I've encountered character encoding issues with Japanese text in Superwhisper 1.44.3. Some Japanese characters are being corrupted and displayed as '����' in the output JSON file.
Details:
- App Version: 1.44.3
- Model: Ultra V3 Turbo
- LLM Model: Standard V1
- OS: macOS
- Language: Japanese
Example :
Original text includes corrupted characters:
"見積もり"→"見��もり"
"ご連絡"→"ご連��"
Steps to Reproduce:
1. Set language to Japanese
2. Record or import audio with Japanese speech
3. Generate transcription
4. Corrupted characters appear in the result
Additional Information:
This appears to be a UTF-8 encoding issue when handling Japanese text. The corruption consistently occurs with certain Japanese characters.
Please let me know if you need any additional information to investigate this issue.