| 列挙子 |
発生タイミングと得られる情報 |
備考 |
| CALLBACK_POLL |
認識時に周期的に呼ばれる |
|
| CALLBACK_STATUS_PARAM |
第1パスの後で呼ばれ、入力の長さなどの情報を得られる |
|
| CALLBACK_ADIN_CAPTURED |
be called at every time a small audio fragment has been read into Julius. |
A/D-in plugin callback to access to the captured input. This callback will be processed first in Julius, and after that Julius will process the content for recognition. This callback can be used to monitor or modify the raw audio input in user-side application. |
| CALLBACK_ADIN_TRIGGERED |
be called for input segments triggered by level and zerocross. |
A/D-in plugin callback to access to the triggered input. After processing this callback, Julius will process the content for recognition. This callback can be used to monitor or modify the triggered audio input in user-side application. |
| CALLBACK_PAUSE_FUNCTION |
be called inside Julius when the engine becomes paused. |
When Julius engine is required to stop by user application, Julius interrupu the recognition and start calling the functions registered here. After all the functions are executed, Julius will resume to the recognition loop. So if you want to use the pause / resume facility of Julius, You should also set callback function to this to stop and do something, else Julius returns immediately. |
| 列挙子 |
発生タイミングと得られる情報 |
備考 |
| CALLBACK_EVENT_PROCESS_ONLINE |
the engine becomes active and start running. |
ex. resume by j_request_resume() |
| CALLBACK_EVENT_PROCESS_OFFLINE |
the engine becomes inactive and stop running. |
ex. pause or terminate by user request |
| CALLBACK_EVENT_STREAM_BEGIN |
(not implemented yet) |
|
| CALLBACK_EVENT_STREAM_END |
(not implemented yet) |
|
| CALLBACK_EVENT_SPEECH_READY |
engine is ready for recognition and start listening to the audio input. |
|
| CALLBACK_EVENT_SPEECH_START |
input speech processing starts. |
This will be called at speech up-trigger detection by level and zerocross. When the detection is disabled (i.e. file input), This will be called immediately after opening the file. |
| CALLBACK_EVENT_SPEECH_STOP |
input speech ends. |
This will be called at speech down-trigger detection by level and zerocross. When the detection is disabled (i.e. file input), this will be called just after the whole input has been read. |
| CALLBACK_EVENT_RECOGNITION_BEGIN |
a valid input segment has been found and speech recognition process starts. |
This can be used to know the actual start timing of recognition process. On short-pause segmentation mode and decoder-based VAD mode, this will be called only once at a triggered long input. @sa CALLBACK_EVENT_SEGMENT_BEGIN. |
| CALLBACK_EVENT_RECOGNITION_END |
a valid input segment has ended up, speech recognition process ends and return to wait status for another input to come. |
On short-pause segmentation mode and decoder-based VAD mode, this will be called only once after a triggered long input. @sa CALLBACK_EVENT_SEGMENT_END. |
| CALLBACK_EVENT_SEGMENT_BEGIN |
ショートポーズによって分割された、各セグメントの最初 |
ショートポーズセグメンテーションとデコーダベースVAD (decoder-based VAD) が有効のとき |
| CALLBACK_EVENT_SEGMENT_END |
ショートポーズによって分割された、各セグメントの最後
(分割された各セグメントのCALLBACK_RESULTよりも後)
|
ショートポーズセグメンテーションとデコーダベースVAD (decoder-based VAD) が有効のとき |
| CALLBACK_EVENT_PASS1_BEGIN |
the 1st pass of recognition process starts for the input. |
|
| CALLBACK_EVENT_PASS1_FRAME |
periodically at every input frame. |
This can be used to get progress status of the first pass at each frame. |
| CALLBACK_EVENT_PASS1_END |
the 1st pass of recognition process ends for the input and proceed to 2nd pass. |
|
| CALLBACK_EVENT_PASS2_BEGIN |
the 2nd pass of recognition process starts. |
|
| CALLBACK_EVENT_PASS2_END |
the 2nd pass of recognition process ends. |
|
| CALLBACK_EVENT_PAUSE |
the engine becomes paused. |
|
| CALLBACK_EVENT_RESUME |
the engine becomes resumed. |
|