科大讯飞语音转文字以及中文分词的 Java 测试代码
发布于: 2021 年 05 月 24 日
我录了一段音存储在这个 test.m4a 文件里，语音内容为"测试一下 Netweaver 对于并发请求的响应性能"。
使用如下 Java 代码进行测试：
package com.iflytek.msp.lfasr;
import java.util.HashMap;import org.apache.log4j.Logger;import com.alibaba.fastjson.JSON;import com.iflytek.msp.cpdb.lfasr.client.LfasrClientImp;import com.iflytek.msp.cpdb.lfasr.exception.LfasrException;import com.iflytek.msp.cpdb.lfasr.model.LfasrType;import com.iflytek.msp.cpdb.lfasr.model.Message;import com.iflytek.msp.cpdb.lfasr.model.ProgressStatus;
// SDK document: http://www.xfyun.cn/doccenter/lfasr#go_sdk_doc_v2public class TestLfasr { // original media path private static final String local_file = "c:\\temp\\test.m4a";
private static final LfasrType type = LfasrType.LFASR_STANDARD_RECORDED_AUDIO; private static int sleepSecond = 20; public static void main(String[] args) { LfasrClientImp lc = null; try { lc = LfasrClientImp.initLfasrClient(); } catch (LfasrException e) { Message initMsg = JSON.parseObject(e.getMessage(), Message.class); System.out.println("ecode=" + initMsg.getErr_no()); System.out.println("failed=" + initMsg.getFailed()); } // get upload task id String task_id = ""; HashMap<String, String> params = new HashMap<>(); params.put("has_participle", "true"); try { Message uploadMsg = lc.lfasrUpload(local_file, type, params); int ok = uploadMsg.getOk(); if (ok == 0) { task_id = uploadMsg.getData(); System.out.println("task_id=" + task_id); } else { System.out.println("ecode=" + uploadMsg.getErr_no()); System.out.println("failed=" + uploadMsg.getFailed()); } } catch (LfasrException e) { Message uploadMsg = JSON.parseObject(e.getMessage(), Message.class); System.out.println("ecode=" + uploadMsg.getErr_no()); System.out.println("failed=" + uploadMsg.getFailed()); } while (true) { try { Thread.sleep(sleepSecond * 1000); System.out.println("waiting ..."); } catch (InterruptedException e) { } try { Message progressMsg = lc.lfasrGetProgress(task_id); if (progressMsg.getOk() != 0) { System.out.println("task was fail. task_id:" + task_id); System.out.println("ecode=" + progressMsg.getErr_no()); System.out.println("failed=" + progressMsg.getFailed()); continue; } else { ProgressStatus progressStatus = JSON.parseObject(progressMsg.getData(), ProgressStatus.class); if (progressStatus.getStatus() == 9) { System.out.println("task was completed. task_id:" + task_id); break; } else { System.out.println("task was incomplete. task_id:" + task_id + ", status:" + progressStatus.getDesc()); continue; } } } catch (LfasrException e) { Message progressMsg = JSON.parseObject(e.getMessage(), Message.class); System.out.println("ecode=" + progressMsg.getErr_no()); System.out.println("failed=" + progressMsg.getFailed()); } } try { Message resultMsg = lc.lfasrGetResult(task_id); System.out.println(resultMsg.getData()); if (resultMsg.getOk() == 0) { System.out.println(resultMsg.getData()); } else { System.out.println("ecode=" + resultMsg.getErr_no()); System.out.println("failed=" + resultMsg.getFailed()); } } catch (LfasrException e) { Message resultMsg = JSON.parseObject(e.getMessage(), Message.class); System.out.println("ecode=" + resultMsg.getErr_no()); System.out.println("failed=" + resultMsg.getFailed()); } }}
复制代码
测试结果
(1) 所有中文均能成功转成文字; 但英文 Netweaver 的语音转换成了 Net ball
(2) 智能分词也能按照期望工作，比如“测试一下”成功地分词成了“测试”和“一下”。
完整的 Java 项目在我的 github 上：https://github.com/i042416/voice2text
划线
评论
复制
发布于: 2021 年 05 月 24 日阅读数: 41
版权声明: 本文为 InfoQ 作者【Jerry Wang】的原创文章。
原文链接:【http://xie.infoq.cn/article/790e10f1b8d380a22f34c7393】。
本文遵守【CC-BY 4.0】协议，转载请保留原文出处及本版权声明。
Jerry Wang
关注
个人微信公众号：汪子熙 2017.12.03 加入
SAP成都研究院开发专家，SAP社区导师，SAP中国技术大使。
评论