Load Language Resources

You can manually load a resource before you start a task. If the resource is not loaded when the task starts, the task automatically loads the resource.

A resource can be loaded only if there are free slots available for that type of resource.


Whether the resource is loaded manually or automatically affects whether the resource is unloaded after it is no longer needed, as is covered in detail in the next section.

The following parameters control the maximum number of each resource type:

If you set any of the parameters to 0, loading of that resource type is unavailable.

If you set any of the parameters to -1, you can load an unlimited number of that resource type, memory permitting.

The more resources that you have loaded at any one time, the more memory the server uses. However, because loading resources can take time, you can save time when running a task for which the required resource is already loaded.


You shonsider memory usage and load times when you determine the limits that you want to place on the number of resources that can be loaded simultaneously.

Memory usage and load times are significant considerations for language packs. Each language pack might take 10 or more seconds to load, and can take up approximately 700 MB of memory. However, speaker identification base packs are much smaller, and load almost instantly; in addition, it is unlikely that you would need to use more than one speaker identification base pack.

Audio fingerprint databases tend to load considerably faster than language packs, and use significantly less memory. However, performance does depend on the amount of data stored in the database, and the cache size selected for database access.