Speech corpus

From A Cat's Wiki

A speech corpus is a database of speech audio files (e.g. WAV or FLAC), and the corresponding text transcriptions of these speech files. Both components can be used to create acoustic models. Those acoustic models can then be used in conjunction with a speech recognition machine. As speech corpus may contain book excerpts, lists of words, sequences of numbers, dialogues between two people, broadcast news, spontaneous speech, read speech, speech with a foreign accent, native speech.

[edit] External Links

• Explanation at VoxForge.org
• Speech corpus at VoxForge - transcribed Wav/FLAC speech files in 48 kHz/16 bit quality
• The Production of Speech Corpora