Wideband Speech Encoder Inactive Frame Bit Rate Optimization

The patent applies parameter changes by dynamically adjusting the bit rate allocation based on frame activity status. Inactive frames are encoded at a lower bit rate (e.g., 16 bits) while active frames use a higher bit rate (e.g., 171 bits). The coding scheme selector changes coding parameters (wideband vs. narrowband mode) based on the activity detection, optimizing the balance between bit rate reduction and perceptual quality maintenance.

The system dynamically adapts the encoding parameters based on the detected speech activity. The speech activity detector continuously monitors incoming frames and dynamically switches between different coding modes (wideband for active frames, narrowband for inactive frames). This dynamic adaptation allows the system to optimize bit rate usage while maintaining perceptual quality by applying the appropriate encoding strategy in real-time based on the current frame's characteristics.

The patent applies local quality by providing different frequency band representations for different frame types. Wideband spectral envelope descriptions (covering extended frequency ranges up to 7-8 kHz) are applied locally to active frames where speech content requires high intelligibility. For inactive frames, only narrowband spectral descriptions are used, reducing bit rate while maintaining sufficient quality for background noise representation. This localized application of wideband encoding optimizes the trade-off between intelligibility and bit rate.

The speech signal is segmented into active and inactive frames based on speech activity detection. This segmentation allows the system to apply different encoding strategies: wideband encoding for active frames containing speech content and narrowband encoding for inactive frames containing only background noise. By segmenting the signal and applying appropriate encoding to each segment, the system achieves high intelligibility when needed while minimizing bit rate during silence periods.

The speech activity detector performs preliminary analysis of each incoming frame to determine whether it contains speech or only background noise before encoding. This preliminary classification allows the coding scheme selector to pre-select the appropriate encoding mode (wideband or narrowband) based on the frame type. By performing this classification in advance, the system avoids complex real-time decision-making during encoding and simplifies the overall processing architecture while achieving significant bit rate reduction.

The coding scheme selector automatically selects the appropriate coding mode based on the speech activity detection results without requiring external control or complex optimization algorithms. The system serves itself by using the speech activity information to directly determine the encoding parameters, eliminating the need for additional complexity in mode selection logic. This self-service approach simplifies the device architecture while achieving efficient bit rate optimization.