Multichannel Speech Signal Compression via Reference Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech signal transmission and reception systems face challenges in efficiently compressing and synchronizing multichannel speech signals, leading to low compression efficiency due to variations in power and synchronization between microphones.
Innovation Solution
A method that adjusts powers and synchronization of multichannel speech signals using correlation between microphones, sets a reference speech signal, calculates power and synchronization parameters, generates offset signals, and encrypts and compresses the signals for transmission, and includes a receiver to restore the signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multichannel speech signals are transmitted directly without processing, then transmission quality is maintained, but transmission bandwidth and data volume increase significantly
Solution Approach 1:
The patent extracts only the essential characteristics from multichannel speech signals - specifically power parameters and synchronization parameters - rather than transmitting the complete signals. By separating and transmitting only these critical features along with reference speech signals, the system achieves significant data reduction while preserving speech quality for reconstruction at the receiver端.
Solution Approach 2:
The patent transforms the speech signals into different parameter representations - converting time-domain signals into power parameters and synchronization parameters. This parameter transformation enables efficient compression by transmitting compact numerical representations instead of raw audio data, reducing bandwidth requirements while maintaining reconstructability.
2Ease of manufacture
If speech signals from multiple microphones are compressed individually, then processing simplicity is maintained, but compression efficiency is low due to power and synchronization variations
Solution Approach 1:
The patent merges the processing of multichannel speech signals by identifying a reference speech signal from one microphone and using it as a basis for processing signals from other microphones. By combining the reference signal with power parameters and synchronization parameters from other channels, the system achieves efficient joint compression that exploits correlations between microphones while maintaining processing organization.
Solution Approach 2:
The patent performs preliminary processing steps before compression - specifically calculating power parameters and synchronization parameters from the multichannel signals, and selecting a reference speech signal in advance. These preliminary actions prepare the data in an optimized format that enables efficient subsequent compression and transmission.
3Adaptability or versatility
If the number of processing channels in DSP is increased, then multichannel processing capability is improved, but memory requirements increase due to additional executable code
Solution Approach 1:
The patent changes the representation of speech signals from time-domain waveforms to compact parameter forms (power parameters and synchronization parameters). This parameter transformation reduces the data volume that needs to be processed and stored, enabling the DSP to handle more channels within fixed memory constraints by working with compressed parameter representations rather than full audio buffers.
Solution Approach 2:
The patent extracts only the essential features from each channel - power parameters and synchronization parameters - rather than processing complete audio buffers. By taking out only these critical characteristics, the system reduces memory requirements for executable code and data structures while maintaining the ability to reconstruct and process multichannel speech signals.
4Productivity
If speech signals are synchronized and power-adjusted using correlation between microphones, then compression efficiency is improved, but processing complexity increases
Solution Approach 1:
The patent implements self-service processing where the reference speech signal itself is used to determine the synchronization and power adjustment parameters for other channels. By using the reference signal as the basis for calculating correlation and deriving adjustment parameters, the system achieves efficient synchronization and normalization without requiring external calibration or complex manual configuration.
Data Source
AI summary
A speech signal transmission apparatus includes an extractor to extract speech signals from speech source signals collected by a plurality of microphones, a power calculator to calculate powers of speech signals of multiple channels and set any one of the speech signals of the multiple channels as a reference speech signal, a synchronization adjustor to adjust synchronization of the other speech signals based on the reference speech signal, a signal generator to generate extraction signals by offsetting the reference speech signal from the other synchronization-adjusted speech signals, an encryptor to compress and encrypt the reference speech signal and the extraction signals, and a transmitter to transmit the compressed and encrypted reference speech signal and extraction signals.