RealTime Active Noise Control Using TI C54

1
Real-Time Active Noise Control Using TI C54

• Ryanth Atmadja
• CIT Honors Research Project
• Advisor Prof. Tsuhan Chen
• Summer 1999

2
Motivation

• Real-time noise reduction.
• DSP-based implementation with a low-powered
  fixed-point microprocessor (Texas Instruments
  TMS320C54.)

3
What is active noise control?

• Noise attenuation by means of destructively
  interfering the target noise with an anti-noise
  generated by an ANC controller.
• The anti-noise is the target noise with opposite
  sign.
• Some ANC applications
• active ear protectors, automobiles (interior
  cabin and exhaust pipe), air ducts.

4
Why active noise control?

• Passive means to block noise, e.g., physical
  enclosures, and barriers, can not effectively
  reduce low-frequency noise (lt 500 Hz), but ANC
  can, provided that the noise is predictable.
• To reduce low-frequency noises the passive
  barriers tend to be bulky, whereas ANC systems
  come as compact packages of DSP controllers,
  speakers, and microphones. Therefore ANC is the
  cost-effective choice.

5
Single-Channel Adaptive Feedback ANC

• The ANC controller predicts the incoming target
  noise and then sends a matching anti-noise
  through a loudspeaker.
• The goal of the ANC controller is to reduce the
  signal picked up by the microphone (i.e., reduce
  the error signal en.)

6
Feedback ANC Algorithm and Secondary-Path Modeling
Phase 1 Offline Secondary- path S(z) modeling.
(Figure taken from S. M. Kuo and D. R.
Morgan, Active Noise Control Systems
Algorithms and DSP Implementations. New York
Wiley, 1996.)
Phase 2 Feedback ANC algorithm
7
Hardware Setup
two-way
communication
Host PC
TI C54 evaluation
Sound
Analog
- User interface
module (EVM)
Card
Interface
- Data collection
- ANC digital
Circuit (AIC)
and analysis
signal processing
microphone
en
yn
S(z)
yn
dn
anti-noise speaker
target noise
speaker
8
Quantitative Result Spectra of the Test Noises
Quantitative ANC performance sinusoid gt fan gt
autoregressive
9
Subjective Testing Result

• The ANC system works best for sinusoids
  (qualitative ANC performance sinusoid gt
  autoregressive gt fan.)
• The reduced noises do not always sound pleasing
  -- this is due to the additional noises that the
  ANC system creates (see plots.)

10
Conclusions and Future Work

• The feedback ANC implementation reduces noise in
  real-time, especially for sinusoids or narrowband
  noises.
• Low-powered DSP-based implementation makes it
  possible to build ANC as an embedded product.
• Possible causes of the additional noise problem
  (to be addressed in future work)
• Mediocre precision of the fixed-point C54
  implementation.
• Secondary-path modeling S(z) is not good.
• May hand-assemble the ANC code for efficiency.
• Other ANC schemes that use more than one
  microphone (feedforward and hybrid ANCs) can be
  explored.