A smart subtitle system for emotional expression in mixed-language communication
MOOD is a real-time subtitle enhancer that detects emotional tone in spoken language and transforms it into visual cues.
It’s built for Deaf users to better understand how something is said, not just what is said—bridging the emotional gap often lost in plain captions.
Detect tone of voice, volume shifts, and speech rhythm
Recognize emotional signals like excitement, calmness, sarcasm, or frustration
Designed to work with live conversation or pre-recorded voice content
Using audio analysis and voice emotion AI, MOOD interprets vocal tone in real time.Subtitles are then visually modified—through color, weight, or animation—to express emotional intent.
Subtitles often miss the emotion.MOOD brings that emotion back—by letting tone speak visually.
Most existing subtitle systems focus solely on transcribing words. But in mixed-language conversations—especially between Deaf and hearing individuals—emotional tone is often lost.Through interviews with Deaf viewers, we learned how challenging it can be to distinguish a joke from sarcasm, or calmness from passive aggression—when tone is invisible.This inspired us to ask:
Can subtitles carry emotion, not just words?
“When hearing people say ‘fine,’ I don’t know if they’re happy, annoyed, or just tired. It all sounds the same in text.”
“Deaf individuals often extract emotions by observing facial expressions from other Deaf people. However, hearing individuals typically do not rely heavily on facial expressions to convey their emotions.”
How might we bridge emotional gaps between Deaf and hearing communities? How might we make emotions more visible and mutual, beyond words, sounds, or spoken tone?
