Active removal of wind noise from outdoor microphones using local velocity measurements

Wind noise in outdoor microphone measurements can significantly degrade acoustic data. Since acoustic and wind spectra often overlap when low frequency acoustic sources are observed, wind noise is generally impossible to remove by band selective filtering. This research examines active removal of wind noise from an outdoor microphone.; Measurements performed outdoors on a horizontally positioned microphone demonstrate that a major part of unscreened microphone wind noise is flow noise caused by interaction between the microphone and adjacent air flow. It is concluded that pressure fluctuations in the passing air volume are a second significant contributor. While wind noise contamination is most often treated with passive spherical foam windscreens, an active wind noise reduction system based upon local velocity measurements applied to a passively screened or unscreened microphone is proposed.; A simple relationship, developed from the Bernoulli equation, is used to transform these wind velocity measurements into microphone wind pressure estimates. Extensive measurements of coherence between these anemometer derived pressure estimates and wind noise in screened and unscreened microphones are performed for a variety of anemometer positions. These measurements indicate the dominant source of wind noise in a passively screened microphone is local pressure fluctuations that are not correlated with pressure estimates based on local velocity fluctuations. This important observation indicates than an active system based upon velocity measurements is unlikely to quiet a microphone that has convential passive screening by a substantial amount and that an active system based upon velocity measurements applied to an unscreened microphone is unlikely to surpass the performance of a passively screened microphone.; An outdoor active wind noise cancellation system is demonstrated for a horizontally positioned unscreened microphone. A signal space approach is used to determine the effectiveness of single and multiple anemometer channels in the system and it is shown that while additional quieting is obtained with additional anemometers the additional quieting achieved diminishes rapidly with new channels. The active system transfer functions are observed to possess significant stability of over periods of tens of minutes to hours. This stability apparently depends upon the stability of the wind environment. It is concluded that adaptive active quieting systems would require only infrequent updates in such an environment. System performance with a four channel anemometer system is observed to be similar to that using a small passive screen. Wind energy reduction performance of the active windscreen is shown to be between 10 and 13 dB for frequencies below 200Hz and this performance decreases smoothly to approximately 5 dB at 500 Hz. The active system has application when the presence or physical size of the passive windscreen is undesirable. Further, it is concluded that the concepts presented also provide an effective approach for studying the behavior of passive wind screens and outdoor microphone wind noise.