Software Map

Julius is an open-source, high-performance large vocabulary continuous speech recognition (LVCSR) engine for speech-related researchs and developments. With HMM acoustic model and language model, you can construct your own speech recognition system.

Moved to github: https://github.com/julius-speech/julius

Linphone is an audio and video Internet phone with GTK+ and console interfaces. It uses the SIP protocol, and is compatible with most SIP clients and gateways. It can use various audio and video codecs such as Speex, GSM, G711, G722, ilbc, amr, Theora, H263-1998, MPEG4, H264, VP8, and snow.

eGuideDogプロジェクトでは、視覚障害者のためのフリーソフトウェアを開発しています。現在は「WebSpeech」、「Ekho TTS」と「WebAnywhere」にフォーカスしています。

Speex is a patent-free compression format designed especially for speech. It is specialized for voice communications at low bit-rates in the 2-45 kbps range. Possible applications include Voice over IP (VoIP), Internet audio streaming, audio books, and archiving of speech data (e.g. voice mail).

The FAAC project includes the AAC encoder FAAC and decoder FAAD2. It supports several MPEG-4 object types (LC, Main, LTP, HE AAC, PS) and file formats (ADTS AAC, raw AAC, MP4), multichannel and gapless en/decoding as well as MP4 metadata tags. The codecs are compatible with standard-compliant audio applications using one or more of these profiles.

PHP Voice (formerly known as PHP VXML) contain four classes that assist in developing voice application using PHP. It supports Speech Synthesis Markup Language 1.0, Speech Recognition Grammar Specification 1.0, Voice Browser Call Control: CCXML 1.0, and Voice Extensible Markup Language (VoiceXML) 2.0.

CMU Sphinx, a Speech Recognition System, is transitioning to Open Source. The distribution contains a library (libsphinx2) and some small examples that link against it.

Sphinx-4 is a speaker-independent, continuous speech recognition system.

MisterHouse is a Unix/Windows home automation program written in Perl. It can respond to voice commands, Web browsers, time of day, serial port and X10 data, external files, etc., and can speak via Text to Speech engines.

機能豊富な辞書検索プログラムであり、複数の形式の辞書をサポート、完全なマークアップにより完璧な記事レンダリング、イラストおよびその他のコンテンツの保持などの機能をサポートします。

Open JTalk は、修正BSDライセンスの元で配布されている日本語テキスト音声合成システムです。Open JTalk は、オープンソースの形態素解析エンジンの MeCab（和布蕪、めかぶ）、奈良先端大学を中心にして開発された形態素解析用辞書の naist-jdic、隠れマルコフモデル(HMM)に基づく音声合成エンジン hts_engine を用いています。