the nearest reflecting surface
at least 1/4 of wavelength, i.e. normally about 1 m
- The
minimum microphone-source distance should be at least:
meters,
where V is the room volume, c is the speed of
sound and T is the expected reverberation time.
More information can be found in the ISO/DIS 3382 normative.
The
room response can be measured by using Sample Champion simply
by starting an average cycle of measurements.
NOTE
1: The length of the MLS signal must be greater than the
expected reverberation time, otherwise the tail of the room
impulse response will wrap at the beginning of the measurement
(see MLS theory page about
this). For example, by using a sample rate of 48 kHz for a
measurement in a large room with reverberation time of about
2 s, a MLS length of 128K = 2719.7 ms must be used (or better
256K, to have a larger margin).
NOTE
2: The measured Impulse Response gives information for the
specific point of the room where the microphone is placed.
RT20, RT30,
RTcustom
The
Room Impulse Response gives information about the reverberation
time by mean of the Schroeder Plot (a detailed explanation
of this analysis technique is described in the Application
Note #12)
Figure
2 - RT30, obtained from the Schroeder Plot.
The white line is the regression line
RT20 is the reverberation time of the room (the time
required for a sound decaying of 60 dB) evaluated over a 20
dB decay range in the Schroeder Plot (from -5 to -25 dB), using
linear regression techniques. It is the time distance between
the -5 dB and the -25 dB points of the Schroeder Plot, multiplied
by 3 (actually the computation is more accurate because the
plugin uses the values on the regression line).
RT30
is the reverberation time of the room evaluated over a 30 dB
decay range (from -5 to -35 dB), using linear regression techniques.
RTcustom
is the reverberation time of the room evaluated over a decay
range set by the user, using linear regression techniques.
For
every parameter, the correlation coefficient r is shown.
A value close to -1 indicates a perfect fit of the linear regression
over the decay curve. When the value is greater than -0.95,
a red alert is shown to indicate a loss of accuracy.
The
plugin allows computating these parameters manually on the unfiltered
Schroeder Plot or on the Schroeder Plot filtered in Octave
or 1/3 Octave bands (standard IEC 1260). A noise
compensation algorithm is implemented and optionally used
by the plugin, when possible. The user can choose whether computing
the noise floor power from the beginning (before the first peak)
or at the end of the Impulse Response (when the impulse has
decayed under the noise level).
The
values of these parameters are computed for all octave bands
in a single step, by pressing the "Show Parameters"
button.
They
can be viewed in tabular form or as a graph.
Figure
3 - RT30, computed for all octave bands
Figure
4 - RT30, graph for all octave bands
The noise compensation algorithm can be applied only when certain
mathematical conditions are satisfied. When this happens (and
the noise compensation option is enabled), a small red rectangle
is shown near the octave band number.
EDT
The
EDT parameter (Early Decay Time) is the
reverberation time, measured over the first 10 dB of the decay.
This gives a more subjective evaluation of the reverberation
time. The user can also change the decay interval of this parameter.
As RT parameters, the EDT is computed for every octave band.
It is expressed in ms.
D50
The
D50 parameter (Definition or Deutlichkeit) is
the early to total sound energy ratio. It is defined
as:
and
expressed in percentage.
C50, C custom
The
C50 parameter (Clarity or Klarheitsmass) is the early
to late arriving sound energy ratio. It is defined as:
and
expressed in dB. It is common practice to calculate this parameter
also over 80 ms (C80). The user can set this time interval (C
custom).
TS
The
TS parameter (Centre Time) is the time of the centre of gravity
of the squared impulse response. An high value is an indicator
of poor clarity. It is defined as:
and
expressed in ms.
Use
of the Room Acoustics Plugin
A
Room Impulse Response must be measured or loaded in the main
program. If you load an I.R. file, select in the Settings|General
window a buffer and MLS lengths equal or greater than the I.R.
length to be loaded, otherwise the end of the impulse will be
lost.
Open
now the Room Acoustics plugin and press F5.
The
impulse response can be shown in different ways (Data, Log Squared,
Schroeder Integration, Step Response and Cumulative Energy).
- Select "View Data as..." Schroeder Plot
- Press the "View ALL data" button (Unzoom)
- Uncheck Autoscale
- Enable the Filter and select the desired Octave or 1/3 Octave
band
Now
press the "RT20" button. The two cursors will be positioned
at the -5 dB and -25 dB points of the plot and the RT20 will
be computed and shown in the right lower corner of the graphic
screen. The position of the cursors can be changed manually
if desired.
NOTE:
The noise compensation option leaves to the user the
possibility to choose whether the noise must be evaluated at
the beginning or at the end of the I.R. file. The choice depends
basically on the considered Impulse Response. Probably when
the initial silence gap before the first peak is short, it's
better to use the "End" option. This often occurs
when measuring Room Impulse Responses in small rooms and the
source-microphone distance is low. When measuring an I.R. in
large rooms or halls, it could be useful to use the "Beginning"
option. Anyway the number of points used for evaluating the
silence portion of the file is written by the plugin below the
Noise Compensation Enable/Disable checkbox. If the number of
points is too low, try changing the option settings. Note that
when the filter is enabled the initial gap could become small
or null; in these cases use the "End" option.
By
pressing the "Show Parameters" button, all parameters
will be computed and shown in a tabular form.
The values of the parameters can also be plotted as bar graphs,
by selecting the "Ac. Par. Graph" in "View"
mode.
All
computed parameters can be exported to Clipboard and the graphic
plots can be saved to disk as image files.
 Download
PDF file (269K) Shift-click
to download
Back to Application Notes
|
