A method and device for locating voice content in batches

Abstract

The invention discloses a method for batch positioning of voice content. The method comprises the following steps: S1, playing a prompt tone before recording starts, and recording a path of stored audio files into a path recording file; S2, traversing and reading all paths recorded in the path recording file; S3, carrying out prompt tone detection and positioning on the content of the front-part time of audio files; S4, carrying out prompt tone detection on the audio file of which the prompt tone is detected and part of audio segments are deleted again, and if the audio file of the prompt toneis not detected again, storing the file until no new prompt tone is detected; and S5, after all the detected audio files are processed, ending. The invention further discloses a device for batch positioning of the voice content. The voice recording positions in all the audio files can be found out by accurately positioning the prompt tone positions, and the voice corpus file processing speed is increased.

Description

technical field [0001] The invention belongs to the technical field of voice intelligent recognition, relates to corpus recognition technology, in particular to a method and device for locating voice content in batches. Background technique [0002] In the existing field of artificial intelligence, speech recognition is becoming more and more mature, and most of the development of artificial intelligence is based on speech recognition and processing. However, due to the research and development of speech recognition, it needs to be based on a large amount of corpus. However, the normal recording process is not full of effective content, and there is a large amount of redundant information. The large amount of corpus and the complicated redundant information have become obstacles in the research and development of speech recognition. [0003] There are following disadvantages in the prior art to the method for corpus redundancy processing: [0004] 1. The corpus of differen...

AI Technical Summary

Problems solved by technology

[0004] 1. The corpus of different frequencies and different channels needs to be processed separately;

[0005] 2. The corpus processing needs to be configured with the same path template, and it cannot process audio with different formats stored in the directory;

[0006] 3. The audio processing speed is slow;

[0007] 4. Audio positioning is easily affected by background noise, resulting in inaccurate positioning;

[0008] 5. Lack of repeatability detection

Images