Revision as of 09:04, 21 September 2013

<<< back to Wiki Home

Introduction

Talking To Machines

A short workshop introducing speech recognition and speech synthesis techniques for the creation of interactive artwork. We use pre-compiled open-source tools (CMU Sphinx ASR, Festival TTS, Processing, Python) and focus on the demonstrable strengths and unexpected limitations of speech technologies as vehicles for creating meaning.

Saturday Sept 21, 2-6pm Centro Cultural de Borges UNTREF.

Background Reading:

• Natalie Jeremijenko. "If Things Can Talk, What Do They Say? If We Can Talk To Things, What Do We Say?" 2005-03-05 [http://www.electronicbookreview.com/thread/firstperson/voicechip
  • also see the responses by Simon Penny, Lucy Suchmann, and Natalie linked from that page.
  • "Dialogue With A Monologue: Voice Chips and the Products of Abstract Speech". http://www.topologicalmedialab.net/xinwei/classes/readings/Jeremijenko/VoiceChips.pdf
• Mel Bochner. "Serial Art, Systems, Solipsism." (pdf)

Automatic Speech Recognition

• Talking to Machines.

Engines

• CMU Sphinx Open Source Toolkit For Speech Recognition Project by Carnegie Mellon University
• Pocketsphinx. A light-weight, portable implementatin of sphinx. pocketsphinx on win32 - http://www.aiaioo.com/cms/index.php?id=28
• Google ASR.
• Google ASR wrapped for processing - http://stt.getflourish.com/

Hands-on with Processing

• Requires STT library. http://stt.getflourish.com/
• Download the library file, unzip it, and copy it to the Processing\libraries folder. I also put it on the thumbdrive.
• Processing example: File:google_listen.zip
• Try switching the recognition language.

Hands-on with Sphinx

Installation

• Download from sourceforge: http://cmusphinx.sourceforge.net/wiki/download/
• If using windows, you need the sphinxbase-0.8-win32.zip and pocketsphinx-0.8-win32.zip files. I already downloaded these for you. They are in the untref_speech folder.

Usage

• open a terminal. Windows, Run->Cmd.
• change to the pocketsphinx directory.
  • cd Desktop\untref_speech\pocketsphinx-0.8-win32\bin\Release
• ENGLISH: run the pocketsphinx command to recognize english:
  • pocketsphinx_continuous.exe -hmm ..\..\model\hmm\en_US\hub4wsj_sc_8k -dict ..\..\model\lm\en_US\cmu07a.dic -lm ..\..\model\lm\en_US\hub4.5000.DMP
• SPANISH: recognize spanish:
  • pocketsphinx_continuous.exe -hmm ..\..\model\hmm\es_MX\hub4_spanish_itesm.cd_cont_2500 -dict ..\..\model\lm\es_MX\h4.dict -lm ..\..\model\lm\es_MX\H4.arpa.Z.DMP
  • this should transcribe live from the microphone.

Language Models

• Acoustic models versus language models.
• Grammars versus Satistical Language Models.
• Available language models for Sphinx:
  • English
  • Mandarin
  • French
  • Spanish
  • German
  • Dutch
  • and more: http://sourceforge.net/projects/cmusphinx/files/Acoustic%20and%20Language%20Models/

Training your own Models

• grammer is trivial.
• slm, can use online tools. or try the sphinxtrain packages.
• the online tool http://www.speech.cs.cmu.edu/tools/lmtool-new.html
  • upload a plain-text file of sentences. it will produce a language model from these!
  • download the results.
  • I can talk you through using the resultant model.

Programming with Speech Recognition

Processing. Sphinx4, the java interface.

Python or c++, command line, android. pocketsphinx.

Text To Speech Synthesis

Engines

• Festival/Festvox. Festival from University of Edinburgh. CMU Speech group.
• freetts. wrapper for processing - http://www.local-guru.net/blog/pages/ttslib
• MARY TTS. http://mary.dfki.de/
• Google TTS. http://amnonp5.wordpress.com/2011/11/26/text-to-speech/
• Mac OS X Built in speech synthesis
• MBROLA voices. - http://tcts.fpms.ac.be/synthesis/
• Siri

Test them online

• Festival online demo - http://www.cstr.ed.ac.uk/projects/festival/onlinedemo.html
  • Spanish (UVIGO Spanish Male)
  • American English
  • Others...
• MARY TTS online demo - http://mary.dfki.de:59125/

Hands-on With Processing

• For Google TTS no library is required. You don't have to install anything. You just need an internet connection to talk to google.

Example 1. Speech

File:google_speak.zip

Example 2. Daisy Bell

• Daisy Bell - http://www.youtube.com/watch?v=41U78QP8nBk
  • "Daisy Bell" was composed by Harry Dacre in 1892. In 1961, the IBM 7094 became the first computer to sing, singing the song Daisy Bell. Vocals were programmed by John Kelly and Carol Lockbaum and the accompaniment was programmed by Max Mathews.

• Processing Daisy Bell example using Google Text To Speech. Requires an internet connection:
  • File:google_daisy.zip

Hands-on with Festival

Installation

• http://festvox.org/packed/festival/2.1/festival-2.1-release.tar.gz
• Tutorial - http://homepages.inf.ed.ac.uk/jyamagis/misc/Practice_of_Festival_speech_synthesizer.html
• windows binaries http://sourceforge.net/projects/e-guidedog/files/related%20third%20party%20software/0.3/festival-2.1-win.7z/download
• voices http://homepages.inf.ed.ac.uk/jyamagis/software/page54/page54.html
• Copy festival folder to C:\

Usage

• run the terminal. Start Menu, Run -> Cmd.
• cd C:\festival
• start festival:
  • festival
• to say something:
  • (SayText "this is what I am going to say")
• to render speech to sound file:
• to switch voices:
• to exit festival:
  • (exit)
• Festival is written in Scheme, a variant of LISP.

Voices

• http://festvox.org/dbs/index.html
• https://github.com/joseguerrero/festival-spanish-voices
• spanish voices - http://sangonz.wordpress.com/2010/05/22/spanish-voices-for-festival/

Making a Voice

• Portraiture
• Robert Voice

Activity: Feedback Loop

Construct a conversation with the machine.

Processing Sketch

File:Listen speak.zip