Making language-listening material
You improve at a language roughly in proportion to how much of it you hear. When a textbook has no audio, or you want to hear the words and sentences you collected yourself, TTS lets you build the listening material. The "multilingual" voices here read English and several other languages, so one file can hold both the Korean explanation and the foreign-language sentences.
About the multilingual voices
Voices marked "multilingual" in the list (Ava, Andrew, Emma, Brian and others) are English-first but will also read sentences in other languages. Among the Korean voices, "Hyunsu · multilingual male" can mix English sentences in as well. Naturalness varies by language, so press ▶ on a sentence or two before committing.
1. Vocabulary audio
The simplest approach is one line per entry — the word, then the meaning — read by a multilingual voice.
If you want the English in a native voice and the meaning in a Korean one, use Script mode and name the characters after the languages.
In Character voices, set EN to Ava or Andrew and KO to SunHi or InJoon, and a gap of 1–1.5 seconds leaves room to repeat it silently.
2. Shadowing sentences
Shadowing means repeating a sentence immediately after you hear it, so you need space between sentences. In Script mode set the gap to 2 seconds and write one sentence per line.
| Level | Speed | Gap |
|---|---|---|
| Beginner | "Crisp" tone, or speed -20% | 2 seconds |
| Intermediate | -10% to default | 1.5 seconds |
| Practice at pace | Default to +10% | 1 second |
3. Dictation material
For dictation, reading each sentence twice works well. Simply write the same line twice; the gap can go up to 2 seconds. If you need longer to write, pause playback instead.
The answer key comes free: the subtitle file behind the CC button contains the sentences exactly, ready to mark against.
4. Listening to a whole conversation
To practise following a conversation, use two or more characters. Mixing a female and a male voice makes it easy to tell who is speaking.
Things to watch
- Turning the body text of a commercial textbook into audio and sharing it can raise copyright problems. Use sentences you wrote or openly available material.
- TTS pronunciation is broadly accurate, but proper nouns and rare words can come out awkwardly. For words that matter, compare against a dictionary's native-speaker audio.
Suggested settings by level
| Level | Speed (tone) | Gap | Use for |
|---|---|---|---|
| Beginner | Crisp (about -20%) | 2 seconds | Hear one sentence, repeat it |
| Elementary | Calm (about -10%) | 1.5 seconds | Vocabulary, short exchanges |
| Intermediate | Default | 1 second | Shadowing, dictation |
| Exam preparation | Default or News | 0.5 seconds | Getting used to real pace |
The same sentence slowly, then at normal speed
Splitting one voice across two characters lets you change speed within a single file. For example, set "SLOW" to Ava with the "Crisp" tone and "EN" to the same Ava on "Default", then write it like this.
Slow and distinct first, then at real speed — good for letting the ear adjust.
Define just two characters, EN and KO, in Script mode and your vocabulary audio is ready.
Make listening material