Back to Documentation

Transcribe & Add Captions

Automatically transcribe videos and add synchronized captions

Transcribe & Add Captions

Automatically transcribe videos and add synchronized captions with word-level timing for perfect subtitle synchronization.

What is Transcription?

Transcription converts spoken audio from your videos into text captions that are perfectly synchronized with the speech. Using AI-powered speech recognition, TellFrame extracts audio from your videos, transcribes it, and creates captions with precise word-level timing (speech marks) for accurate subtitle display.

Key Features

  • Automatic Transcription - AI-powered speech-to-text conversion
  • Word-Level Timing - Precise synchronization with speech marks
  • Language Detection - Automatically detects the spoken language
  • Multiple Formats - Works with video files and audio tracks
  • Seamless Integration - Captions automatically added to your video
  • High Accuracy - Powered by OpenAI Whisper for 99%+ accuracy

How to Use

Transcribing a Video

  1. Select Your Video
    • In the editor, select a video element or scene
    • Or right-click on a video in the timeline
  2. Start Transcription
    • Click Transcribe & Add Captions in the toolbar
    • Or right-click and select Transcribe & Add Captions
  3. Wait for Processing
    • The system will show progress through 5 steps:
      • Step 1/5: Loading video metadata
      • Step 2/5: Extracting audio from video
      • Step 3/5: Transcribing audio
      • Step 4/5: Processing transcript
      • Step 5/5: Creating captions
  4. Review Results
    • Captions are automatically added to your video
    • Text appears with perfect timing synchronization
    • Speech marks ensure word-by-word highlighting

From the Toolbar

  1. Go to the Video Editing toolbar
  2. Click Transcribe and add captions
  3. Select the video you want to transcribe
  4. Wait for the transcription to complete

From the Timeline

  1. Right-click on a video block in the timeline
  2. Select Transcribe & Add Captions from the context menu
  3. Transcription starts automatically

How It Works

Step 1: Loading Video Metadata

The system loads your video file and determines its duration and properties.

Step 2: Extracting Audio

Audio is extracted from the video file. This process:
  • Separates audio from video tracks
  • Converts to a format suitable for transcription
  • Shows progress percentage during extraction
Note: If your video has no audio track, the process will stop with a notification.

Step 3: Transcribing Audio

The extracted audio is sent to OpenAI Whisper API for transcription:
  • Language Detection - Automatically detects the spoken language
  • Word-Level Timing - Creates precise timestamps for each word
  • High Accuracy - 99%+ accuracy for clear speech

Step 4: Processing Transcript

The transcription is processed to create:
  • Full Text - Complete transcript of all spoken words
  • Speech Marks - Word-level timing data for synchronization
  • Language Code - Detected language for proper formatting

Step 5: Creating Captions

Captions are created and added to your video:
  • New Block - Creates a new text block with transcribed content
  • Or Updates Existing - Updates existing block if transcribing a video that already has text
  • Perfect Sync - Speech marks ensure captions match speech timing

Supported Formats

Video Formats

  • MP4 - Most common format
  • WebM - Web-optimized format
  • MOV - QuickTime format
  • AVI - Windows format
  • Any format with an audio track

Audio Formats

  • MP3 - Compressed audio
  • WAV - Uncompressed audio
  • M4A - Apple audio format
  • OGG - Open-source format

Language Support

Transcription automatically detects the language, but works best with:
  • English (all variants)
  • Spanish
  • French
  • German
  • Italian
  • Portuguese
  • And 50+ other languages
The system automatically detects the language and uses it for improved accuracy.

File Size Limits

Audio Size Limit

  • Maximum: 25MB (OpenAI Whisper API limit)
  • Recommended: Under 20MB for faster processing

What Happens if File is Too Large?

If your extracted audio exceeds 25MB:
  • You'll see an error message
  • The file size will be displayed
  • Options:
    • Use a shorter video segment
    • Compress the audio before uploading
    • Split the video into smaller parts

Understanding Speech Marks

Speech marks are timing data that enable:
  • Word-by-Word Highlighting - Each word appears at the exact right time
  • Synchronized Captions - Perfect timing with speech
  • Subtitle Styles - Enables word-by-word, highlight, and other subtitle types

How Speech Marks Work

Each word in your transcription gets:
  • Start Time - When the word begins (in milliseconds)
  • Word Text - The actual word
  • Type - Marked as "word" for processing
This data enables advanced caption features like:
  • Word-by-word appearance
  • Highlighting individual words
  • Color-coding words
  • Size animations

Use Cases

Adding Captions to Existing Videos

  1. Upload a video with speech
  2. Transcribe it
  3. Captions are automatically added
  4. Perfect for making videos accessible

Creating Videos from Recordings

  1. Record a video with your phone or camera
  2. Upload to TellFrame
  3. Transcribe to get text
  4. Edit and enhance the video

Converting Audio to Video

  1. Upload an audio file
  2. Transcribe to get the text
  3. Add visuals to create a video
  4. Perfect for podcasts and interviews

Improving Accessibility

  1. Transcribe your videos
  2. Captions make content accessible
  3. Helps viewers in noisy environments
  4. Supports multiple languages

Best Practices

For Best Accuracy

  • Clear Audio - Ensure speech is clear and audible
  • Minimize Background Noise - Reduce background sounds
  • Single Speaker - Works best with one speaker at a time
  • Good Microphone - Use quality recording equipment
  • Normal Speech Pace - Not too fast or too slow

For File Size

  • Shorter Videos - Keep videos under 10 minutes when possible
  • Compress Audio - Use compressed formats when possible
  • Split Long Videos - Break long videos into segments
  • Check Before Upload - Verify file size if concerned

For Processing Speed

  • Shorter Videos - Process faster
  • Compressed Audio - Smaller files process quicker
  • Good Connection - Stable internet helps
  • One at a Time - Process videos sequentially

Editing Transcribed Captions

After transcription, you can:

Edit Text

  1. Select the text block with captions
  2. Edit the text in the editor
  3. Speech marks remain synchronized
  4. Perfect for fixing errors or improving clarity

Adjust Timing

  1. Use the timeline to adjust block timing
  2. Speech marks automatically adjust
  3. Maintains word-level synchronization

Change Subtitle Style

  1. Select the text block
  2. Choose a subtitle style:
    • Basic
    • Word-by-word
    • Highlight
    • Modern
    • Single word
  3. Speech marks enable advanced styles

Customize Appearance

  1. Change font, size, and color
  2. Adjust position and layout
  3. Add animations
  4. All while maintaining synchronization

Troubleshooting

"No audio track detected"

Cause: Video has no audio or audio track is missing
Solutions:
  • Check that your video has audio
  • Try a different video file
  • Ensure audio isn't muted in the source

"Audio file too large"

Cause: Extracted audio exceeds 25MB limit
Solutions:
  • Use a shorter video segment
  • Compress the audio before uploading
  • Split the video into smaller parts

"Transcription failed"

Cause: Various possible issues
Solutions:
  • Check your internet connection
  • Verify the video has clear audio
  • Try again (may be temporary API issue)
  • Check if audio is in a supported format

"No speech detected"

Cause: Audio has no speech or speech is too quiet
Solutions:
  • Ensure audio contains speech
  • Check audio volume levels
  • Verify microphone was working during recording
  • Try a different video segment

Transcription is inaccurate

Cause: Audio quality or language issues
Solutions:
  • Improve audio quality (clearer speech)
  • Reduce background noise
  • Ensure language is supported
  • Check if multiple speakers are confusing the system

Processing is slow

Cause: Large file or slow connection
Solutions:
  • Wait for processing to complete
  • Check internet connection
  • Use shorter videos when possible
  • Process during off-peak times

Tips for Best Results

  1. Clear Audio - Ensure speech is clear and audible
  2. Single Speaker - Works best with one person speaking
  3. Good Quality - Use good recording equipment
  4. Appropriate Length - Keep videos reasonable length
  5. Check Results - Always review transcribed text
  6. Edit as Needed - Fix any errors or improve clarity
  7. Use Speech Marks - Take advantage of word-level timing
  8. Test Subtitle Styles - Try different caption styles

Examples

Example 1: Tutorial Video

Original: Video with instructor explaining a process
After Transcription:
  • Full transcript of all speech
  • Word-level timing for perfect sync
  • Ready for subtitle styling
Result: Professional captions that match speech perfectly

Example 2: Interview Video

Original: Interview with multiple speakers
After Transcription:
  • Complete transcript of conversation
  • Synchronized captions
  • Easy to edit and enhance
Result: Accessible interview with perfect captions

Example 3: Presentation Video

Original: Screen recording with narration
After Transcription:
  • All spoken content transcribed
  • Perfect timing with visuals
  • Can be edited and styled
Result: Professional presentation with synchronized captions

Advanced Features

Language Detection

The system automatically:
  • Detects the spoken language
  • Uses appropriate language model
  • Provides best accuracy for detected language

Word-Level Timing

Speech marks enable:
  • Word-by-word appearance animations
  • Highlighting individual words
  • Color-coding words
  • Size animations per word

Multiple Speakers

While best with one speaker, the system can handle:
  • Multiple speakers (may need editing)
  • Different accents
  • Various speech patterns

Integration with Other Features

With Subtitle Styles

Transcribed captions work with all subtitle styles:
  • Basic - Simple text display
  • Word-by-Word - Uses speech marks for word timing
  • Highlight - Highlights words as they're spoken
  • Modern - Contemporary styling
  • Single Word - One word at a time

With Text Editing

After transcription:
  • Edit text freely
  • Speech marks adjust automatically
  • Maintain synchronization
  • Improve clarity and accuracy

With Voice Settings

Transcribed content can be:
  • Used with text-to-speech
  • Matched to voice settings
  • Synced with audio tracks
  • Enhanced with effects
The Transcribe feature makes it easy to add professional, synchronized captions to any video. Simply select your video and let AI do the work!
Video OptimizerVoice & Text-to-Speech