我们在开发中，用到实时语音的时候，会有讯飞、百度、阿里，今天主要讲解的是百度语音之语音合成：
public class YuYinUtil {
?? ?private static final Logger logger = LogManager.getLogger(YuYinUtil.class);
?? ?public static final String YU_YIN_HE_CHENG_URL = "http://tsn.baidu.com/text2audio";
?? ?/**
?? ? * 语音合成
?? ? *?
?? ? * @param token 百度token
?? ? * @param text ?发音文本
?? ? * @param vol ? 音量 0-15
?? ? * @param per ? 发音人 0-普通女声、1-普通男声、3-度逍遥、4度丫丫
?? ? * @param spd ? 语速 0-15
?? ? * @param aue ? 格式 3为mp3格式(默认)； 4为pcm-16k；5为pcm-8k；6为wav（内容同pcm-16k）;
?? ? * ? ? ? ? ? ? ?注意aue=4或者6是语音识别要求的格式，但是音频内容不是语音识别要求的自然人发音，所以识别效果会受影响。
?? ? * @throws UnsupportedEncodingException
?? ? */
?? ?@SuppressWarnings("deprecation")
?? ?public void generateYuYin(String token, String text, String vol, String per, String spd, String pit, String aue,
?? ??? ??? ?String outputFile) {
?? ??? ?logger.debug("调用语音合成接口");
?? ??? ?logger.debug("text:" + text);
?? ??? ?logger.debug("vol:" + vol);
?? ??? ?logger.debug("per:" + per);
?? ??? ?logger.debug("spd:" + spd);
?? ??? ?logger.debug("pit:" + pit);
?? ??? ?logger.debug("aue:" + aue);
?? ??? ?CloseableHttpClient httpClient = HttpClients.createMinimal();
?? ??? ?// 百度语音要求tex字段进行两次url encode，post实体中会进行一次encode
?? ??? ?String encodeText;
?? ??? ?try {
?? ??? ??? ?encodeText = URLEncoder.encode(text, "utf-8");
?? ??? ?} catch (UnsupportedEncodingException e1) {
?? ??? ??? ?throw new RuntimeException(e1);
?? ??? ?}
//?? ??? ?encodeText = URLEncoder.encode(encodeText, "utf-8");
//?? ??? ?String url = YU_YIN_HE_CHENG_URL + "?lan=zh&ctp=1&cuid=abcdxxx&tok=" + token + "&tex=" + encodeText
//?? ??? ??? ??? ?+ "&vol=9&per=0&spd=5&pit=5&aue=3";
//?? ??? ?url = URLEncoder.encode(url, "utf-8");
//?? ??? ?HttpGet get = new HttpGet(url);
?? ??? ?HttpPost post = new HttpPost(YU_YIN_HE_CHENG_URL);
?? ??? ?List<NameValuePair> params = new ArrayList<>();
?? ??? ?// 语言 目前只支持zh固定值
?? ??? ?params.add(new BasicNameValuePair("lan", "zh"));
?? ??? ?// 客户端类型选择 web端写1
?? ??? ?params.add(new BasicNameValuePair("ctp", "1"));
?? ??? ?// cuid 用户唯一标识，用来计算UV值。建议填写能区分用户的机器 MAC 地址或 IMEI 码，长度为60字符以内
?? ??? ?params.add(new BasicNameValuePair("cuid", "kelvinylon"));
?? ??? ?// token
?? ??? ?params.add(new BasicNameValuePair("tok", token));
?? ??? ?// text 语音文本
?? ??? ?params.add(new BasicNameValuePair("tex", encodeText));
?? ??? ?// 以下是选填项
?? ??? ?// vol 音量 0-15
?? ??? ?if (vol != null) {
?? ??? ??? ?params.add(new BasicNameValuePair("vol", vol));
?? ??? ?}
?? ??? ?// per 发音人 0、1、3、4
?? ??? ?if (per != null) {
?? ??? ??? ?params.add(new BasicNameValuePair("per", per));
?? ??? ?}
?? ??? ?// spd 语速 0-15
?? ??? ?if (spd != null) {
?? ??? ??? ?params.add(new BasicNameValuePair("spd", spd));
?? ??? ?}
?? ??? ?
?? ??? ?// pit 语调
?? ??? ?if(pit != null) {
?? ??? ??? ?params.add(new BasicNameValuePair("pit", pit));
?? ??? ?}
?? ??? ?// aue 3为mp3格式(默认)； 4为pcm-16k；5为pcm-8k；6为wav（内容同pcm-16k）;
?? ??? ?// 注意aue=4或者6是语音识别要求的格式，但是音频内容不是语音识别要求的自然人发音，所以识别效果会受影响。
?? ??? ?if (aue != null) {
?? ??? ??? ?params.add(new BasicNameValuePair("aue", aue));
?? ??? ?}
?? ??? ?UrlEncodedFormEntity entity;
?? ??? ?try {
?? ??? ??? ?entity = new UrlEncodedFormEntity(params);
?? ??? ?} catch (UnsupportedEncodingException e1) {
?? ??? ??? ?throw new RuntimeException(e1);
?? ??? ?}
?? ??? ?post.setEntity(entity);
?? ??? ?RequestConfig config = RequestConfig.custom().setConnectionRequestTimeout(5000).setConnectTimeout(5000)
?? ??? ??? ??? ?.setSocketTimeout(20000).build();
//?? ??? ?get.setConfig(config);
?? ??? ?post.setConfig(config);
?? ??? ?try {
?? ??? ??? ?CloseableHttpResponse resp = httpClient.execute(post);
?? ??? ??? ?HttpEntity respEntity = resp.getEntity();
?? ??? ??? ?String contentType = respEntity.getContentType().getValue();
?? ??? ??? ?System.out.println(contentType);
?? ??? ??? ?if (contentType.equals("application/json")) {
?? ??? ??? ??? ?throw new RuntimeException("错误：contentType为json，此请求的contentType应为音频文件");
?? ??? ??? ?}
?? ??? ??? ?InputStream contentIn = respEntity.getContent();
?? ??? ??? ?File outFile = new File(outputFile);
?? ??? ??? ?if(!outFile.getParentFile().exists()) {
?? ??? ??? ??? ?boolean b = outFile.getParentFile().mkdirs();
?? ??? ??? ??? ?if(!b) {
?? ??? ??? ??? ??? ?throw new RuntimeException("创建父文件夹失败");
?? ??? ??? ??? ?}
?? ??? ??? ?}
?? ??? ??? ?
?? ??? ??? ?FileOutputStream fout = new FileOutputStream(outputFile);
?? ??? ??? ?try {
?? ??? ??? ??? ?IOUtils.copy(contentIn, fout);
?? ??? ??? ?} finally {
?? ??? ??? ??? ?IOUtils.closeQuietly(contentIn);
?? ??? ??? ??? ?IOUtils.closeQuietly(fout);
?? ??? ??? ?}
?? ??? ?} catch (IOException e) {
?? ??? ??? ?throw new RuntimeException(e);
?? ??? ?} finally {
?? ??? ??? ?try {
?? ??? ??? ??? ?httpClient.close();
?? ??? ??? ?} catch (IOException e) {
?? ??? ??? ??? ?e.printStackTrace();
?? ??? ??? ?}
?? ??? ?}
?? ?}
?? ?public void generateYuYin(String token, String text, String outputFile) {
?? ??? ?generateYuYin(token, text, null, null, null, null, null, outputFile);
?? ?}
?? ?/**
?? ? * 语音识别成文字
?? ? *?
?? ? * @param in
?? ? * @param token
?? ? * @return
?? ? */
?? ?public String reconizeYuYin(InputStream in, String token) {
?? ??? ?return null;
?? ?}
}
?