Speech Sequence Filtering With Voiceprint Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech processing methods fail to effectively filter and screen speech sequences due to background noise and multiple voices, leading to unreliable voice transmission.
Innovation Solution
A speech processing method involving framing, speech detection, and voiceprint recognition to identify and forward target speech, using models like LSTM and voiceprint recognition to enhance speech quality and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If TCP speech data transmission mode is used, then transmission reliability is improved, but speech filtering and screening capability deteriorates
Solution Approach 1:
The speech sequence is divided into multiple speech sub-sequences through framing processing, enabling individual detection and filtering of each segment. This segmentation allows the system to identify and remove harmful speech segments while preserving valid speech, resolving the contradiction between maintaining transmission reliability and filtering harmful content.
Solution Approach 2:
A server is introduced as an intermediary between the client and other clients. The server receives speech sequences, performs detection and filtering using speech detection models and voiceprint recognition models, and forwards only valid target speech. This intermediary structure enables effective speech filtering while maintaining reliable transmission through the TCP protocol.
2Ease of operation
If speech sequences are transmitted without filtering, then transmission simplicity is maintained, but speech quality and usability deteriorate
Solution Approach 1:
Speech detection and voiceprint recognition are performed in advance on the server before forwarding speech to other clients. This preliminary filtering action ensures that only valid target speech is transmitted, improving speech quality without requiring complex filtering operations at the receiving end, thus maintaining transmission simplicity.
3Loss of information
If all speech data is forwarded, then completeness of information is maintained, but bandwidth usage and processing efficiency deteriorate
Solution Approach 1:
The system extracts and forwards only the valid target speech segments from the complete speech sequence, removing harmful segments such as background noise and other people's voices. This extraction process maintains the completeness of useful information while significantly reducing bandwidth usage by eliminating redundant and harmful speech data.
4Reliability
If speech filtering is implemented, then speech quality is improved, but system complexity increases
Solution Approach 1:
The server acts as an intermediary that centralizes the complex speech processing functions including framing, speech detection, and voiceprint recognition. This concentrates system complexity in one location while keeping client devices simple, resolving the contradiction between improving speech quality through filtering and maintaining system simplicity.
Data Source
AI summary
A speech processing method and apparatus, a computer storage medium, and an electronic device, relating to the technical field of speech processing. The speech processing method includes: acquiring a speech sequence, obtaining a plurality of speech sub-sequences by performing framing processing on the speech sequence, and extracting a target feature of each speech sub-sequence of the plurality of speech sub-sequences; detecting each speech sub-sequence of the plurality of speech sub-sequences by a speech detection model according to each target feature, and determining valid speech based on a detection result; inputting a target feature corresponding to the valid speech into a voiceprint recognition model, and screening out target speech from the valid speech by the voiceprint recognition model; and controlling the target speech to be forwarded to a client.