| c |
|
 |
|
Adaptive Dual Microphone (ADM) is a digital signal processing technology creating a directional or noise-canceling microphone using just two omnidirectional microphones. ADM automatically changes its directional characteristics to provide the best noise attenuation in varying environments while fully preserving the signal quality. The adaptation process is very fast and frequency selective so that multiple interferences are cancelled simultaneously. In addition to its superior directional characteristics, ADM is less effected by wind generated noise than conventional acoustic directional microphones. It has no proximity effect and it is easy to integrate into mobile voice communication and other devices. ADM technology allows two types of microphones' configurations: "end-fire" and "broadside". In the end-fire configuration, the signal source (user’s mouth) is intended to be on the axis (line connecting the two microphones). In the broadside configuration it is aimed on the line transverse to the axis. In the end-fire configuration, ADM has two modes of operation; "far-talk" and "close-talk". In the far-talk mode, ADM functions as an optimal directional microphone that attenuates signals arriving from the back and side directions while preserving the front signal. It is ideal for applications like conference, reporter and surveillance microphones, camcorders, and hearing aids, assistive listening devices and mobile phones in "speakerphone mode".
In the close-talk mode, ADM works as the best possible noise-canceling microphone creating a virtual "bubble of silence" around it effectively eliminating distant sounds while preserving sounds in the close-talk “bubble”. Relative freedom of acoustic design makes ADM ideal for mobile handsets with "soft" switch between speakerphone (far) and handset (close) modes of operation. |
|
All processing is done in frequency sub-bands meaning that the input full-band signals are split into narrow frequency bands, processed individually and combined into the full-band signal again afterwards. This greatly improves the system performance as the signal and noise properties may differ significantly in different frequency regions. The functionality of the sub-band processing blocks depends on the operational mode. In the far-talk mode, the signal and noises are differentiated by the direction of arrival. The delays are chosen to attenuate all signals except the signal from the main direction. This is achieved by limiting the range of values that delays can take by the values corresponding to noise directions. In the close-talk mode noises can come from any direction, but from a large distance. The signal is supposed to come from a close distance creating a significantly higher sound pressure level on the main microphone. ADM works to eliminate all sounds that create equal sound pressure level (with some specified tolerance) on the two sound pressure sensors. Thus all distant sounds coming from all directions are cancelled. Simultaneously, close sounds that create significant sound pressure difference are preserved without any distortion.
Use control buttons on the interactive screenshots to play/stop/navigate |
Two examples of real-life recordings made in noisy environments using dual-microphone headset device.
|
UNPROCESSED (RAW) SIGNAL,
as captured by the first microphone |
OUTPUT SIGNAL,
processed using Alango ADM |
|
UNPROCESSED (RAW) SIGNAL,
as captured by the first microphone |
OUTPUT SIGNAL,
processed using Alango ADM |
|
Mobile phones, Bluetooth headsets, hands-free kits, video/audio conferencing equipment, voice recorders, intercom systems, voice controlled devices and many others.
Alango Adaptive Dual Microphone can be integrated with other Alango technologies into one Voice Communication Package (VCP). VCP includes acoustic echo cancellation, stationary and transient noise suppression, dynamic range compressor, automatic volume and equalization control (optional), speech enhancement and several other functions.
114th AES Convention paper (PDF) 116th AES Convention paper (PDF) 117th AES Convention paper (PDF) Presentation given at IBM speech seminar, June 2003, Haifa, Israel (PPS) |