GOLDBERG AND MÄKIVIRTA
AUTOMATED IN-SITU EQUALISATION
AES 23RD CONFERENCE, May 23-25, 2003
7
rithm discussed in Section 3 and statistical data for
each measurement before and after equalisation was
recorded. The statistical data is analysed to study how
the objective quality of the system magnitude re-
sponse has been improved by using the proposed algo-
rithm for setting the room response controls.
Table 10. Acoustic measurement system parameters.
Parameter
Equipment / Setting
Measurement System
WinMLS2000 [39]
Microphone
Neutrik 3382 [40]
Sample rate,
f
s
48
kHz
MLS sequence order
14 (16)
Averages 1
Impulse response length
0.341 s (1.36 s)
Time window
Half-cosine
FFT size
16384 (65536)
Frequency resolution
2.93 Hz (0.733 Hz)
4.1. Statistical Data Analysis
A further statistical analysis was conducted on all of
the loudspeakers in the study. The bandwidths of the
frequency bands used are shown in Table 11. The
bandwidths ‘
LF
’, ‘
MF
’ and ‘
HF
’ are later referred to
collectively as the ‘
subbands’
and correspond roughly
to the bandwidths for each driver in the three-way sys-
tems.
Table 11. Frequency band definitions the statistical
data analysis;
f
LF
is the frequency of the lower –3 dB
limit of the frequency range.
Frequency Range Limit
Bandwidth Name
Low
High
Broadband
f
LF
15
kHz
LF
f
LF
400
Hz
MF
400 Hz
3.5 kHz
HF
3.5 kHz
15 kHz
For each loudspeaker, the broadband (Table 11) mag-
nitude response data median value is standardised to
0 dB.
The statistical descriptors recorded before and after
equalisation for each loudspeaker and in each fre-
quency band defined in Table 11 are the minimum,
maximum and range of the magnitude dB values. Also
for the magnitude pressure values in each bandwidth
(Table 11), the median, 5% & 95% percentiles and
quartiles are recorded. In addition, the root-mean-
square (RMS) deviation of the pressure from the me-
dian in each bandwidth is calculated: the value is ex-
pressed in dB.
These statistical descriptors are compared for each
subband to study the in-band flatness improvement
due to equalisation. The median values for each sub-
band are compared to study the broadband tonal bal-
ance improvement. This is indicated by a reduction of
the median value differences.
4.2. Example of Statistical Data Analysis
Figure 7 in Appendix C shows a case example where
room response control settings are calculated accord-
ing to the optimisation algorithm. The equalisation
target is a flat magnitude response (straight line at
0 dB level). The in-situ frequency response of the
loudspeaker was recorded before equalisation, i.e.
when all the room response controls were set to their
default position, which has no effect to the response.
The appropriate room response control settings were
calculated using the optimisation algorithm, applied to
the loudspeaker and the corrected in-situ frequency
response plotted. The loudspeaker’s passband (trian-
gles) and the frequency band of equalisation (crosses)
are indicated on the graphical output. The proposed
room response control settings are shown and the ef-
fect of these settings is visualised in the response plot.
The treble tilt, midrange level and bass tilt controls
have been set. The equalisation corrects the low fre-
quency alignment and improves the linearity across
the whole passband.
Figure 8 in Appendix C shows a statistical analysis of
the same loudspeaker presented in graphical form.
The upper three plots were calculated before equalisa-
tion and the lower three plots after equalisation. The
plots display the values of percentiles in the magni-
tude value distribution (box plot), the histogram of
values and the fit of the magnitude values to normal
distribution before and after equalisation. These plots
clearly show that the distribution in magnitude data
has been reduced. This is illustrated by the reduced
range in the box plot and the value histogram, as well
as a better fit to a normal distribution in the normal
probability plot.
4.3. Results
A total of 63 loudspeakers were measured before and
after equalisation. Of these, 12 were small two-way
systems, 22 were two-way systems, 30 were three-
way systems and three were large systems.
Depending on the product type, not all of the room
response controls are available (Tables 1–4). Table 12
shows the number times the controls were used when
available on the loudspeaker. The midrange level con-
trol is used most frequently and the bass roll-off the
least.
