基于SoundPool的金额语音播报（无SDK，一次解码，可控制语速）

最新推荐文章于 2024-05-15 11:42:20 发布

四处碰壁的小黄

最新推荐文章于 2024-05-15 11:42:20 发布

阅读量307

点赞数

分类专栏：语音播报 android 文章标签： android 算法

本文链接：https://blog.csdn.net/qq_28741899/article/details/106585492

版权

android 同时被 2 个专栏收录

11 篇文章 0 订阅

订阅专栏

语音播报

1 篇文章 0 订阅

订阅专栏

金额语音播报，做过支付行业的人都接触过这关功能，实现方法有很多种

比较常见的如 TTS(系统自带声音)，语音引擎SDK(一些收费，这里不讨论)，MP3（资源文件顺序播报）

TTS

最简单的实现方法吧，不过声音的改变比较困难，需要下载声音引擎

参考上一篇博客《金钱转中文字符串》得到中文字符串

然后使用tts进行播报

 textToSpeech = TextToSpeech(this,TextToSpeech.OnInitListener {
            if (it == TextToSpeech.SUCCESS){
                textToSpeech.language = Locale.CHINESE
                textToSpeech .speak("收款"+MoneyUtil.readString("30.57"),TextToSpeech.QUEUE_FLUSH,null,"1")
            }
        })

一个简单的语音播报就这样实现了

基于MP3资源实现

简书 Android 语音播报实现方案（无SDK），写得非常棒，不过这里使用的是Mediaplayer播放器，这里做了一些改进

由于是基于Mediaplayer，基本是监听播放完成和解码完成的监听器，然后播放完成后立刻解码下一个音频文件。

这种每次需要播放时才去解码语音文件进行播放，文字和文字加的语速间隔是不可控的。比如一些低端机解码慢

资源准备

https://ai.baidu.com/tech/speech/tts_online

百度AI语音合成，也可以自己录~

准备 0~9 的语音点，千，百，十，亿，元，***收款成功，等语音

SoundPool

android专门用来播放短时长提示音，缺点是短~，无法知道啥时候播放完成。优点的轻量级，一次load，后面即可实时播放

基于资源的语音播报，一般一个音频大概是0.5s-1s，应该没人会把1 拖到一秒种那么长吧，SoundPool刚好符合需求，短

1、SoundPool的初始化

 /**
     * SoundPool ID ArrayMap
     * 数据量小使用ArrayMap比HashMap性能更好
     * */
    private var soundId = ArrayMap<String,Int>()

/**
     * 初始化，第一次使用时自动初始化，无需手动调用
     * @param context
     * @param moneyString 播报的金额中文串（通过MoneyUtil获得）
     * @param autoPlay 初始化后是否自动播放
     * */
    private fun init(context: Context,moneyString:String ="",autoPlay:Boolean=false){

        var audioAttributes =  AudioAttributes.Builder()
            .setUsage(AudioAttributes.USAGE_MEDIA)
            .setContentType(AudioAttributes.CONTENT_TYPE_MUSIC)
            .build();
        var assets = context.assets
        mSoundPool = SoundPool.Builder().setAudioAttributes(audioAttributes).setMaxStreams(22).build()

        mSoundPool.setOnLoadCompleteListener { _, sampleId, status ->
            if (status == 0 &&sampleId == soundId.size){ //所有资源初始化成功
                isPrepare = true //解码成功
                if (autoPlay){ //是否自动播放
                    playVoicePool(moneyString,context)
                }
            }
        }
        soundId.clear()
        soundId["sound/tts_1.mp3"] = mSoundPool.load(assets.openFd("sound/tts_1.mp3"),1)
        soundId["sound/tts_2.mp3"] = mSoundPool.load(assets.openFd("sound/tts_2.mp3"),1)
        soundId["sound/tts_3.mp3"] = mSoundPool.load(assets.openFd("sound/tts_3.mp3"),1)
        soundId["sound/tts_4.mp3"] = mSoundPool.load(assets.openFd("sound/tts_4.mp3"),1)
        soundId["sound/tts_5.mp3"] = mSoundPool.load(assets.openFd("sound/tts_5.mp3"),1)
        soundId["sound/tts_6.mp3"] = mSoundPool.load(assets.openFd("sound/tts_6.mp3"),1)
        soundId["sound/tts_7.mp3"] = mSoundPool.load(assets.openFd("sound/tts_7.mp3"),1)
        soundId["sound/tts_8.mp3"] = mSoundPool.load(assets.openFd("sound/tts_8.mp3"),1)
        soundId["sound/tts_9.mp3"] = mSoundPool.load(assets.openFd("sound/tts_9.mp3"),1)
        soundId["sound/tts_0.mp3"] = mSoundPool.load(assets.openFd("sound/tts_0.mp3"),1)
        soundId["sound/tts_dot.mp3"] = mSoundPool.load(assets.openFd("sound/tts_dot.mp3"),1)
        soundId["sound/tts_hundred.mp3"] = mSoundPool.load(assets.openFd("sound/tts_hundred.mp3"),1)
        soundId["sound/tts_receipt.mp3"] = mSoundPool.load(assets.openFd("sound/tts_receipt.mp3"),1)
        soundId["sound/tts_success.mp3"] = mSoundPool.load(assets.openFd("sound/tts_success.mp3"),1)
        soundId["sound/tts_ten_million.mp3"] = mSoundPool.load(assets.openFd("sound/tts_ten_million.mp3"),1)
        soundId["sound/tts_ten_thousand.mp3"] = mSoundPool.load(assets.openFd("sound/tts_ten_thousand.mp3"),1)
        soundId["sound/tts_ten.mp3"] = mSoundPool.load(assets.openFd("sound/tts_ten.mp3"),1)
        soundId["sound/tts_thousand.mp3"] = mSoundPool.load(assets.openFd("sound/tts_thousand.mp3"),1)
        soundId["sound/tts_yuan.mp3"] = mSoundPool.load(assets.openFd("sound/tts_yuan.mp3"),1)
    }

这里可以第一次播放时自动初始化，也可以程序开始直接初始化，都可

2、进行语音播报

fun playVoicePool(moneyString: String,context: Context) {
        thread { //异步线程
            synchronized(this){ //同步锁
                if (isPrepare){ //判断SoudPool是否加载完成
                    var soundList = getListSound(moneyString) //通过monneyString获取播放的音频
                    for (i in soundList){
                        if (soundId[i]!=null){ 
                            mSoundPool.play(soundId[i]!!,1f,1f,1,0,1f) //最后一个参数是语速
                            if (i.contains("tts_success")){ //tts_succes为***收款,间隔长一点
                                Thread.sleep(1100)
                            }else{
                                Thread.sleep(400) // 1,2 十，百，千，亿等等
                            }
                        }
                    }
                }else{
                    init(context,moneyString,true) //还未初始化则初始化
                }

            }
        }
    }

3、根据MoneyString获取对应的语音播放链表

/**
     * 根据@link MoneyUtil生成的中文串来获取链表
     * @param moneyString 播报的金额中文串（通过MoneyUtil获得）
     * @return
     * */
    private fun getListSound(moneyString: String):LinkedList<String>{
        var soundList = LinkedList<String>()
        var moneyChar = moneyString.toCharArray()

        soundList.add("sound/tts_success.mp3")

        for (i in moneyChar){
            if (NUM.contains(i)){
                soundList.add(getStringFormNUM(i))
            }
            if (CHIN_NUM.contains(i)){
                soundList.add(getStringFormCHINNUM(i))
            }
        }
        return soundList
    }

    /**
     * 获取中文字符资源
     * @param char
     * */
    private fun getStringFormCHINNUM(char: Char):String{
        var chineseNumPath =""
        when(char){
            '元' -> chineseNumPath = "sound/tts_yuan.mp3"
            '拾' -> chineseNumPath = "sound/tts_ten.mp3"
            '佰' -> chineseNumPath = "sound/tts_hundred.mp3"
            '仟' -> chineseNumPath = "sound/tts_thousand.mp3"
            '万' -> chineseNumPath = "sound/tts_ten_thousand.mp3"
            '亿' -> chineseNumPath = "sound/tts_ten_million.mp3"
            '点' -> chineseNumPath = "sound/tts_dot.mp3"
        }
        return chineseNumPath
    }

    /**
     * 获取数字资源
     * @param char
     * */
    private fun getStringFormNUM(char:Char):String{
        return String.format("sound/tts_%s.mp3",char)
    }

/**
     * 释放资源
     * */
    fun destroyPool(){
        if (mSoundPool!=null){
            mSoundPool.autoPause()
            mSoundPool.release()
            mSoundPool == null
        }
    }

使用单例子模式封装代码（完整代码）

/**
 * @author HuangJiaHeng
 * @date 2020/6/5.
 * 金额语音播报类

 * playVoicePool SoundPool 可以有效防止卡顿
 */
class VoiceSpeak {

    /**
     * SoundPool 使用SoundPool方式才初始化
     * */
    private lateinit var mSoundPool:SoundPool
    /**
     * SoundPool ID ArrayMap
     * 数据量小使用ArrayMap比HashMap性能更好
     * */
    private var soundId = ArrayMap<String,Int>()
    /**
     * SoundPool是否加载完成
     * */
    private var isPrepare = false

    companion object{

        private val NUM = arrayListOf('0', '1', '2', '3', '4', '5', '6', '7', '8', '9')
        private val CHIN_NUM = arrayListOf('元', '拾', '佰', '仟', '万', '亿','点')

        val util by lazy(LazyThreadSafetyMode.SYNCHRONIZED){
            VoiceSpeak()
        }
    }

    /**
     * 释放资源
     * */
    fun destroyPool(){
        if (mSoundPool!=null){
            mSoundPool.autoPause()
            mSoundPool.release()
            mSoundPool == null
        }
    }

    /**
     * 以池的方式播放，可以有效减少资源解码卡顿情况 (同步锁)
     * @param moneyString 播报的金额中文串（通过MoneyUtil获得）
     * @param context
     * */
    fun playVoicePool(moneyString: String,context: Context) {
        thread {
            synchronized(this){
                if (isPrepare){
                    var soundList = getListSound(moneyString)
                    for (i in soundList){
                        if (soundId[i]!=null){
                            mSoundPool.play(soundId[i]!!,1f,1f,1,0,1f)
                            if (i.contains("tts_success")){
                                Thread.sleep(1100)
                            }else{
                                Thread.sleep(400)
                            }
                        }
                    }
                }else{
                    init(context,moneyString,true)
                }

            }
        }
    }


    /**
     * 根据@link MoneyUtil生成的中文串来获取链表
     * @param moneyString 播报的金额中文串（通过MoneyUtil获得）
     * @return
     * */
    private fun getListSound(moneyString: String):LinkedList<String>{
        var soundList = LinkedList<String>()
        var moneyChar = moneyString.toCharArray()

        soundList.add("sound/tts_success.mp3")

        for (i in moneyChar){
            if (NUM.contains(i)){
                soundList.add(getStringFormNUM(i))
            }
            if (CHIN_NUM.contains(i)){
                soundList.add(getStringFormCHINNUM(i))
            }
        }
        return soundList
    }

    /**
     * 获取中文字符资源
     * @param char
     * */
    private fun getStringFormCHINNUM(char: Char):String{
        var chineseNumPath =""
        when(char){
            '元' -> chineseNumPath = "sound/tts_yuan.mp3"
            '拾' -> chineseNumPath = "sound/tts_ten.mp3"
            '佰' -> chineseNumPath = "sound/tts_hundred.mp3"
            '仟' -> chineseNumPath = "sound/tts_thousand.mp3"
            '万' -> chineseNumPath = "sound/tts_ten_thousand.mp3"
            '亿' -> chineseNumPath = "sound/tts_ten_million.mp3"
            '点' -> chineseNumPath = "sound/tts_dot.mp3"
        }
        return chineseNumPath
    }

    /**
     * 获取数字资源
     * @param char
     * */
    private fun getStringFormNUM(char:Char):String{
        return String.format("sound/tts_%s.mp3",char)
    }

    /**
     * 初始化，第一次使用时自动初始化，无需手动调用
     * @param context
     * @param moneyString 播报的金额中文串（通过MoneyUtil获得）
     * @param autoPlay 初始化后是否自动播放
     * */
    private fun init(context: Context,moneyString:String ="",autoPlay:Boolean=false){

        var audioAttributes =  AudioAttributes.Builder()
            .setUsage(AudioAttributes.USAGE_MEDIA)
            .setContentType(AudioAttributes.CONTENT_TYPE_MUSIC)
            .build();
        var assets = context.assets
        mSoundPool = SoundPool.Builder().setAudioAttributes(audioAttributes).setMaxStreams(22).build()

        mSoundPool.setOnLoadCompleteListener { _, sampleId, status ->
            if (status == 0 &&sampleId == soundId.size){
                isPrepare = true
                if (autoPlay){
                    playVoicePool(moneyString,context)
                }
            }
        }
        soundId.clear()
        soundId["sound/tts_1.mp3"] = mSoundPool.load(assets.openFd("sound/tts_1.mp3"),1)
        soundId["sound/tts_2.mp3"] = mSoundPool.load(assets.openFd("sound/tts_2.mp3"),1)
        soundId["sound/tts_3.mp3"] = mSoundPool.load(assets.openFd("sound/tts_3.mp3"),1)
        soundId["sound/tts_4.mp3"] = mSoundPool.load(assets.openFd("sound/tts_4.mp3"),1)
        soundId["sound/tts_5.mp3"] = mSoundPool.load(assets.openFd("sound/tts_5.mp3"),1)
        soundId["sound/tts_6.mp3"] = mSoundPool.load(assets.openFd("sound/tts_6.mp3"),1)
        soundId["sound/tts_7.mp3"] = mSoundPool.load(assets.openFd("sound/tts_7.mp3"),1)
        soundId["sound/tts_8.mp3"] = mSoundPool.load(assets.openFd("sound/tts_8.mp3"),1)
        soundId["sound/tts_9.mp3"] = mSoundPool.load(assets.openFd("sound/tts_9.mp3"),1)
        soundId["sound/tts_0.mp3"] = mSoundPool.load(assets.openFd("sound/tts_0.mp3"),1)
        soundId["sound/tts_dot.mp3"] = mSoundPool.load(assets.openFd("sound/tts_dot.mp3"),1)
        soundId["sound/tts_hundred.mp3"] = mSoundPool.load(assets.openFd("sound/tts_hundred.mp3"),1)
        soundId["sound/tts_success.mp3"] = mSoundPool.load(assets.openFd("sound/tts_success.mp3"),1)
        soundId["sound/tts_ten_million.mp3"] = mSoundPool.load(assets.openFd("sound/tts_ten_million.mp3"),1)
        soundId["sound/tts_ten_thousand.mp3"] = mSoundPool.load(assets.openFd("sound/tts_ten_thousand.mp3"),1)
        soundId["sound/tts_ten.mp3"] = mSoundPool.load(assets.openFd("sound/tts_ten.mp3"),1)
        soundId["sound/tts_thousand.mp3"] = mSoundPool.load(assets.openFd("sound/tts_thousand.mp3"),1)
        soundId["sound/tts_yuan.mp3"] = mSoundPool.load(assets.openFd("sound/tts_yuan.mp3"),1)
    }
}

使用：

        VoiceSpeak.util.playVoicePool(MoneyUtil.readString("30.57"),this)
        VoiceSpeak.util.playVoicePool(MoneyUtil.readString("108"),this)
        VoiceSpeak.util.playVoicePool(MoneyUtil.readString("556.78"),this)
        VoiceSpeak.util.playVoicePool(MoneyUtil.readString("104850000.56"),this)

githubDemo ：这么简单就不要demo了吧【滑稽】

资源放在 assets/sound目录，文件名和路径要和代码里面一致

四处碰壁的小黄

关注

0
点赞
踩
2

收藏

觉得还不错? 一键收藏
0
评论
基于SoundPool的金额语音播报（无SDK，一次解码，可控制语速）

金额语音播报，做过支付行业的人都接触过这关功能，实现方法有很多种比较常见的如 TTS(系统自带声音)，语音引擎SDK(一些收费，这里不讨论)，MP3（资源文件顺序播报）TTS最简单的实现方法吧，不过声音的改变比较困难，需要下载声音引擎参考上一篇博客《金钱转中文字符串》得到中文字符串然后使用tts进行播报 textToSpeech = TextToSpeech(this,TextToSpeech.OnInitListener { if (it == TextT
复制链接

扫一扫