SSM2000 Analog, SSM2000 Datasheet - Page 7

• Manufacturer Part Number: SSM2000
• Description: HUSH Stereo Noise Reduction System with Adaptive Threshold
• Manufacturer: Analog

REV. 0
APPLICATIONS INFORMATION
Introduction
The SSM2000 is a dual channel audio noise reduction IC which
reduces noise through a combination of variable filtering and
downward expansion in conjunction with a unique adaptive
noise threshold detector. These two techniques yield an overall
noise reduction of up to 25 dB on AM and FM radio, open reel
and cassette tape, CD, Dolby B encoded programming, broad-
cast studio-transmitter links, telephone lines, and other audio
sources without the need for any additional manual adjustment.
The HUSH Noise Reduction System as implemented in the
SSM2000 has been shown to substantially reduce noise in PC
multimedia, intercom systems, teleconferencing systems, mobile
communications, automotive audio, home stereos and televi-
sions, and other consumer and professional audio applications
while preserving full signal fidelity and transparency.
A brief explanation of the nature of audio signals is useful before
discussing the details of the SSM2000’s operation. First, audio
signals have both amplitude and frequency content. Music and
voice are created by changing both the amplitude and frequency
of sound waves as shown in Figure 17. Second, the highest
audio signal amplitudes typically occur at low frequencies
(100 Hz–1 kHz) and taper off exponentially as frequency in-
creases. Third, when an audio signal is recorded, audio noise is
also generated. This noise is the white noise “hiss” or waterfall
sound that is easily heard on taped material. Audio noise for a
given source and bandwidth does not, in general, change with
frequency or amplitude. However, noise levels do change be-
tween different sources, because of differences in recording
equipment, media, and the surrounding environment. Fourth,
psychoacoustic effects mask noise that occurs at or near the fre-
quency of the audio signal. Finally, audio noise is usually con-
sidered to be most objectionable in the 3 kHz–8 kHz bandwidth
region.
Figure 17. A Frequency-Amplitude Plot of the Audio
Envelope
An analog noise reduction system must first distinguish between
the desired source material and the undesirable noise. It must
then attenuate the noise while leaving the source material unaf-
fected. One approach to noise reduction is to assume that signal
below a predetermined amplitude is noise, and to then attenuate
the noise by using a Voltage Controlled Amplifier (VCA). A
variation of this noise reduction method is found in Dolby B
cassette tape systems. This method achieves about 10 dB of
improvement in signal-to-noise ratio (SNR). This system incor-
porates a high-frequency compressor on the recording side, and
a high frequency expander during playback. Another noise
reduction technique senses and reduces noise by measuring the
frequency content of the audio signal and then filters noise that
NORMAL AUDIO SYSTEM FREQUENCY
AND AMPLITUDE ENVELOPE
NOISE
AND AMPLITUDE
FREQUENCY
FREQUENCY
ENVELOPE
MUSIC'S
–7–
occurs above the highest signal frequency. This noise reduction
method utilizes a Voltage Controlled Filter (VCF) and is the
basic method of operation in the DNR
about 10 dB of noise reduction. The HUSH system shown in
Figure 34 combines elements of both of these techniques to
achieve 25 dB of noise reduction, and also has significant
improvements.
Figure 18. SSM2000 Block Diagram (One Audio Channel
Shown)
Recalling that the noise floor changes with different audio
sources due to recording equipment, media, and the environ-
ment, we realize that fixed threshold approaches cannot yield
optimal results. The HUSH Noise Reduction System incorpo-
rates an automatic noise threshold detector that senses these
changes and adapts the VCA and VCF to become more or less
aggressive depending on the amplitude of the noise floor. To
determine the amplitude of the noise floor the SSM2000 as-
sumes that the averaged amplitude during short periods of no
audio is equal to the noise floor. This assumption works well in
audio applications, but does have implications for evaluation
and testing that are discussed in the “Evaluating the SSM2000”
section.
Variable Low-Pass Filtering
The audio signal is first passed through a single-pole low-pass
Voltage Controlled Filter (VCF). Both the Left and Right
VCFs are controlled by a detector which places their cutoff fre-
quencies just beyond the highest audio signal frequency. Since
the highest audio signal frequency constantly changes, the
VCF’s cutoff frequency must also change in concert with the
audio signal to avoid attenuating desired signal.
For example, with signal levels below the filter threshold (pre-
sumed to be mostly noise), the VCF shuts down to about 1 kHz,
providing noise reduction in the critical 3 kHz– 8 kHz band, as
Figure 19 illustrates. The VCF progressively “opens up” as
higher frequency amplitudes are detected at the inputs. The
VCF’s cutoff reaches 20 kHz when the high frequency signal
amplitude is 30 dB above the threshold. At this point the VCF
is acoustically transparent. The VCF’s cutoff frequency range is
between 1 kHz to 35 kHz. The minimum range of the VCF is
limited to 1 kHz for two reasons: one, to avoid high frequency
loss at the leading edge of transients, because the lower the
minimum cutoff the longer it takes the VCF to slew “open,” and
two, noise is most objectionable at mid and high range frequen-
cies. Once again, if the VCF characteristics are to be measured,
then please refer to the section on “Evaluating the SSM2000.”
DNR is a registered trademark of National Semiconductor Corporation.
PRE-DETECTOR
CONDITIONING
SIGNAL
L/R IN
1,2
10
8
9
x3
DETECTOR
VCF
3,22
VCF FILTER
VCF
CAP
4,21
THRESHOLD
DETECTOR
VCA EXTERNAL
NOISE
CONTROL
PORT
7
®
system, which provides
DETECTOR
VCA
MUTE
SSM2000
VCA
17
SSM2000
AMP
OUT
16
24,23
HUSH
DEFEAT
L/R OUT