SPRAAK
|
If SPRAAK is not yet installed on your system, pls. refer to the Installation Guide for details on how to install SPRAAK first.
Your local SPRAAK administrator may provide extra information in the Local Pages on how to use SPRAAK in your local computing environment.
Assuming that
SPR_HOME is the home directory of the SPRAAK installation OPSYS is the operating system and architecture you are running on, e.g. Linux_x86_64
then you should verify that following PATH variables have been set correctly, such that the commands and scripts of the SPRAAK package are included in your PATHs.
PATH should include $SPR_HOME/bin:$SPR_HOME/bin/$OPSYS LD_LIBRARY_PATH should include $SPR_HOME/lib/$OPSYS
Developers who want to include Python routines from the SPRAAK package will also need to set their PYTHON_PATH appropriately.
You may run the command
> eval `spr_version.py <vnum>`
to set to a different version if you have multiple version on your machine. Or you may run the same command with the empty string to get all PATH settings correct on the basis of SPR_HOME.
Most of SPRAAK and its functionalities can be used in a WINDOWS envirnoment, though certain aspects such as multi-threading will not operate properly. Hence SPRAAK has mainly been used on WINDOWS in "run-time" mode, i.e. to perform recognition or alignments and not for developing of new acoustic models.
Pls. refer to the description on UNIX.
Users who previously worked with the ESAT speech recognition software HMM V7.x and now are converting to SPRAAK should read What's new in SPRAAK V1.0.
There are 2 main sections to the manual, which you see as the Tabs on the top line of your current screen.
This part of the manual is structured into different sections with different target uses and users and tries to get you started in SPRAAK. 99% of the users will find everything they need here.
What is SPRAAK? | What is SPRAAK ? |
Features of the SPRAAK Toolkit | List of the features of SPRAAK toolkit |
Installation Guide | Describes how to install SPRAAK |
User Manual | User manual for the everyday user of SPRAAK |
Reference Manual | Reference manual with lists of programs, scripts, configuration files, data files, file structures, ... |
Developers Manual | Developers manual who want to dig deep into SPRAAK or want to modify the internals |
This part of the manual is all automatically extracted from the source code and is ented as such much closer onto the code. It contains lists of files, lists of variables, functions, enumerations, etc. .. directory structure of the code, ... This part is mainly handy for developers.
Depending on your background in speech recognition and experience with other speech recognition toolkits your further moves may be quite different.
The next section of the user manual (Introduction to SPRAAK concepts, conventions and file structures) contains introductory material about many different aspects of the SPRAAK toolkit. The section after that contains baseline tutorials on how to set up speech recognition experiments with the TIMIT database (Tutorials and Reference Experiments on TIMIT).