CN105469656A - A spoken language learning system and its operation method - Google Patents

Abstract

The invention relates to a spoken language learning system, comprising: the audio decoding module is connected with the voice breakpoint searching module, and the voice breakpoint searching module is connected with the audio playing module. The spoken language learning system effectively solves the problem of simultaneously and interactively training listening and speaking abilities in English learning.

Description

A spoken language learning system and its operation method

technical field

The invention relates to a spoken language learning system and an operation method of the system.

Background technique

The learning of spoken English needs repeated listening and speaking training in order to improve learning efficiency. At present, people generally use a one-way audio player for the English audio files purchased or downloaded. Users can only train their ability to listen, but cannot train their ability to speak in time.

In view of this, it is indeed necessary to provide a system that allows users to use commonly used audio files or network audio to implement comprehensive training in listening, speaking, and confirmation, so as to improve learning efficiency.

Contents of the invention

A kind of spoken language learning system that the present invention provides in order to solve the above-mentioned problem, comprises: the audio frequency decoding module that is used for audio file decoding, is used for automatic calculation and seeks the speech interruption point search module of speech breakpoint in audio frequency, is used for playing and replaying audio frequency An audio playback module for data, an adaptive recording module for adaptively recording the user's voice, and a recording playback module for playing back recordings, the audio decoding module is connected with the voice interruption point search module, and the voice interruption The point search module is connected with the audio playing module.

Preferably, the audio decoding module supports decoding of audio files such as MP3 or MVA or online audio streams.

Preferably, the audio decoding module supports reading decoding data of any length each time.

Preferably, the adaptive recording module has a noise reduction module supporting speech noise reduction processing.

Preferably, the adaptive recording module saves the voice in a recording file MicFile, and the recording playback module can automatically trigger the recording file MicFile.

The present invention also provides an operation method of the above-mentioned language learning system, the operation method comprising:

Step 1, the audio decoding module decodes the audio file;

Step 2, the voice interruption point search module automatically calculates and finds the voice interruption point in the audio;

Step 3, the audio playback module plays and replays audio data;

Step 4, the adaptive recording module adaptively records the voice of the user;

Step 5. The recording playback module plays back the user's recording.

Preferably, in the step 2, the speech interruption point search module searches for the speech interruption point in the decoded data stream based on the entire audio data buffer or based on automatic calculation of part of the data stream.

Preferably, in the step 2, the speech break point search module uses an energy threshold speech break point detection algorithm.

Preferably, in said step 4, if there is no effective voice in the first time length T1, then automatically end the recording; if there is an effective voice in the first time length T1, then enter the silent section judgment, if last a second 2. If there is a silent segment in the time length T2, the recording will end automatically.

Preferably, after said step 5, the following steps are also included:

Step 6. The audio decoding module and the voice breakpoint search module perform subsequent data decoding and breakpoint detection.

The beneficial effect of the present invention is that: the spoken language learning system effectively solves the problem of simultaneous interactive training of listening and speaking abilities in English learning. As long as it is based on ordinary audio files or network audio streams, it can realize the sentence-by-sentence cycle training of listening, repeating, and confirming. In addition, it supports the single-sentence repeat playback function, which can significantly improve the learning efficiency of oral English.

Description of drawings

FIG. 1 is a schematic diagram of a learning system framework provided by an embodiment of the present invention.

detailed description

The present invention will be further elaborated below in conjunction with accompanying drawing:

The invention provides a spoken language learning system. The input object of the spoken language learning system is audio files, which are mainly speech and do not include continuous background music.

As shown in Figure 1, this spoken language learning system comprises audio decoding module, is used for the decoding of audio file; Speech interruption point search module, is used for automatic calculation and finds the speech interruption point in audio frequency; Audio playback module, is used for playing and replaying audio frequency data; an adaptive recording module for adaptively recording the user's voice; a recording playback module for playing back the user's recording.

The audio decoding module is connected with the voice interruption point search module, and transmits the decoded decoded data stream to the voice interruption point search module. The voice interruption point search module is connected with the audio playback module, and transmits the data of the voice segment to the audio playback module.

The present invention also provides the operation method of the above-mentioned spoken language learning system, comprising the following steps:

Step 1, the audio decoding module decodes the audio file;

Step 2, the voice interruption point search module automatically calculates and finds the voice interruption point in the audio;

Step 3, the audio playback module plays and replays audio data;

Step 4, the adaptive recording module adaptively records the voice of the user;

Step 5. The recording playback module plays back the user's recording.

The audio decoding module supports the decoding processing of audio files such as MP3 or MVA, and also supports the decoding of online audio streams, and supports reading decoded data of any length each time. For different platforms, you can choose an appropriate cache size, and read the decoded data PcmData of an appropriate length each time.

The speech interruption point search module can be based on the entire audio data cache, or can automatically calculate and find the speech interruption point in the decoded data stream based on part of the data stream. The algorithm used includes but is not limited to commonly used energy threshold speech breakpoint detection algorithms. For example: based on the previously obtained decoded data PcmData, calculate speech energy and zero-crossing rate in units of 20ms or 40ms frames, and then judge whether there is a speech interruption point through sliding window and threshold judgment. If there is a voice break point, then record the break point information, and start the recording module after the audio playback module plays the voice clip. If there is no voice interruption point, the data is directly passed to the audio playback module to play the voice.

The voice playback module plays directly after receiving the previous data, and automatically stops playing if there is no data. The voice playback module can play the voice segment data output by the previous voice interruption point search module; it can also play a specified voice segment repeatedly.

The self-adaptive recording module can self-adaptively control the length of recording time and save the recording of the user's voice input as an audio file. At the same time, the self-adaptive recording module has a noise reduction module to support speech noise reduction processing. The algorithms for adaptive control duration include but are not limited to voice endpoint detection, adaptive silence segment length control, and the like. After the adaptive recording module receives the start command, it starts the recording process. The adaptive recording module caches the data MicData output by the microphone device, saves it in a recording file MicFile, and performs breakpoint detection on the data MicData. If there is no valid voice within the first duration T1, the recording is automatically terminated. If a valid voice appears within the first time length T1, it will enter the silent segment judgment, and if the silent segment occurs for the second time length T2, the recording will be automatically ended. After the recording is received, the recording playback module is automatically started.

The recording playback module can automatically trigger the playback of the user's recording file MicFile, which is used for the user's self-confirmation of his repeated voice. The recording playback module starts to play the recording file MicFile after receiving the instruction. After the playback is completed, the following steps are included: notify the audio decoding module and the voice breakpoint search module to perform subsequent data decoding and breakpoint detection.

During this period, if the user inputs an instruction, the audio decoding module is notified to start decoding data from the previously saved breakpoint position.

The spoken language learning system effectively solves the problem of simultaneous interactive training of listening and speaking abilities in English learning. As long as it is based on ordinary audio files or network audio streams, it can realize the sentence-by-sentence cycle training of listening, repeating, and confirming. In addition, it supports the single-sentence repeat playback function, which can significantly improve the learning efficiency of oral English.

The above-described embodiments are only preferred examples of the present invention, and are not intended to limit the scope of the present invention, so all equivalent changes or modifications made according to the structure, features and principles described in the patent scope of the present invention should be Included in the patent application scope of the present invention.

Claims (10)

1. A spoken language learning system, characterized in that, said spoken language learning system comprises: an audio decoding module for audio file decoding, a voice breakpoint search module for automatically calculating and finding a voice breakpoint in the audio frequency, for playing and an audio playback module for playing back audio data, an adaptive recording module for adaptively recording the user's voice, and a recording playback module for playing back the recording, The audio decoding module is connected with the voice interruption point search module, and the voice interruption point search module is connected with the audio playback module. 2. The spoken language learning system as claimed in claim 1, wherein the audio decoding module supports the decoding of audio files such as MP3 or MVA or online audio streams. 3. The spoken language learning system according to claim 1, wherein the audio decoding module supports reading decoding data of any length each time. 4. The spoken language learning system as claimed in claim 1, 2 or 3, wherein the adaptive recording module has a noise reduction module supporting speech noise reduction processing. 5. The spoken language learning system as claimed in claim 4, wherein the adaptive recording module saves the voice into a recording file MicFile, and the recording playback module can automatically trigger the recording file MicFile. 6. A method of operation of any spoken language learning system as claimed in claims 1-5, wherein said method of operation comprises: Step 1, the audio decoding module decodes the audio file; Step 2, the voice interruption point search module automatically calculates and finds the voice interruption point in the audio; Step 3, the audio playback module plays and replays audio data; Step 4, the adaptive recording module adaptively records the voice of the user; Step 5. The recording playback module plays back the user's recording. 7. The operation method according to claim 6, wherein in the step 2, the voice interruption point search module is based on the entire audio data cache or based on partial data stream automatic calculation to find the decoded data stream inside Speech break point. 8. The operation method according to claim 6, characterized in that, in the step 2, the speech break point search module uses an energy threshold speech break point detection algorithm. 9. The method of operation as claimed in claim 6, wherein in said step 4, if there is no valid voice within the first duration T1, the recording is automatically terminated; If the voice is valid, it enters the judgment of the silent segment, and if the silent segment occurs for a second duration T2, the recording is automatically ended. 10. The operation method according to claim 6, characterized in that, after said step 5, further comprising the following steps: Step 6. The audio decoding module and the voice breakpoint search module perform subsequent data decoding and breakpoint detection.