|AUTHORS||H A D||31-Dec-2008||4.8 KiB||138||95|
|COPYING||H A D||31-Dec-2008||1.7 KiB||36||35|
|ChangeLog||H A D||31-Dec-2008||840 ||37||22|
|INSTALL||H A D||31-Dec-2008||9.3 KiB||237||179|
|Makefile.am||H A D||31-Dec-2008||3.9 KiB||110||73|
|Makefile.in||H A D||16-Aug-2019||24.3 KiB||747||642|
|NEWS||H A D||31-Dec-2008||593 ||19||11|
|README||H A D||31-Dec-2008||5.9 KiB||168||122|
|aclocal.m4||H A D||01-Jan-2009||263.6 KiB||7,526||6,766|
|autogen.sh||H A D||31-Dec-2008||4.5 KiB||153||131|
|config.guess||H A D||23-Oct-2007||43.5 KiB||1,517||1,305|
|config.sub||H A D||23-Oct-2007||32 KiB||1,627||1,483|
|configure||H A D||01-Jan-2009||749.6 KiB||23,896||19,176|
|configure.in||H A D||01-Jan-2009||5.7 KiB||214||186|
|depcomp||H A D||31-Dec-2008||11.8 KiB||412||275|
|install-sh||H A D||31-Dec-2008||9 KiB||324||189|
|ltmain.sh||H A D||08-Feb-2008||195.3 KiB||6,965||5,506|
|missing||H A D||31-Dec-2008||10.8 KiB||361||268|
|sphinx3.pc.in||H A D||31-Dec-2008||365 ||15||13|
1This is Sphinx 3.7 (s3.7), one of Carnegie Mellon University's open
2source large vocabulary, speaker-independent continuous speech
5There has always been confusion of the versioning of Sphinx-3. Here
6is a brief description.
8Over the course of development, Sphinx-3 had branched into two major
9versions. They are the flat lexicon decoder, (or aka "s3-slow") and
10the tree lexicon decoder (or aka "s3-fast"). Performance wise,
11s3-slow is ~10% relative better than s3-fast empirically. However,
12s3-fast could be around 10 times faster than s3-slow. From sphinx 3.0
13to Sphinx 3.4. The two code-bases were separated.
15During the development of Sphinx 3.5, conscious decision were made to
16merge the two branches together. This merging is stabilized in Sphinx
173.6. Internally, both the fast decoder and slow decoder share over
1870% of the codebase. From the user perspective though, the old
19sphinx3_decode_anytopo (s3-slow) and sphinx3_decode (s3-fast) exist as
20two separate interface of the decoding engine. This will allow
21backward compatibility to the past.
23In 3.7, we did a site-wide change that allow sphinx3 and pocketsphinx
24(and perhaps later SphinxTrain) to share the same codebase.
26THIS IS A RESEARCH SYSTEM. After several extensions and upgrade, we
27 still find that Sphinx is better regarded as an early release of a
28research system. We know the APIs and function names are likely to
29change, and that several tools need to be made available to make this
30all complete. With your help and contributions, this can progress in
31response to the needs and patches provided.
33If you install the distribution in the default location, you can
34access the documentation here:
38There are couple of useful documents you could find in the ./doc
39directory. These include:
41s3_codework.html : Code review document prepared by Ravi Mosur.
42s3_description.html : Code review document prepared by Ravi Mosur.
43s3_overview.html : Code review document prepared by Ravi Mosur.
44s3.2.ppt : The powerpoint presentation prepared by Ravi Mosur.
45FAQ.html : Frequently asked question page prepared by Rita Singh
46sphinxman_manual.html : Manual for training prepared by Rita Singh.
47sphinxman_FAQ.html : FAQ for training prepared by Rita Singh.
48sphinxman_misc.html : Miscellaneous information for training prepared by Rita Singh.
49models.html : Description of the default broad cast news model
51(Obsolete) cmdhelp.txt : Sphinx 3.3 command line arguments.
53For up-to-date information, please see the web site at
58A brief note of the API: From s3.5, the live mode API of the decoder
59is stabilized. The decoder routine is stored in a library called
60libs3decoder.a. The decoder function takes blocks of speech samples
61and returns partial hypotheses for the blocks decoded during
62runtime. Examples of its use are given in main_live_example.c and
65This directory contains the scripts and instructions necessary for
66decoding speech input using the CMU Sphinx Recognizer.
68This distribution is free software, see LICENSE for license.
73This sections contain installation guide for various platforms.
77Starting from Sphinx 3.7, you need to compile sphinxbase before
78compilation of sphinx3. The layout of the directory should look
83By default, sphinx3 will search for ../sphinxbase/ as the path of the
84include files and libraries. You need to compile sphinxbase FIRST
85before compiling sphinx3. Please read the README in the sphinxbase
86package for details.
88This distribution now uses GNU autoconf to find out basic information
89about your system, and should compile on most Unix and Unix-like
90systems, and certainly on Linux.
92If you downloaded directly from the Subversion repository, you need to
93create the "configure" file
98(Make sure you do this twice.)
100If you downloaded a release version, or if you have already run
101"autogen.sh", you can build simply by running
106(Or whatever you call GNU Make). This should configure everything
107automatically. After a successful compilation, you may test the system
110 >make check
112and then install it with
114 >make install
116This defaults to installing Sphinx-3 under /usr/local. You may
117customize it by running ./configure with an argument, as in
119 >./configure --prefix=/my/own/installation/directory
123Starting from sphinx 3.7, you need to install sphinxbase before
124installation of sphinx. The general layout looks like this
129To compile the sphinx 3.7 in Visual C++ 6.0,
1301, unzip the file.
1312, click sphinx3.dsw
1323, build the application projects by clicking "Build" at the tool bar.
134Step 3 can be done easily in VC6.0 by clicking the "Build" menu at
135the top bar, and selecting "Batch Build...". Make sure all projects
136are selected, preferably the "Release" version of each. You can also
137run with the "Debug" version, but this version is usually slower. To
138run the executable, please remember to set the corresponding active
141If you are using cygwin, the installation procedure is very similar to
142the Unix installation. However, there is no audio driver support in
143cygwin currently so one can only use the batch mode recognzier.
149This work was built over a long years at CMU by most of the people in
150the Sphinx Group. Some code goes back to 1986. The most recent wor for
151release after sphinx 3.0 includes the following, listed
154Alan W Black (email@example.com)
155Arthur Chan (firstname.lastname@example.org)
156Evandro Gouvea (email@example.com)
157Ricky Houghton (firstname.lastname@example.org)
158David Huggins-Daines (email@example.com)
159Kevin Lenzo (firstname.lastname@example.org)
161Rita Singh (email@example.com)
162Yitao Sun (firstname.lastname@example.org)