miniDSP Ltd, Hong Kong /
/ Features and specifications subject to change without prior notice
18
7.1.1
UMA-8 DSP MODE
Bypass_on [0,1] / Modes for DSP processing
0: normal operation / Both microphone processing and speaker processing is enabled
1: bypass microphone processing, speaker processing remains
2: pure bypass where both the speaker processing AND microphone processing is disabled
Mic_shift [0..4]: left shift of microphone input signal can be used to boost the gain of the MEMS. Use this
caution as it could create internal DSP saturation*/
7.1.2
BEAMFORMER
The UMA-8 is using a 2ch highly directive end-fire beamformer to isolate voice from noise. The beamformer uses
a time-frequency adaptive LMS algorithm with differential microphone pre-processing. The beamformer is
suitable for non-stationary sound stations, and integrates seamlessly with AEC and NS algorithms.
BF: [0, 1]: 0: Beamformer off, Beamformer 1: on
BF_direction: [0..3]: 0: 360 degree , 1: end-fire 1, 2: end-fire 2, 3: line array
BF_focus [0..10]:Beamformer focus controls the endfire polar pattern of the microphone. See below for chart
BF_diffgain_dB; /* [-20..0]: diffuse sound gain [dB] controls the de-reverb processing.
