GUIDE

An alternative to RHVoice for natural voice-overs

RHVoice is a free, open-source speech synthesizer that started with Russian and became a favorite of screen reader users: it works with NVDA, Windows SAPI5, Android and Linux, answers instantly and takes very little space. Its statistical parametric voices are clear, but nobody mistakes them for a person.

The free SuperClonica engine is built for the opposite job: natural neural voices for videos, audiobooks and announcements, offline on an ordinary processor. This guide compares them honestly and says when RHVoice is the right tool.

RHVoice in short

For a screen reader that must answer every key press instantly, even at high speed, RHVoice is hard to beat.

  • Languages: Russian, American and Scottish English, Brazilian Portuguese, Esperanto, Georgian, Ukrainian, Kyrgyz, Tatar, Macedonian, Albanian and Polish.
  • It works with SAPI5 on Windows, Speech Dispatcher on Linux and the Android text-to-speech system, and NVDA uses it through its own driver.
  • Its voices are statistical parametric models based mainly on HTS, which keeps them small and fast.
  • The engine is under GPL-2.0.

Compare the sound free

On the page of any voice, paste a sentence you know in RHVoice, 30 to 100 characters, choose the language and press Voice it. In a few seconds you hear the difference on your own text, with no account.

What SuperClonica does differently

A short phrase takes about 0.4 s on a processor: fine for files, too slow for reading a screen at three times normal speed.

  • Neural voices with natural intonation, made to be listened to, not only understood.
  • 31 languages, including Russian, Ukrainian, Polish, English, German, Spanish, Japanese and Arabic.
  • Ten stock voices and more than 170 catalog voices, from calm narrators to cartoon and game characters, plus your own trained voice.
  • A command-line tool, a Python library and a Docker HTTP server, installed with python -m pip install superclonica-cpu.

Audiobooks and voice-overs from long texts

For a book, a course or a long article, save the text to a file and voice it in one go: superclonica --text-file book.txt --voice F3 --lang en --normalize --output book.mp3. The engine splits long text at natural pauses, keeps the loudness even and writes one file; an hour of audio takes several minutes on an ordinary desktop processor. In Russian, write ё where it belongs and numbers as words, because digits are read as plain numbers and are not inflected.

Which one to choose

Choose RHVoice as the voice of your screen reader, on Android or for Tatar, Kyrgyz, Georgian, Esperanto, Macedonian or Albanian. Choose SuperClonica for audio other people listen to: audiobooks, video voice-overs, lessons, announcements and character voices, in 31 languages and in a voice of your own.