refactor: Improve text conditioning logic in transcription-filter.cpp

locaal-ai · May 10, 2024 · 0e8c28d · 0e8c28d
1 parent fc9383f
commit 0e8c28d
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 5 deletions.
diff --git a/src/transcription-filter.cpp b/src/transcription-filter.cpp
@@ -147,9 +147,14 @@ void set_text_callback(struct transcription_filter_data *gf,
 	}
 	gf->last_sub_render_time = now;
 
-	// recondition the text
-	std::string str_copy = fix_utf8(result.text);
-	str_copy = remove_leading_trailing_nonalpha(str_copy);
+    std::string str_copy = result.text;
+
+	// recondition the text - only if the output is not English
+    if (gf->whisper_params.language != "en") {
+        str_copy = fix_utf8(str_copy);
+    } else {
+        str_copy = remove_leading_trailing_nonalpha(str_copy);
+    }
 
 	// if suppression is enabled, check if the text is in the suppression list
 	if (!gf->suppress_sentences.empty()) {

diff --git a/src/whisper-utils/whisper-processing.h b/src/whisper-utils/whisper-processing.h
@@ -6,9 +6,9 @@
 // buffer size in msec
 #define DEFAULT_BUFFER_SIZE_MSEC 3000
 // overlap in msec
-#define DEFAULT_OVERLAP_SIZE_MSEC 150
+#define DEFAULT_OVERLAP_SIZE_MSEC 125
 #define MAX_OVERLAP_SIZE_MSEC 1000
-#define MIN_OVERLAP_SIZE_MSEC 150
+#define MIN_OVERLAP_SIZE_MSEC 125
 
 enum DetectionResult {
 	DETECTION_RESULT_UNKNOWN = 0,