利用百度语音识别技术实现文字转语音的应用（Java版附源码）

迷南。 2021-08-31 02:10 671阅读 0赞

# 一、基本介绍 #

语音交互的需求场景越来越多，很多开发者都想做一款把语音转换成文字的游戏或者应用，这需要用到语音识别技术，这里会对如何使用完全免费的百度语音开放平台来快速开发，拥有高大上的语音识别技术的应用。

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2d1b3J1aV9qYXZh_size_16_color_FFFFFF_t_70][]

# 二、引入jar包 #

下载jacob-1.18.zip  
并导入jacob.jar、json-20160810.jar、log4j-1.2.17.jar  
将解压后的文件中jacob-1.18-x64.dll复制到对应的JDK中（我的是C:Program FilesJavajdk1.8.0\_152jrein）

# 三、代码实例 #

## 1、工具类，ConnUtil ##

package baidu.restapi.common;
     
    import java.io.ByteArrayOutputStream;
    import java.io.IOException;
    import java.io.InputStream;
    import java.io.UnsupportedEncodingException;
    import java.net.HttpURLConnection;
    import java.net.URLEncoder;
     
    /**
     * 与连接相关的Util类
     */
    public class ConnUtil {
     
        /**
         * UrlEncode， UTF-8 编码
         *
         * @param str 原始字符串
         * @return
         */
        public static String urlEncode(String str) {
            String result = null;
            try {
                result = URLEncoder.encode(str, "UTF-8");
            } catch (UnsupportedEncodingException e) {
                e.printStackTrace();
            }
            return result;
        }
     
        /**
         * 从HttpURLConnection 获取返回的字符串
         *
         * @param conn
         * @return
         * @throws IOException
         * @throws DemoException
         */
        public static String getResponseString(HttpURLConnection conn) throws IOException, DemoException {
            return new String(getResponseBytes(conn));
        }
     
        /**
         * 从HttpURLConnection 获取返回的bytes
         * 注意 HttpURLConnection自身问题， 400类错误，会直接抛出异常。不能获取conn.getInputStream();
         *
         * @param conn
         * @return
         * @throws IOException   http请求错误
         * @throws DemoException http 的状态码不是 200
         */
        public static byte[] getResponseBytes(HttpURLConnection conn) throws IOException, DemoException {
            int responseCode = conn.getResponseCode();
            if (responseCode != 200) {
                System.err.println("http 请求返回的状态码错误，期望200， 当前是 " + responseCode);
                if (responseCode == 401) {
                    System.err.println("可能是appkey appSecret 填错");
                }
                throw new DemoException("http response code is" + responseCode);
            }
     
            InputStream inputStream = conn.getInputStream();
            byte[] result = getInputStreamContent(inputStream);
            return result;
        }
     
        /**
         * 将InputStream内的内容全部读取，作为bytes返回
         *
         * @param is
         * @return
         * @throws IOException @see InputStream.read()
         */
        public static byte[] getInputStreamContent(InputStream is) throws IOException {
            byte[] b = new byte[1024];
            // 定义一个输出流存储接收到的数据
            ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream();
            // 开始接收数据
            int len = 0;
            while (true) {
                len = is.read(b);
                if (len == -1) {
                    // 数据读完
                    break;
                }
                byteArrayOutputStream.write(b, 0, len);
            }
            return byteArrayOutputStream.toByteArray();
        }
    }

## 2、token的获取类，TokenHolder   ##

package baidu.restapi.common;
     
    import org.json.JSONObject;
     
    import java.io.IOException;
    import java.net.HttpURLConnection;
    import java.net.URL;
     
    /**
     * token的获取类
     * 将apiKey和secretKey换取token，注意有效期保存在expiresAt
     */
    public class TokenHolder {
     
        public static final String ASR_SCOPE = "audio_voice_assistant_get";
     
        public static final String TTS_SCOPE = "audio_tts_post";
     
        /**
         * url , Token的url，http可以改为https
         */
        private static final String url = "http://openapi.baidu.com/oauth/2.0/token";
     
        /**
         * asr的权限 scope 是  "audio_voice_assistant_get"
         * tts 的权限 scope 是 "audio_tts_post"
         */
        private String scope;
     
        /**
         * 网页上申请语音识别应用获取的apiKey
         */
        private String apiKey;
     
        /**
         * 网页上申请语音识别应用获取的secretKey
         */
        private String secretKey;
     
        /**
         * 保存访问接口获取的token
         */
        private String token;
     
        /**
         * 当前的时间戳，毫秒
         */
        private long expiresAt;
     
        /**
         * @param apiKey    网页上申请语音识别应用获取的apiKey
         * @param secretKey 网页上申请语音识别应用获取的secretKey
         */
        public TokenHolder(String apiKey, String secretKey, String scope) {
            this.apiKey = apiKey;
            this.secretKey = secretKey;
            this.scope = scope;
        }
     
        /**
         * 获取token，refresh 方法后调用有效
         *
         * @return
         */
        public String getToken() {
            return token;
        }
     
        /**
         * 获取过期时间，refresh 方法后调用有效
         *
         * @return
         */
        public long getExpiresAt() {
            return expiresAt;
        }
     
        /**
         * 获取token
         *
         * @return
         * @throws IOException   http请求错误
         * @throws DemoException http接口返回不是 200, access_token未获取
         */
        public void resfresh() throws Exception {
            String getTokenURL = url + "?grant_type=client_credentials"
                    + "&client_id=" + ConnUtil.urlEncode(apiKey) + "&client_secret=" + ConnUtil.urlEncode(secretKey);
     
            // 打印的url出来放到浏览器内可以复现
            System.out.println("token url:" + getTokenURL);
     
            URL url = new URL(getTokenURL);
            HttpURLConnection conn = (HttpURLConnection) url.openConnection();
            conn.setConnectTimeout(5000);
            String result = ConnUtil.getResponseString(conn);
            System.out.println("Token result json:" + result);
            parseJson(result);
        }
     
        /**
         * @param result token接口获得的result
         * @throws DemoException
         */
        private void parseJson(String result) throws Exception {
            JSONObject json = new JSONObject(result);
            if (!json.has("access_token")) {
                // 返回没有access_token字段
                throw new DemoException("access_token not obtained, " + result);
            }
            if (!json.has("scope")) {
                // 返回没有scope字段
                throw new DemoException("scopenot obtained, " + result);
            }
            if (!json.getString("scope").contains(scope)) {
                throw new DemoException("scope not exist, " + scope + "," + result);
            }
            token = json.getString("access_token");
            expiresAt = System.currentTimeMillis() + json.getLong("expires_in") * 1000;
        }
    }

## 3、异常类，DemoException  ##

package baidu.restapi.common;
     
    public class DemoException extends Exception {
        public DemoException(String message) {
            super(message);
        }
    }

## 4、测试类，TtsMain  ##

package baidu.restapi.ttsdemo;
     
    import baidu.restapi.common.ConnUtil;
    import baidu.restapi.common.TokenHolder;
    import java.io.File;
    import java.io.FileOutputStream;
    import java.net.HttpURLConnection;
    import java.net.URL;
     
    public class TtsMain {
     
        public static void main(String[] args) throws Exception {
            (new TtsMain()).run();
        }
     
        //  填写网页上申请的appkey 如 $apiKey="g8eBUMSokVB1BHGmgxxxxxx"
        private final String appKey = "4E1BG9lTnlSeIf1NQFlrSq6h";
     
        // 填写网页上申请的APP SECRET 如 $secretKey="94dc99566550d87f8fa8ece112xxxxx"
        private final String secretKey = "544ca4657ba8002e3dea3ac2f5fdd241";
     
        // text 的内容为"欢迎使用百度语音合成"的urlencode,utf-8 编码
        // 可以百度搜索"urlencode"
        private final String text = "欢迎使用百度语音";
     
        // 发音人选择, 0为普通女声，1为普通男生，3为情感合成-度逍遥，4为情感合成-度丫丫，默认为普通女声
        private final int per = 0;
        // 语速，取值0-9，默认为5中语速
        private final int spd = 5;
        // 音调，取值0-9，默认为5中语调
        private final int pit = 5;
        // 音量，取值0-9，默认为5中音量
        private final int vol = 5;
     
        public final String url = "http://tsn.baidu.com/text2audio"; // 可以使用https
     
        private String cuid = "1234567JAVA";
     
        private void run() throws Exception {
            TokenHolder holder = new TokenHolder(appKey, secretKey, TokenHolder.ASR_SCOPE);
            holder.resfresh();
            String token = holder.getToken();
     
            String url2 = url + "?tex=" + ConnUtil.urlEncode(text);
            url2 += "&per=" + per;
            url2 += "&spd=" + spd;
            url2 += "&pit=" + pit;
            url2 += "&vol=" + vol;
            url2 += "&cuid=" + cuid;
            url2 += "&tok=" + token;
            url2 += "&lan=zh&ctp=1";
            // System.out.println(url2); // 反馈请带上此url，浏览器上可以测试
            HttpURLConnection conn = (HttpURLConnection) new URL(url2).openConnection();
            conn.setConnectTimeout(5000);
            String contentType = conn.getContentType();
            if (contentType.contains("mp3")) {
                byte[] bytes = ConnUtil.getResponseBytes(conn);
                File file = new File("result.mp3"); // 打开mp3文件即可播放
                // System.out.println( file.getAbsolutePath());
                FileOutputStream os = new FileOutputStream(file);
                os.write(bytes);
                os.close();
                System.out.println("mp3 file write to " + file.getAbsolutePath());
            } else {
                System.err.println("ERROR: content-type= " + contentType);
                String res  = ConnUtil.getResponseString(conn);
                System.err.println(res);
            }
        }
    }

##  5、实现效果 ##

![20200618105524947.jpg][]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2d1b3J1aV9qYXZh_size_16_color_FFFFFF_t_70]: /images/20210813/5f70449e0be547e39795c8043c3bce54.png
[20200618105524947.jpg]: /images/20210813/42b87ed350a84021a69b25367bb8fb6f.png