NOISE CANCELLATION IN HEADPHONES

1
NOISE CANCELLATION IN HEADPHONES
M. Tech. Credit Seminar Presentation

• Vikash Sethia
• (Roll no. 02307023)
• Supervisor Prof. P. C. Pandey
• 18th November 2002

2
Presentation overview

• Introduction
• Passive attenuation
• Active reduction
• Adaptive algorithms used in ANC headphones
• Practical applications
• Conclusion

3
Introduction

• Acoustic noise
• Solutions
• Passive reduction techniques
• Sound absorption and isolation - earmuffs,
  earplugs, audio-limiters
• Good attenuation - about 15 dB and more at middle
  and high frequencies
• Acoustic wavelength dependence of its size and
  mass
• Hence bulky and heavy at low frequencies
• Cont

4

• Active attenuation
• Known as active noise control (ANC) - sound field
  modification, particularly sound field
  cancellation, by electro-acoustical means
• Principle of destructive interference cancel
  noise with anti-noise signal of the same
  magnitude but of opposite phase at some specific
  points
• Adaptation to noise

5
Passive attenuation
Headphone Error
microphone -
Audio signal

• Fig. 1 Passive headphone attenuation with noise
  filtering
• Filters usually bandpass filters and
  audio-limiters
• Poor attenuation inside the passband i.e.
  frequency band of interest
• Cont..

6

• Fig. 2 Typical passive attenuation
• Good attenuation at middle and higher
  frequencies

7
Active ReductionAnalog feedback control
Headphone Error microphone
_ Audio
signal

• Fig. 3 Active headphone attenuation with feedback
  control
• Error signal at the error microphone used as an
  error correction to the noise
• Control elements lead, lag and notch filters
  Cont

8

• Merits Good performance against broadband noise
  at low frequencies
• Demerits
• Possibility of excessive phase-shift leading to
  instability
• Poor performance against narrowband or tonal noise

Fig. 4 Typical active attenuation
Cont
9
Active digital control
Headphone Error
microphone Noise reference
- microphone Audio
signal
Adaptive filter
Fig. 5 Active headphones with feedforward
control Cont..
10

• Requires two microphones
• Noise reference microphone picks the
  disturbance or noise
• Error microphone picks the error signal and the
  noise that is not sensed by the reference
  microphone
• Adaptive filters - use adaptive algorithms on DSP
  devices
• Good performance against narrowband or tonal noise

11
Combined analog and digital control
Audio d(n)
signal
y(n) x(n)
- - To ear
e(n)
Fig. 6 Block diagram of combined analog and
digital control
12
Adaptive algorithms used in ANC headphones

• Least mean square (LMS) algorithm
• For feedforward structure
• Loudspeaker output
• Error signal
• Weight update
• where

13

• Filtered-x LMS algorithm
• For feedback structure
• Error signal
• Estimated reference noise
• Loudspeaker output
• Weight update
• Faster rate of convergence than LMS algorithm
• Complicated because of the accuracy of the
  loudspeaker model

14
Fig. 7 Spectrum of cockpit noise in a jet
aircraft (upper curve), the noise at the ear of
the pilot when wearing a conventional headset
(middle curve) , and at the pilots ear when
using a headset with active noise control.
(Elliott, 1993)
15
Practical applications

• Audiometry to diagnose hearing defects and
  speech discrimination
• In industry particularly in noisy areas
• For military uses
• Commercial applications listening to music

16
Conclusion

• Good passive attenuation at middle and high
  frequencies, and better active control at low
  frequencies
• Good performance of analog control at broadband
  frequencies, and digital control at narrowband
  frequencies for ANC headphones
• Possible combination of analog and digital model
  for use at wide range of frequencies

17
Thank you