summaryrefslogtreecommitdiffstats
path: root/Whisper/API
diff options
context:
space:
mode:
authorKonstantin <const@const.me>2023-01-28 15:18:02 +0100
committerKonstantin <const@const.me>2023-01-28 15:18:02 +0100
commite1e3ac09a97d602a6ea60ff1928de77de81d99a7 (patch)
treedb38b886ef5d93b76aaecfe89de8fc421616646e /Whisper/API
parent214aacaa5c0a685f8be1cbe4fe06f5a1af8ad2d4 (diff)
DLL API for diarize feature
Diffstat (limited to 'Whisper/API')
-rw-r--r--Whisper/API/TranscribeStructs.h9
-rw-r--r--Whisper/API/iContext.cl.h2
-rw-r--r--Whisper/API/iContext.h2
3 files changed, 13 insertions, 0 deletions
diff --git a/Whisper/API/TranscribeStructs.h b/Whisper/API/TranscribeStructs.h
index ac28357..29bb9ab 100644
--- a/Whisper/API/TranscribeStructs.h
+++ b/Whisper/API/TranscribeStructs.h
@@ -124,4 +124,13 @@ namespace Whisper
{
return 0 != ( (uint32_t)a & (uint32_t)b );
}
+
+ // Output value for iContext.detectSpeaker method
+ enum struct eSpeakerChannel : uint8_t
+ {
+ Unsure = 0,
+ Left = 1,
+ Right = 2,
+ NoStereoData = 0xFF,
+ };
} \ No newline at end of file
diff --git a/Whisper/API/iContext.cl.h b/Whisper/API/iContext.cl.h
index fdb15ce..18e56de 100644
--- a/Whisper/API/iContext.cl.h
+++ b/Whisper/API/iContext.cl.h
@@ -31,6 +31,8 @@ namespace Whisper
virtual HRESULT COMLIGHTCALL runCapture( const sFullParams& params, const sCaptureCallbacks& callbacks, const iAudioCapture* reader ) = 0;
virtual HRESULT COMLIGHTCALL getResults( eResultFlags flags, iTranscribeResult** pp ) const = 0;
+ // Try to detect speaker by comparing channels of the stereo PCM data
+ virtual HRESULT COMLIGHTCALL detectSpeaker( const sTimeInterval& time, eSpeakerChannel& result ) const = 0;
virtual HRESULT COMLIGHTCALL getModel( iModel** pp ) = 0;
diff --git a/Whisper/API/iContext.h b/Whisper/API/iContext.h
index d6ca29c..fc38a53 100644
--- a/Whisper/API/iContext.h
+++ b/Whisper/API/iContext.h
@@ -28,6 +28,8 @@ namespace Whisper
HRESULT __stdcall runCapture( const sFullParams& params, const sCaptureCallbacks& callbacks, const iAudioCapture* reader );
HRESULT __stdcall getResults( eResultFlags flags, iTranscribeResult** pp ) const;
+ // Try to detect speaker by comparing channels of the stereo PCM data
+ HRESULT __stdcall detectSpeaker( const sTimeInterval& time, eSpeakerChannel& result ) const;
HRESULT __stdcall getModel( iModel** pp );