Principles
We have 3 separate tiers where speech related information is stored:
- Every user has his/her own home directory.
You can access someones homedirectory using either ∼<username> or using /home/speech/<username>.
To allow easy collaboration, the default is that data in your homedirectory can be accessed by the other group members.
Use the command chmod (see useful commands) to control who can access what.
- Shared data is managed by a set of 3 meta-users:
- spchdata:
- Here you find all databases.
- spchprog:
- Here you find all our software packages.
- spch:
- Here you can find other shared resources such as presentations, published papers, conference processings, ...
You can access this shared data using either ∼<spchdata,spchprog,spch> or using /speech/data/, /speech/prog/, or /speech/spch.
- A large storage pool that is not backed-up. Most experiments involve converting a database in yet antoher format.
While it is relevant to backup all scripts and results of an experiments, backing-up yet another copy of a databases is just a waste of resources.
Hence, you are encouraged to store intermediate results in a subdirectory of this disk system. By convention, this would be in /speechdump/<username>/<whatever_subdir_you_want>.
Databases
All databases are managed by the meta-user
spchdata. The databases can be found using either
- ls -l ∼spchdata/
- ls -l /speech/data/
Software
Speech related tools are managed by the meta-user
spchprog. The installed software packages can be found using either
- ls -l ~spchprog/
- ls -l /speech/prog
In case you use the advised login scripts, you should be able to access all software directly.
Other common resources
Other shared resources are managed by the meta-user
spchg. You can find the common resources using either
- ls -l ~spch/
- ls -l /speech/spch