xiaooquanwu

Results 23 comments of xiaooquanwu

我这边试了下wav音频文件,也是一样,准确性为0 [call.zip](https://github.com/alphacep/kaldi-android-demo/files/4627235/call.zip)

请注意接听钉钉免费电话

我的是rk3399 android7.1 ,除了更换模型文件,其他代码不需要修改吗? 比如这个地方 KaldiRecognizer(activityReference.get().model, 16000.f, "你 好");

1. 转成单声道,依旧不行 Input File : 'mono.wav' Channels : 1 Sample Rate : 44100 Precision : 16-bit Duration : 00:00:03.20 = 141312 samples = 240.327 CDDA sectors File Size : 283k...

改成这个样子可以设别wav文件了 KaldiRecognizer(activityReference.get().model, 44100 ); 那如果mic需要怎么改?

就是MIC的识别准确性不怎么高,大概只有20%的准确性

public SpeechRecognizer(Model model) throws IOException { this.recognizer = new KaldiRecognizer(model, 16000.0F); this.sampleRate = 16000; this.bufferSize = Math.round((float)this.sampleRate * 0.4F); this.recorder = new AudioRecord(6, this.sampleRate, 16, 2, this.bufferSize * 2); if(this.recorder.getState()...

vosk_api 我这边编译不过android 版本 CMake Error: CMake can not determine linker language for target: kaldi_jni

应该是我在windows android stduio 编译的原因,linux 编译环境没有搭建