Transcription Problems

Having issues with transcription quality or processing? This guide helps you diagnose and fix common transcription problems.

Transcription Failed

Cause

Transcription can fail for several reasons:

• Corrupted or invalid audio/video file
• Unsupported codec or container format
• File too large for processing (over 2GB)
• Modal.com transcription service error
• Network timeout during processing
• R2 storage access issues

Solution

1. Verify file integrity:
   • Try playing the file in VLC or another media player
   • If it won't play, the file is corrupted - try re-recording or re-exporting
2. Check file format:
   • Use common formats: MP4 (H.264), MP3, WAV
   • Avoid exotic codecs or uncommon container formats
3. Reduce file size:
   • If file is over 2GB, compress it using HandBrake or similar tools
   • For very long recordings, consider splitting into chunks
4. Retry the upload:
   • Delete the failed content and upload again
   • Failed attempts don't count toward your daily limit

---

Poor Transcription Accuracy

Cause

Low accuracy can result from:

• Poor audio quality (background noise, low volume, echo)
• Heavy accents or dialects
• Technical jargon or uncommon terminology
• Multiple overlapping speakers
• Fast speech or mumbling
• Non-standard language usage (slang, abbreviations)

Solution

1. Use Initial Prompt

Provide context to improve accuracy when uploading:

• Technical terms: "This is a software engineering discussion about React, TypeScript, and API endpoints"
• Names: "Participants include Dr. Sarah Johnson, Michael Chen, and Emma Williams"
• Industry jargon: "Medical consultation discussing diabetes, insulin resistance, and HbA1c levels"

2. Improve Audio Quality

• Use external microphones instead of built-in laptop mics
• Record in quiet environments
• Position microphone 6-12 inches from speaker
• Avoid background noise (fans, air conditioning, traffic)
• Use audio filters to reduce noise if needed

3. Speaker Considerations

• Speak clearly at normal pace
• Avoid talking over each other (enable speaker diarization to help)
• Pause between sentences
• Spell out uncommon names or terms when first mentioned

4. Post-Processing

• Edit the transcript directly in VideoToBe after completion
• Use the Chat feature to ask AI to fix specific errors
• Download and edit in your preferred text editor

---

Speaker Identification Not Working

Cause

Speaker diarization (identifying different speakers) can fail when:

• Voices are too similar (same gender, age, accent)
• Poor audio quality makes voices indistinguishable
• Speakers talk over each other frequently
• Multiple people in same room sharing one microphone

Solution

• Use separate microphones: Best results come from individual mic per speaker
• Enable speaker diarization: Make sure to check "Speaker Diarization" when uploading
• Edit speaker names: After transcription, rename "Speaker 1", "Speaker 2", etc. to actual names
• Accept limitations: Very similar voices may be grouped together - you can manually split sections after transcription

---

Processing Takes Too Long

Normal Processing Time

Expected transcription times:

• Short files (0-10 min): 1-3 minutes
• Medium files (10-30 min): 3-8 minutes
• Long files (30-60 min): 8-15 minutes
• Very long files (60+ min): 15-30 minutes

If Processing Exceeds Expected Time

1. Check status: Refresh the page to see current status
2. Wait for email: You'll receive email notification when ready
3. Check after 30 minutes: If no email after 30 minutes, the job may have failed
4. Retry: Delete the content and upload again

---

Missing or Incomplete Transcript

Cause

• File has no audio track (video only)
• Audio is completely silent or extremely quiet
• Processing interrupted before completion

Solution

1. Verify audio exists:
   • Play the file and ensure audio is audible
   • Check volume levels - should be clearly audible
2. Check file format:
   • Some video files have separate audio tracks that may not be detected
   • Try converting to MP4 with AAC audio
3. Re-upload: Delete and upload again to retry processing

---

Translation Issues

Translation to English Not Working

Cause

• Audio is already in English (no translation needed)
• Language not supported by Whisper
• Mixed languages in single file

Solution

• Check language support: Whisper supports 90+ languages including Spanish, French, German, Chinese, Japanese, and more
• Single language per file: For mixed-language content, Whisper will transcribe the dominant language
• Verify translation enabled: Make sure you checked "Translate to English" when uploading

---

Transcript Has Wrong Language

Cause

Whisper auto-detects language. It may misidentify if:

• Audio quality is very poor
• File contains multiple languages
• Speaker has heavy accent
• Very short audio clip (under 30 seconds)

Solution

• Use the "Translate to English" option to force translation
• Ensure audio is at least 30 seconds long for accurate detection
• Use initial prompt to specify language context

---

Best Practices for Accurate Transcription

Before Recording

• Test your microphone and audio levels
• Choose a quiet location
• Close windows and turn off fans/AC if possible
• Use headphones for video calls to reduce echo

During Recording

• Speak clearly at moderate pace
• Avoid interrupting or talking over others
• State names clearly when introducing people
• Spell out acronyms or technical terms when first used

When Uploading

• Use initial prompt for technical content
• Enable speaker diarization for multi-person recordings
• Select "Translate to English" if audio is in another language
• Use high-quality audio formats (WAV, FLAC) when possible