WeChat applet Speech Recognition development process records WeChat applet silk to mp3 silk to wav and ffmpeg use, silkffmpeg

Small Program Speech Recognition development process record small program silk to mp3 silk to wav and ffmpeg use, silkffmpeg
Let's talk about the problems encountered in the development of small program speech recognition recently. First, we can use the small program recording control to get the silk format. Later, the official support for the mp3 format, but after we get these formats, can not be directly used, do voice recognition, because the current Baidu Speech Recognition format does not support mp3 format of Baidu php Speech Recognition interface http://yuyin.baidu.com/docs/asr/188 but there are support wav format, so I had to find a way to convert the mp3 format to wav format. I searched Baidu for a large circle and finally learned to use an ffmpeg component. So I continued to search for information to see how to install ffmpeg on Linux servers. if you use php to call the transcoding command, after transcoding is successful, you can pass the transcoded wav format to Baidu's Speech Recognition interface. Then, after Baidu uploads the audio to him, I am not telling you the recognition result right away. I still need to call back to you. So you need to make a callback interface, get the recognition result in the callback interface, and then get the recognition result for comparison, in this way, you can make a speech. Sound Recognition has gone through a lot of learning procedures and has refined a lot of technologies. The key is tossing. I learned a lot from this meal. I, Qiu Feng, and phper. I am a startup, and I am working on project system development. I am learning from each other. I am improving my account: qiufeng2983.