A selection method and storage medium of a modem voice codec
A codec, codec technology, applied in the selection method of modem voice codec and the field of storage media, can solve the problems of silent call drop, noise, voice frame jitter, etc.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0069] The present invention provides a method for selecting a voice codec of a modem. In this embodiment, the method is applied to a personal terminal. figure 1 It is the flowchart of the selection method of the modem speech codec of the embodiment of the present invention, below in conjunction with figure 1 Each step of the method is described in detail. Such as figure 1 As shown, the method mainly includes the following steps:
[0070] S1: Select a plurality of characteristic parameters used to form training data from the BP side and the AP side, and set the value range of each characteristic parameter.
[0071] Specifically, the characteristic parameters include: the signal-to-noise ratio (Signal to Noise Ratio / Signal to Interferenceplus Noise Ratio, SNR / SINR) of the physical layer (Physical Layer, PHY) on the baseband processor (Base-Band Processor, BP) side, the received signal Channel Power (Receive Signal Channel Power, RSCP), Received Signal Strength Indicator (Rec...
Embodiment 2
[0109] In this embodiment, on the basis of Embodiment 1, step S4 also includes the process of using the comprehensive tester to provide test data and fine-tuning the speech codec classifier obtained in Embodiment 1 to prevent overfitting of training data together affect the training results. image 3 It is a flowchart of obtaining a trained speech codec classifier according to Embodiment 2 of the present invention. Such as image 3 As shown, step S4 specifically includes:
[0110] S41: Based on the training data, through preliminary training, obtain a first speech codec classifier and a first bit rate prediction regressor after preliminary training;
[0111] S42: Under the test network, respectively predict the corresponding first speech codec and the first bit rate through the first speech codec classifier and the first bit rate prediction regressor;
[0112] S43: Based on the type of the test network, perform a matching test on the first speech codec and the first bit rat...
Embodiment 3
[0118] In this embodiment, different from the second embodiment, the process of matching test is not included in step S4, but the effect of preventing over-fitting of the training data is also achieved. Figure 4 It is a flow chart of obtaining a trained speech codec classifier according to Embodiment 3 of the present invention. Such as Figure 4 As shown, step S4 specifically includes:
[0119] S41": Based on the training data, through preliminary training, obtain the first speech codec classifier after preliminary training;
[0120] S42": Provide test data through the comprehensive tester, fine-tune the first speech codec classifier, obtain the fine-tuned fifth speech codec classifier, and use the fifth speech codec classifier as the trained speech codec classifier Decode the classifier.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


