使用FreeTTS和JSAPI

Question

我正在从事一个项目，其中我需要文本到语音转换。我下载了这个FreeTTS 1.2，为了熟悉它的工作，我开始编写一个简单的拼写蜜蜂程序，其中的单词将由应用程序说出，您必须正确拼写它。但我在功能上遇到了问题，比如不管我做了什么改变，声音都是一样的。与最初一样，我找到的代码在Voice()构造函数中有以下参数

Voice kevin = new Voice("kevin", 
    Voice.GENDER_DONT_CARE, Voice.AGE_DONT_CARE, null);

但后来我把它改成

Voice kevin = new Voice("Different1", 
        Voice.GENDER_FEMALE, Voice.AGE_MIDDLE_ADULT, null);

声音仍然是一样的。FreeTTS附带了一个speech.properties文件，为了运行代码，我必须将该文件粘贴到我的主文件夹中。我已经搜索了很多链接来解决这个问题，甚至还检查了Javadoc中的JSAPI和FreeTTS。我唯一的共同点是一个叫MBrola的东西。我不知道这是什么，但当我运行代码时，控制台也会给出一条与它相关的消息。

System property "mbrola.base" is undefined.  Will not use MBROLA voices.

下面是完整的代码：

package game;

    import java.beans.PropertyChangeEvent;
import java.beans.PropertyChangeListener;
import java.io.File;
import java.util.Locale;

import javax.speech.Central;
import javax.speech.synthesis.SpeakableAdapter;
import javax.speech.synthesis.SpeakableEvent;
import javax.speech.synthesis.Synthesizer;
import javax.speech.synthesis.SynthesizerModeDesc;
import javax.speech.synthesis.Voice;

public class MixedVoices {

    private final static void usage() {
    System.out.println("MixedVoices [-showEvents] [-showPropertyChanges]");
    }

    /**
    * Returns a "no synthesizer" message, and asks 
    * the user to check if the "speech.properties" file is
    * at <code>user.home</code> or <code>java.home/lib</code>.
    *
    * @return a no synthesizer message
    */
    static private String noSynthesizerMessage(String synthesizer)
    {
        String message =
            "Cannot find " + synthesizer + ".\n" +
            "This may be the result of any number of problems.  It's\n" +
            "typically due to a missing \"speech.properties\" file that\n" +
            "should be at either of these locations: \n\n";
        message += "user.home    : " + System.getProperty("user.home") + "\n";
        message += "java.home/lib: " + System.getProperty("java.home") +
        File.separator + "lib\n\n" +
            "Another cause of this problem might be corrupt or missing\n" +
            "voice jar files in the freetts lib directory.  This problem\n" +
            "also sometimes arises when the freetts.jar file is corrupt\n" +
            "or missing.  Sorry about that.  Please check for these\n" +
            "various conditions and then try again.\n";
        return message;
    }

    public static void main(String[] argv) 
    {
        boolean showEvents = false;
        boolean showPropertyChanges = false;

        for (int i = 0; i < argv.length; i++) 
        {
            if (argv[i].equals("-showEvents")) {
                showEvents = true;
        } else if (argv[i].equals("-showPropertyChanges")) {
        showPropertyChanges = true;
        } else {
        usage();
        System.exit(0);
        }
    }

    System.out.println(" ** Mixed Voices - JSAPI Demonstration program **");
        /* kevin in an 8khz general domain diphone voice
        */
    Voice kevin = new Voice("kevin", 
        Voice.GENDER_DONT_CARE, Voice.AGE_DONT_CARE, null);

        /* kevin16 in a 16khz general domain diphone voice
        */
    Voice kevinHQ = new Voice("kevin16", 
        Voice.GENDER_DONT_CARE, Voice.AGE_DONT_CARE, null);

    try {
        /* Find a synthesizer that has the general domain voice
            * we are looking for.  NOTE:  this uses the Central class
            * of JSAPI to find a Synthesizer.  The Central class
            * expects to find a speech.properties file in user.home
            * or java.home/lib.
            *
            * If your situation doesn't allow you to set up a
            * speech.properties file, you can circumvent the Central
            * class and do a very non-JSAPI thing by talking to
            * FreeTTSEngineCentral directly.  See the WebStartClock
            * demo for an example of how to do this.
            */
        SynthesizerModeDesc generalDesc = new SynthesizerModeDesc(
        null,          // engine name
                "general",     // mode name
                Locale.US,     // locale
                null,          // running
                null);         // voice

        final Synthesizer synthesizer1 =
                Central.createSynthesizer(generalDesc);

        if (synthesizer1 == null) {
        System.err.println(
                    noSynthesizerMessage("general domain synthesizer"));
        System.exit(1);
        }

        /* Find a synthesizer that has the time domain voice.
        */
        SynthesizerModeDesc limitedDesc = new SynthesizerModeDesc(
                null,          // engine name
                "time",        // mode name
                Locale.US,     // locale
                null,          // running
                null);         // voice

        final Synthesizer synthesizer2 =
                Central.createSynthesizer(limitedDesc);

        if (synthesizer2 == null) {
        System.err.println(
                    noSynthesizerMessage("time domain synthesizer"));
        System.exit(1);
        }

        System.out.print("  Allocating synthesizers...");
        synthesizer1.allocate();
        synthesizer2.allocate();

        /* get general domain synthesizer ready to speak
            */
        System.out.print("Loading voices...");
        synthesizer1.getSynthesizerProperties().setVoice(kevinHQ);
        synthesizer1.getSynthesizerProperties().setVoice(kevin);

        if (showPropertyChanges) {
        synthesizer1.getSynthesizerProperties().addPropertyChangeListener(
                    new PropertyChangeListener() {
                        public void propertyChange(
                            PropertyChangeEvent pce) {
                            if (pce.getNewValue() instanceof Voice) {
                                String newVoice = 
                                    ((Voice) pce.getNewValue()).getName();
                                    System.out.println(
                                        "  PCE Voice changed to " + newVoice);
                            }
                            else {
                                System.out.println(
                                    "  PCE " + pce.getPropertyName()
                                    + " changed from " 
                                    + pce.getOldValue() + " to " +
                                    pce.getNewValue() + ".");
                            }
            }
            });
        }

        if (showEvents) {
        synthesizer1.addSpeakableListener(
            new SpeakableAdapter() {
            public void markerReached(SpeakableEvent e) {
                dumpEvent(e);
            }
            public void speakableCancelled(SpeakableEvent e) {
                dumpEvent(e);
            }
            public void speakableEnded(SpeakableEvent e) {
                dumpEvent(e);
            }
            public void speakablePaused(SpeakableEvent e) {
                dumpEvent(e);
            }
            public void speakableResumed(SpeakableEvent e) {
            dumpEvent(e);
            }
            public void speakableStarted(SpeakableEvent e) {
                dumpEvent(e);
            }
            public void topOfQueue(SpeakableEvent e) {
                dumpEvent(e);
            }
            public void wordStarted(SpeakableEvent e) {
                dumpEvent(e);
            }
            private void dumpEvent(SpeakableEvent e) {
            System.out.println(" EVT: " + e.paramString() 
                                            + " source: " + e.getSource());
        }   
            });
        }

        System.out.println("And here we go!");
        synthesizer1.resume();
        synthesizer2.resume();

        // speak the "Hello world" string
        synthesizer1.speakPlainText("Hello! My name is Kevin.", null);
        synthesizer1.speakPlainText("I am a die phone synthesizer", null);
        synthesizer1.speakPlainText("I have a friend named Alan.", null);
        synthesizer1.speakPlainText("Listen to him count!", null);

        // get synth2 ready to speak
    synthesizer2.waitEngineState(Synthesizer.ALLOCATED);
        synthesizer2.resume();

        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);
        synthesizer2.speakPlainText("1 2 3 4 5 6 7 8 9 ten", null);

        synthesizer2.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.speakPlainText("Now listen to me count!", null);
        synthesizer1.speakPlainText("1 2 3 4 5 6 7 8 9 10.", null);

        synthesizer1.speakPlainText(
            "Now, let's try that a little bit faster.", null);
    synthesi    zer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

    synthesizer1.getSynthesizerProperties().setSpeakingRate(240.0f);
        synthesizer1.speakPlainText("1 2 3 4 5 6 7 8 9 10.", null);
        synthesizer1.speakPlainText("That's pretty fast.", null);
        synthesizer1.speakPlainText("Now lets go very slow.", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.getSynthesizerProperties().setSpeakingRate(80.0f);
        synthesizer1.speakPlainText("1 2 3 4 5 6 7 8 9 10.", null);
        synthesizer1.speakPlainText("That is pretty slow.", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

    synthesizer1.getSynthesizerProperties().setSpeakingRate(150.0f);
        synthesizer1.speakPlainText("Now back to normal", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.getSynthesizerProperties().setPitch(200);
        synthesizer1.speakPlainText("I can talk very high.", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.getSynthesizerProperties().setPitch(50);
        synthesizer1.speakPlainText("and I can talk very low.", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.getSynthesizerProperties().setPitch(100);
    synthesizer1.getSynthesizerProperties().setVolume(.8f);
        synthesizer1.speakPlainText("and I can talk very softly.", null);
    synt    hesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.getSynthesizerProperties().setVolume(1.0f);
        synthesizer1.speakPlainText(
        "I can talk with a higher quality voice", null);
        synthesizer1.speakPlainText(
            "Here is a low quality tongue twister. "
            + "She sells seashells by the seashore.", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

    synthesizer1.getSynthesizerProperties().setVoice(kevinHQ);
        synthesizer1.speakPlainText("And this is high quality. "
            + "She sells seashells by the seashore.", null);
        synthesizer1.speakPlainText(
            "The funny thing is, I do not have a tongue.", null);
        synthesizer1.speakPlainText(
        "Hey Alan, what time is it where you are right now?", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer2.speakPlainText(
            "the time is now twenty past six.", null);
        synthesizer2.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.speakPlainText("Is that the exact time?", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer2.speakPlainText("Almost", null);
    synthesizer2.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.speakPlainText(
        "Is it twenty past six In the morning or the evening?", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer2.speakPlainText("in the morning.", null);
        synthesizer2.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.speakPlainText(
        "Alan and I can talk at the same time", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.speakPlainText("1 2 3 4 5 6 7 8 9 11 12", null);
        synthesizer2.speakPlainText("1 2 3 4 5 6 7 8 9", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);
        synthesizer2.waitEngineState(Synthesizer.QUEUE_EMPTY);

        synthesizer1.speakPlainText( "That is a bit confusing.", null);
        synthesizer1.speakPlainText( "Well, thanks. This was fun.", null);
        synthesizer1.speakPlainText("Goodbye everyone.", null);
        synthesizer1.waitEngineState(Synthesizer.QUEUE_EMPTY);

        // clean up
        synthesizer1.deallocate();
        synthesizer2.deallocate();
    }
    catch (Exception e) {
        e.printStackTrace();
    }

    System.exit(0);
    }
}

Answer 1

您可以访问的声音由VoiceManager类控制。它将通过几种不同的方法搜索安装在您系统中的语音包。默认情况下，你会安装3个声音：“阿兰”，“凯文”和"kevin16“。您需要安装其他语音包(例如CMUArctic或MBROLA)，然后才能使用除3种默认语音之外的语音。

但是，一旦您安装了新的语音(查看相应的网站以获取安装说明)，您还需要将该语音目录添加到正确的位置，并将其名称添加到一个文本文件中，该文本文件列出了所有已安装的语音。请看一下安装语音包下面的Programmer's Guide。作为参考，我最终将我的新语音包路径添加到了我正在使用的freetts.jar旁边的voices.txt文件中(这有点棘手-我忘记了我已经将它添加到了我的Mac上的Library/Java/Extensions文件夹中，所以你可能必须在那里添加voices.txt文件)。

您可以使用以下代码查看哪些语音可用：

VoiceManager voiceManager = VoiceManager.getInstance();
Voice[] voices = voiceManager.getVoices();
for (int i = 0; i < voices.length; i++){
    System.out.print( "voice name is: " + voices[i].getName() +"\n");
}

成功添加已安装语音的Voice目录后，在运行该代码时应该会看到列出的语音目录。