From Openitware
Jump to: navigation, search

System Description

Software for Speech System

  • The following chart shows the speech system tools that are used throughout this project. The chart displays the current version that is running on caesar and the location online to find the current version. The newest version column displays what the newest version available of the the speech tool is and where to find it online. The next column displays whether or not the newest version works with openSUSE11.3 (if the information is available).
  • As of the beginning of Spring 2015 OpenSUSE11.3 is the operating system that is used on caesar. The last column lists the difference between the current version and the newest version. Resources where this information can be found are also listed under the table.
  • Redhat Enterprise 6.6 replaced OpenSuse at the end of the 2015 Spring semester

Tool Purpose Current Version Newest Version Does the newest version work with Redhat Enterprise 6.8? Differences between current and newest versions Should the Software be Upgraded?
Sphinx Decoder Sphinx is a large vocabulary, speaker independent speech recognition codebase and suite of tools. Sphinx3 can use fully continuous observation densities. Sphinx3.7 Sphinx Version 5 Date: 9/24/16 Yes
  • Sphinx 3 is still considered the most accurate for large vocabulary tasks according to CMU.
  • Sphinx 3 is intended more for researchers.
  • Word rate error has decreased with Sphinx 4.
  • The real time (ratio of processing time to audio time) has decreased.
  • Sphinx 4 was written in Java, where Sphinx 3 was written in C.
  • Sphinx 3 generally run in a command line, where Sphinx 4 generally runs in a Java program.
No major reasons to upgrade now.
CMU Language Model Toolkit The Carnegie Mellon Statistical Language Modeling Toolkit is a set of unix software tools designed to facilitate language modeling work in the research community. CMU-Cambridge Statistical Language Modeling Tookit v2 CMUclmtk 0.7 Yes
  • Various bug fixes
  • Documentation fixes
  • No major differences between the versions.
No major reasons to upgrade now.
Sphinx Trainer Part of the CMU toolkit used to run experiments based on the user's specifications. SphinxTrain 1.0 SphinxTrain 1.0.8 Yes
  • SphinxTrain 1.0.8 can access all training processes by single 'sphinxtrain' command
  • SphinxTrain 1.0.8 fixes memory leaks and build issues
  • SphinxTrain 1.0.8 is meant for use with Sphinx4
No major reasons to upgrade now.
CMU Dictionary The Carnegie Mellon University Pronouncing Dictionary is an open-source machine-readable pronunciation dictionary for North American English that contains over 134,000 words and their pronunciations. CMU Dictionary 0.6 CMU Dictionary 0.7b Date: 11/19/14 Yes
  • New dictionary file format
  • More words
It would be beneficial to have considerably more words within the dictionary but it may be best to stick with version 0.6 so we can perfect that first.
Sclite The program sclite is a tool for scoring and evaluating the output of speech recognition systems. SCLITE 2.3, installed with SCTK V1.3 SCLITE 2.4.10 (included in Speech Recognition Scoring Toolkit (SCTK) Version 2.10) Date: 11/11/2018 Yes New version has mainly bug fixes in the software, nothing major. No major reasons to upgrade now.
Emacs An extensible, customizable, free/libre text editor GNU Emacs-nox 23.1 GNU Emacs-nox 26.1 No Unknown No major reasons to upgrade now.
Screen Screen is a full-screen window manager that multiplexes a physical terminal between several processes, typically interactive shells. screen-3.9.15 i386(32bit version) screen-4.6.2 Date: 10/22/17 No Unknown No major reasons to upgrade now.
SOX SoX is a command line utility that can convert various formats of computer audio files in to other formats. Version 14.3.1 [1] Version 14.4.2 Date: 2/22/2015 Yes Adds read support for Ogg Opus files and RF64 WAV files. Deemph can now be used at 48kHz sample rates. Rate is now much faster in many cases. Now supports multi-channel LADSPA plugins and optional latency compensation. Includes many other bug fixes. No major reasons to upgrade now.
Tree Tree is a recursive directory listing command that produces a depth indented listing of files. Tree v 1.7.0 v 1.8.0 Date: 11/16/2018 Yes Mainly fixed issues and bugs with the program. No big improvements or new features. No major reasons to upgrade now.