tts-edge-java

java sdk for Edge Read Aloud

click me have a try

Setup

maven

<dependency>
  <groupId>io.github.whitemagic2014</groupId>
  <artifactId>tts-edge-java</artifactId>
  <version>version</version>
</dependency>

gradle

implementation group: 'io.github.whitemagic2014', name: 'tts-edge-java', version: 'version'

short
implementation 'io.github.whitemagic2014:tts-edge-java:version'

demo

  @Test
    void should_convert_to_mp3_file_success_with_single_content() {
        String voiceName = "zh-CN-XiaoyiNeural";
        Optional<Voice> voiceOptional = TTSVoice.provides()
                .stream()
                .filter(v -> voiceName.equals(v.getShortName()))
                .findFirst();
        if (!voiceOptional.isPresent()) {
            throw new IllegalStateException("voice not found：" + voiceName);
        }
        Voice voice = voiceOptional.get();
        String filename = voiceName + "-" + "test-tts";
        String content = "你好，有什么可以帮助你的吗，今天的天气很不错呢";
        TTS tts = new TTS(voice, content)
                .findHeadHook()
                .isRateLimited(true) // Set to true to resolve the rate limiting issue in certain regions.
                .fileName(filename)// You can customize the file name; if omitted, a random file name will be generated.
                .overwrite(true) // When the specified file name is the same, it will either overwrite or append to the file.
                .formatMp3();  // default mp3.
//                .formatOpus() // or opus
//                .voicePitch()
//                .voiceRate()
//                .voiceVolume()
//                .storage()  // the output file storage ,default is ./storage
//                .connectTimeout(0) // set connect timeout
        tts.trans();
        // you can find the voice file in storage folder
    }
    
    void should_convert_to_mp3_file_success_with_multi_content() throws IOException {
        String voiceName = "zh-CN-XiaoyiNeural";
        Optional<Voice> voiceOptional = TTSVoice.provides()
                .stream()
                .filter(v -> voiceName.equals(v.getShortName()))
                .findFirst();
        if (!voiceOptional.isPresent()) {
            throw new IllegalStateException("voice not found：" + voiceName);
        }
        Voice voice = voiceOptional.get();
        List<TransRecord> recordList = new ArrayList<>();
        String store = "./storage";

        // create batch task
        for (int i = 0; i < 100; i++) {
            TransRecord record = new TransRecord();
            record.setContent(i + ", hello tts, 你好，有什么可以帮助你的吗");
            record.setFilename(i + ".test-tts");
            recordList.add(record);
            Files.deleteIfExists(Paths.get(buildFilename(store, record)));
        }
        new TTS(voice)
                .findHeadHook()
                .isRateLimited(true)
                .overwrite(true)
                .batch(recordList) // set batch task
                .parallel(12) // set up 12 parallel threads
                .storage(store)
                .formatMp3()
                .batchTrans(); // trans
        for (TransRecord record : recordList) {
            Path path = Paths.get(buildFilename(store, record));
            Assertions.assertTrue(Files.exists(path), "file not found in " + path.toString());
        }
    }

    private static String buildFilename(String store, TransRecord record) {
        return store + File.separator + record.getFilename() + ".mp3";
    }

Version

1.3.1

• Optimize: Optimize parallel tasks PR#15

1.3.0

• New: Now supports batch conversion of multiple text tasks within a single TTS transaction.
• Optimize: A new parameter, enableVttFile, has been added. This parameter determines whether VTT file support is enabled, with the default setting being false
• Optimize: Refactored some of the code.
• Update: Upgraded some dependency packages.
• Thanks to user lifeopsgo for making nearly all the contributions to this version update.PR#13

1.2.6

• Resolve Conversion Issues Caused by Special Characters.PR#9

1.2.5

• Optimize: Setting the parameter isRateLimited to true can resolve rate limiting issues in certain regions.
• Thanks to user PoliceZ, the answer in the issue was helpful to me.

1.2.4

• Optimize: The default value for the overWrite parameter is now true.

1.2.3

• Optimize: A new parameter, overWrite, has been added. When the same file name is provided, if overWrite = true, it will overwrite the original audio file and VTT subtitle file. If overWrite = false, it will continue to append to the original audio file and VTT subtitle file.

1.2.2

• Optimize: TTS can now set connectTimeout.

1.2.1

• New: In this update, 17 new supported voices have been added, as follows:

    en-US-AvaMultilingualNeural
    en-US-AndrewMultilingualNeural
    en-US-EmmaMultilingualNeural
    en-US-BrianMultilingualNeural
    en-US-AvaNeural
    en-US-AndrewNeural
    en-US-EmmaNeural
    en-US-BrianNeural
    fr-CA-ThierryNeural
    fr-FR-VivienneMultilingualNeural
    fr-FR-RemyMultilingualNeural
    de-DE-SeraphinaMultilingualNeural
    de-DE-FlorianMultilingualNeural
    it-IT-GiuseppeNeural
    ko-KR-HyunsuNeural
    pt-BR-ThalitaNeural
    es-ES-XimenaNeural

1.2.0

• Optimize: Now You can customize the file name; if omitted, a random file name will be generated.
• New: Now, while generating audio, a VTT subtitle file with the same name will be created.issue:3

1.1.1

• Optimize: Fix high CPU usage while waiting for a response. PR#2

1.1.0

• Optimize: Now, the TTS.trans function offers the choice of receiving the output file in two formats, MP3 or opus.
• Optimize: Add two methods to parse voice file

1.0.1

• Optimize: Now, the TTS.trans function will return the name of the voice file.

1.0.0

• Edge Read Aloud Text To Speech

License

This project is an open-sourced software licensed under the MIT license.