Julius

Julius/Julian rev.3.5.3 is an update release. The most outstanding
feature of this release is that some code optimization around acoustic
computation made the recognition process faster up to 1.5 times.
Other improvements has been done for graph precision, usability,
stability and code modurality. All users using previous versions of
Julius/Julian are encouraged to upgrade to this release.

Summary of changes in 3.5.3:

o Recognition speed-up by code optimization for acoustic
computation and memory allocation.
o Support another way of specifying acoustic parameter conditions:
- can read an HTK Config file directly by "-htkconf".
- can embed acoustic analysis condition parameters into binary AM.
o New grammar-related tools: "dfa_minimize", "dfa_determinize".
(An HTK-to-Julian automatic grammar convertion tool "slf2dfa" is
also released)
o Other refinements:
- support generating separate candidates of different phone
context in word graph output.
- preliminary support for emulating energy normalization on live input.
o Bug fixes and code improvements.

Details of the changes are listed in "Release.txt".

Please note that if you want to compile Julius with DirectSound
support, you need DirectX headers installed in your mingw / cygwin
environment. If not exist, configure will choose an old interface.
Please see install instruction on the Web.

3.5.3 (2006.12.29)
===================
o Improved Performance:
- acoustic computation optimized: now becomes 20%-40% faster!
- optimize memory access: re-use work area of deleted hypothesis
in the 2nd pass.
- some memory allocation improvement on dictionary and word trellis.

o New Grammar Tools:
- "dfa_minimize", "dfa_determinize" will minimize/determinize DFA.
mkdfa.pl now calls dfa_mimize in it.
- "slf2dfa": a toolkit to convert HTK slf to Julian dfa (separate kit)

o Embedding HTK Acoustic Parameters:
- add option to load HTK Config file to set correct acoustic parameter
configuration at recognition time.
- the acoustic parameter configuration can be embedded into
header of a binary HMM file.

o Improved Word Graph:
- add an option to completely separate graph words: words with
different phone contexts can be output separatedly by
"-graphrange -1".

o Support for online energy normalization:
- Preliminary support for live recognition using acoustic model with
energy normalization. (approximate with maximum energy of last input)

o Code refinements:
- re-organize libsent/src/wav2mfcc.
- modularize acoustic parameter (Value) handling.
- output compile-time configuration of libsent with "--setting" option.
- Doxygen 1.5.0 support.
- "julius-info@lists.sourceforge.jp" becomes the official contact address.
- fixed typo on copyright notice.

o Fixed bugs:
- sometimes unable to read a binary LM on "--enable-words-int".
- memory leaks around option handling, global variables and local buffers.
- segmentation fault on very long input.
- doublely counted initial state of DFA.
- mkdfa.pl: unable to find mkfa on some OS.
- adintool: makes empty output file on termination.
- adintool: miss last inputs when killed.
- other small changes.