Visual Tools For Counseling And Verification; Percentile Analysis - Interacoustics Affinity 2.0 Additional Information

2.15 Visual tools for counseling and verification

2.15.1 Percentile analysis

The Percentile Analysis can be added in the REM440 setup (see section 0.1)
The Percentile Analysis is implemented according to the IEC 60118-15 standard.
By def ault, the 30th, 65
FFT analysis of 100 measurements.
The 30th percentile (indicated in blue shading) is the sound pressure level below which 30% of the
measured levels are f ound f or each f requency. The remaining 70% of the measured sound pressure levels
are higher. The 30th percentile thereby represents the lower valleys or the sof t components of the speech
signal. This helps you ensure that even the sof t valleys are above thresholds.
The 99th percentile (indicated in green shading) is the sound pressure level below which 99% of the
measured levels are f ound f or each f requency. The 99th percentile may be interpreted as a peak indicator
which helps to ensure that no peaks come too close to the MPO and/or UCL's. In other words, the 99
percentile represents the loud components of the speech signal.
Together the 30th and 99th percentile provides an "area of audibility". The 65th percentile serves as a
borderline between the 30th and 99th percentiles and is visually represented at the point where the blue and
green areas meet. The 65
the FFT analysis of the 100 measurements.
Picture note: On this illustration the percentile analysis reveals important inf ormation about the hearing aid
f itting. At 3-4 kHz the peaks of the signal (the 99
that even though we cannot f it above thresholds the client is still getting some speech inf ormation in this
f requency area since the 99
even though the curve is above threshold, and we match target, the lower valleys (30
above thresholds and the client is thereby missing inf ormation.
D-0004576-Q – 2022/05
Affinity/Equinox2.0 – Additional Information
th , and 99th percentile are selected. These percentiles are calculated based on an
percentile represents the median whereas the measurement is the average of
th
th percentile is above thresholds. In the low f requency area, we can see that
th percentile) are too close to the UCL. However, we also see
th
percentile) are not
th
Page 257