Announcement voice generator for PA systems
Store closing in fifteen minutes, a platform change, the start of an event, a lost child at the information desk: announcements have to be clear the first time, in a calm voice, often in several languages. Recording each one takes a speaker and a quiet room every time the text changes.
SuperClonica turns typed announcements into audio files with natural voices, free to test online and offline on your own computer. This guide shows how to write announcements people understand, how to voice a whole set of them and which voice to choose.
Voices for announcements
Most catalog voices were recorded in Russian and keep a slight accent in other languages, so test your announcement in your language before you choose.
- The ten stock voices of the model, M1 to M5 and F1 to F5: neutral, free in the engine and needing no extra files.
- Pleasant Announcer: a low, restrained male voice made for instructions.
- Business Male Voice, Eva and Nika: calm, even voices with no character attached.
- Your own voice or a colleague’s who agreed, trained in the studio, for a recognizable house voice.
Voice one announcement free online
On the page of any voice, type a short announcement of 30 to 100 characters, choose the language and press Voice it. In a few seconds you can listen and download the recording, with no account.
Write announcements people understand
Put a short chime before the voice in an audio editor: listeners turn their attention to it before the first word.
- Start with an attention phrase, such as “Attention, please.”, and give the key fact first.
- Short sentences with full stops: speakers in large halls echo, and long sentences blur.
- Repeat the most important detail, such as a time, a platform or a gate.
- Write times and numbers the way they should be heard; the engine spells digits out in 27 languages, so check by ear.
A whole set of announcements offline
Install the free SuperClonica engine: python -m pip install superclonica-cpu. Voice each announcement into its own file, for example superclonica --text "Attention, please. The store will close in fifteen minutes. Please bring your purchases to the checkout." --voice M2 --lang en --normalize --output closing.mp3. The --normalize option brings every file to the same loudness, −18 LUFS, so no announcement is suddenly louder than the others, and the same voice can read the same text in up to 31 languages. When opening hours change, edit one sentence and voice it again.
Before it goes on the speakers
Listen to every file on the actual speakers, in the actual room: an announcement that is clear on headphones can blur in a hall full of echo. Slow it down with --speed if needed. For safety and evacuation messages, follow the regulations that apply to your building and test with the people responsible. Do not use character or parody voices for public announcements.