Text-to-Speech handler for Lumen Robot Friend with expressive intonation support.

1. Buat folder C:\mbroladb

2. Download MBROLA for PC/DOS, dan extract ke C:\mbrola. Sumber: http://tcts.fpms.ac.be/synthesis/mbrola.html

3. Install eSpeak. Penting: pada waktu install eSpeak tulis mb-id1 pada salah satu voice name yang kosong.

4. Install MbrolaTools35.exe. eSpeak for Windows membutuhkannya untuk dapat memanggil MBROLA.

5. Download id1: Indonesian Male (4Mb) Arry Arman dari http://tcts.fpms.ac.be/synthesis/mbrola/mbrcopybin.html, extract lalu copy folder id1-001010 ke dalam folder C:\mbroladb

6. Copy file id1 (file di dalam folder id1-001010, bukan foldernya itu sendiri) ke dalam C:\Program Files (x86)\eSpeak\espeak-data\mbrola

7. Download ffmpeg for Windows, extract ke D:\ lalu rename foldernya menjadi ffmpeg, di dalam folder ffmpeg sudah harus ada hasil ekstrak(D:\ffmpeg).

8. Tambahkan MBROLA, eSpeak, dan ffmpeg ke environment PATH, contoh: C:\mbrola;C:\Program Files (x86)\eSpeak\command_line;D:\ffmpeg\bin

9. Test: Jalankan Command Prompt (harus dibuka baru, tidak boleh yang sudah jalan), lalu pastikan semua perintah berikut menghasilkan output (bukan "not recognized as command"):

    mbrola
    espeak "Hello world"
    espeak -v mb-id1 "Apa kabar?"
   

   Bila muncul Can't load mbrola.dll. mbrola voice not found, artinya Mbrola Tools belum diinstall.

1. Install MBROLA, voice English, voice bahasa Indonesia, eSpeak, dan libavcodec54, caranya:

    sudo apt-get install mbrola mbrola-us1 mbrola-id1 espeak libavcodec54
   

2. Test: Jalankan Terminal/Konsole, lalu pastikan semua perintah berikut menghasilkan output (bukan "not recognized as command"):

    mbrola
    espeak "Hello world"
    espeak -v mb-id1 "Apa kabar?"
   

1. Pastikan MBROLA, MBROLA voice Indonesia, dan eSpeak sudah terinstall dan berfungsi baik

2. Clone project lumen-sdk:

    sudo apt-get install mbrola mbrola-us1 mbrola-id1
   

3. Clone project speech-synthesis

4. Buka project speech-synthesis di IntelliJ IDEA sebagai Maven project

5. New Module from Existing Sources > buka folder speech-synthesis

6. Klik speech-synthesis, pilih src>main>java, klik kanan di SpeechSynthesisApp, pilih create 'speech synthesis' pada bagian working directory isi $MODULE_DIR$, ceklis "Single Instance Only"

7. Klik speech-synthesis, pilih src>main>java, klik kanan di SpeechSynthesisApp pilih run

Not working yet (eSpeak version mismatch with arabic data version) :(

2015-10-07 19:48:08.987  INFO 6072 --- [abbitMQConsumer] o.l.l.s.e.SpeechExpressionRouter         : Got speech lang-legacy=ar: CommunicateAction{inLanguage=ar-SA, object='السَّلاَمُ عَلَيْكُمْ وَرَحْمَةُ اللهِ وَبَرَكَاتُهُ', actionStatus=null, avatarId='nao1', emotionKind=null}
2015-10-07 19:48:09.163  INFO 6072 --- [abbitMQConsumer] o.l.l.s.e.SpeechExpressionRouter         : [espeak, -s, 130, -v, mb-ar1, -w, C:\Users\ceefour\AppData\Local\Temp\lumen-speech-expression_2395788186719845122.wav, "السَّلاَمُ عَلَيْكُمْ وَرَحْمَةُ اللهِ وَبَرَكَاتُهُ"]: Wrong version of espeak-data 0x14709 (expects 0x14801) at C:\Program Files (x86)\eSpeak\espeak-data