Multilingual Teleprompter: Reading Scripts in 9 Languages

Most teleprompter apps assume you're speaking English. They work fine for that. But try using one for a Japanese or Korean script, and you'll run into problems fast. Speech recognition gets confused, character rendering looks wrong, and scroll pacing is off because the app was never designed for languages with different character densities.

What makes non-English teleprompting hard

Languages are surprisingly different in ways that matter for teleprompters:

• Chinese and Japanese express ideas in fewer characters than English, so the same content takes up less screen space. Scroll pacing designed for English moves too fast.
• Japanese and Chinese don't use spaces between words. Word-level highlighting, a common teleprompter feature, doesn't map cleanly to these languages.
• Speaking pace varies by language. Japanese averages around 7.84 syllables per second vs. English at about 6.19, according to a 2011 study published in Language. That's a meaningful difference for scroll timing.

How VoiceScroll handles this

VoiceScroll supports 9 languages:

• English, Korean (한국어), Japanese (日本語)
• Chinese Simplified (简体中文), Chinese Traditional (繁體中文)
• German (Deutsch), French (Français), Portuguese (Português), Spanish (Español)

Each language runs on Apple's native speech recognition model for that language. It's not English recognition with a translation layer. The app understands pronunciation patterns, accents, and word boundaries specific to each language.

Since scrolling is driven by voice matching rather than a fixed speed, the pacing issue goes away. The app follows your speech regardless of which language you're speaking in.

Notes on specific languages

Korean

Korean's syllable blocks (like 안녕하세요) work well with word-level matching. Korean grammar is agglutinative, so individual words can get long. Use a slightly larger font size than you would for English.

Japanese

Kanji, hiragana, and katakana all have different character widths. This makes line lengths unpredictable. Voice matching sidesteps the problem because it tracks your speech rather than counting characters on screen.

Chinese

Each character is roughly the same width, which actually makes Chinese easier to display than Japanese. But without spaces, the app needs to use speech recognition rather than text analysis to figure out where you are in the script.

European languages

German has notoriously long compound words (like Geschwindigkeitsbegrenzung). French and Portuguese have lots of short connecting words. Both work fine, but you might need to adjust font size depending on the language.

Multi-language workflow tips

1. Don't machine-translate scripts. Write each version as a separate script adapted for natural speech in that language.
2. Test the speech recognition with a quick run-through. Accents can affect accuracy.
3. Adjust font size per language. What works for English won't necessarily work for Chinese characters or German compound words.