To create a task, send an AddTask
ACI action to your IDOL Speech Server host and ACI port.
For example:
http://host:port/action=AddTask&Type=task&requiredParams&optionalParams
where:
host
|
is the IP address or name of the machine where IDOL Speech Server is installed. |
port
|
is the ACI port by which actions are sent to IDOL Speech Server (set by the Port parameter in the IDOL Speech Server configuration file's [Server] section). |
task
|
is the name of the configuration section where the options for the task are defined in the IDOL Speech Server tasks configuration file. |
requiredParams
|
You must supply requiredParams for the action you request. (Not all tasks have required parameters.) |
optionalParams
|
You can supply optionalParams for the action you request. (Not all tasks have optional parameters.) |
For example:
http://localhost:15000/action=AddTask&Type=SpeechToText&File=Speech.wav&Out=Text.ctm
This action creates a SpeechToText
task on the IDOL Speech Server located on the local machine, with the ACI port 15000
.
Some tasks can run across multiple cores for faster performance. The audio to process is split into chunks, which are shared out between the IDOL Speech Server task managers (Micro Focus recommends that you set one task manager for each core). Each task manager processes its allocated chunks, and the results from all task managers are combined at the end.
The following modules support multicore processing. If you request multicore processing for a task that contains modules that are not in this list, IDOL Speech Server returns an error.
audiopreproc
|
frontend
|
postproc
|
audio
|
mixer
|
stt
|
ctm
|
normalizer
|
wout
|
filter
|
plh
|
When using the stt
module, multiple core processing is semantically meaningful only in fixed mode; relative mode is not supported.
Each running task manager counts as an IDOL Speech Server instance. If insufficient licenses are available for the task, IDOL Speech Server returns an error.
Multicore processing is not supported if server queuing is enabled (see Configure Task Queues).
To run a task on multiple cores, you must set the following three action parameters when you send the AddTask
action.
TaskManagers
|
The number of task managers to split the task across. If you specify more task managers than are available, IDOL Speech Server returns an error. |
SplitSize
|
The size of the chunks, in seconds, to divide the audio into. For most tasks, Micro Focus recommends between 60 and 300 seconds. |
Overlap
|
The number of seconds of overlap between chunks. Increasing the overlap improves accuracy but reduces the overall processing speed. For most tasks, Micro Focus recommends an overlap of about 5 seconds. |
For example:
http://localhost:15000/action=AddTask&Type=SpeechToText&File=Speech.wav&Out=Text.ctm&TaskManagers=3
This action creates a SpeechToText
task across three task managers.
|