Balabolka

Description

Balabolka is a free text-to-speech software for Windows that reads written text aloud. The name is derived from a Russian word that means “chatterer.” The program is created and maintained by Ilya Morozov, an independent developer. Balabolka runs completely offline — it does not send any data to the internet, and requires no account, subscription or cloud connection of any kind. Everything it does takes place on the user’s own computer.

Balabolka does not have a voice built into it. Instead, it works with any speech voices already installed on Windows, using Microsoft’s Speech Application Programming Interface (SAPI) to connect to them. This design makes Balabolka more like a powerful front end for whatever voices a user has available than a standalone speaking engine. Users who wish to have more natural sounding voices can install additional SAPI compatible voices from Microsoft, or from third-party providers such as Acapela, CereVoice, IVONA, or Nuance, and Balabolka picks them up automatically.

SPEECH ENGINES AND VOICES

Balabolka works with two main Microsoft voice frameworks: SAPI 4 and SAPI 5. Windows comes with a default English voice (Microsoft Anna, or the newer Microsoft David and Zira on Windows 10 and later) that works right after installation. Users who require other languages or higher quality voices install them separately, and Balabolka lists all the installed voices in its voice selector dropdown.

Microsoft’s Speech Platform, a separate framework targeted at developers, also integrates with Balabolka and provides access to a larger set of voices in many languages. Third-party SAPI 5 voices from commercial providers had historically been more natural sounding than the default Windows voices; newer Windows 10 and 11 neural voices from Microsoft have improved the default quality considerably.

Voice quality is highly variable depending on the voices installed. Default Windows voices sound synthetic and robotic to most people. Neural voices from Microsoft and paid third-party voices sound more natural. The output Balabolka produces reflects whatever voice is selected — the program itself adds no processing of its own.

READING TEXT

Users can import text into Balabolka in a number of ways. They can open a document file directly, paste or type text into the main editing area, or turn on Clipboard Watch, which causes the program to read aloud anything the user copies to the clipboard automatically.

Balabolka opens a long list of document formats including TXT, DOC, DOCX, RTF, ODT, WPD, PDF, HTML, EPUB, MOBI, AZW, AZW3, FB2, FB3, CHM, DjVu, LIT, PDB, PRC, TCR, PPT, PPTX, XLS, XLSX, ODS, ODP, EML and MD. This range covers most common document types — Word files, PDFs, e-books in several formats, web pages, spreadsheets, presentations and plain text. The program reads the text from the file and displays it in its editing area, where the user can make changes before the text is played.

As the voice reads the words, Balabolka highlights the current word on screen, making it easy to follow along or to find where in the text the voice has reached.

VOICE CONTROLS

Three sliders are used to control the voice output: rate (speed), pitch (tone), and volume. Users drag the sliders to control how quickly the voice reads, how higher or lower in pitch it sounds, and how loud the output is. These adjustments are made in real time while playing the video or at the start of the next read.

For more precise control, Balabolka supports SSML (Speech Synthesis Markup Language), a standard format that uses tags embedded in the text to control pauses, emphasis, pitch changes, and pronunciation at specific points. Users who are comfortable with markup language can add SSML tags directly to the text to fine-tune how the voice handles certain words or passages.

PRONUNCIATION CORRECTION

Balabolka has a substitution list in which users can set custom pronunciation rules. Each rule instructs the program to replace one word or phrase with another spelling that causes the voice engine to pronounce the word correctly. For instance, if the voice mispronounces a brand name, a product term, or a foreign word, the user adds a rule that replaces the original text with a phonetic spelling before the voice reads it. The substitution list also accepts regular expressions for users who need to make pattern-based substitutions in many similar words at once.

This feature addresses one of the most common annoyances with text-to-speech software — voices that always mispronounce domain-specific terms, proper names or abbreviations. Once a rule is in the substitution list, Balabolka applies it each time that word appears in any document.

SAVING AUDIO FILES

Balabolka can save the output of the speech as an audio file instead of just playing it back through the speakers. Supported output formats include WAV, MP3, OGG, WMA, MP4 (M4A), M4B, and AWB. Users select the format and file location and the program renders the audio. The resulting file is playable on any device or media player that supports the format chosen.

The M4B format, which is used for audiobooks, supports chapter markers. Balabolka can add bookmarks as chapter points to M4B files, making long documents easier to navigate when played back in an audiobook-capable player.

SYNCHRONIZED Text (LRC and MP3 TAGS)

One unique feature of Balabolka is the ability to save the text that was spoken in sync with the audio. The program is able to place synchronized text into the ID3 tags of an MP3 file, or save it as a separate LRC file along with the audio. When the audio is played back on a compatible media player, the text scrolls along with the narration — the same way song lyrics scroll on music players that support LRC files.

This feature is useful for creating audiobooks or narrated documents where the listener benefits from seeing the text as it is spoken, or for creating closed caption files to accompany audio content.

BATCH CONVERSION

Balabolka is able to process multiple files at once with its batch conversion tool. Users select a list of documents, choose an output audio format, and the program converts each file in sequence without requiring any further input. Batch conversion is useful for converting a folder of e-books, reports or articles into audio files, or for creating narration files for a series of documents.

The program also supports command-line operation, enabling advanced users and system administrators to perform conversions from scripts, batch files, or scheduled tasks without opening the program’s interface.

PORTABLE VERSION

Balabolka comes in two versions: a regular installer and a portable archive. The portable version runs directly from a folder or a USB drive without installing anything on the computer. Users who use shared or restricted computers, or who wish to run the program on more than one computer from one drive, can use the portable version without needing administrator access to install software.

INTERFACE AND USABILITY

The interface has a classic Windows desktop layout with a menu bar, toolbar, and main editing area. Users who have worked with older versions of Windows productivity software will find the layout familiar. The program has customizable skins, font size and color, and background color for the reading area.

The number of options and settings that Balabolka exposes is large. The menus include voice selection, speech parameters, file handling, substitution rules, batch processing, clipboard behavior, bookmarks, hotkeys, and audio output settings. New users can be overwhelmed by the range of options. The basic functionality — open a file, press play, save audio — is pretty simple, but it takes time to dig into the advanced functionality.

Global hotkeys allow users to control playback, skip forward, pause or perform other actions from anywhere on the desktop without bringing the Balabolka window to the front. System tray integration allows the program to be kept accessible without occupying taskbar space.

ACCESSIBILITY USES

Text-to-speech is of practical value to people with visual impairments, dyslexia, reading difficulties, or fatigue from long reading sessions. Balabolka addresses these uses directly with its clipboard watch mode, which reads whatever the user copies from any application — web pages, emails, documents — without requiring the user to open files manually in the program.

Students and researchers use the program to listen to long documents while they take notes or commute. Writers use it to hear their own text read back to them, which catches awkward phrasing and errors that the eye tends to skip over when reading silently. Language learners use it to hear the correct pronunciation of foreign language text.

LIMITATIONS

Balabolka does not have speech recognition. It is not speech to text but text to speech. Users who want to transcribe a spoken audio into written text require a different program.

The program is Windows-only. No version is available for macOS, Linux, Android or iOS. Audio files generated by Balabolka play on any device, but the program itself only works on Windows.

Output quality is all down to the installed voices. The default voices that come with Windows are quite synthetic sounding compared to modern neural voices or cloud-based services. Users who want more natural output have to install more voices separately and good third-party voices are not free.

SYSTEM REQUIREMENTS

Operating System: Windows XP, Vista, 7, 8, 8.1, 10 or 11 (32-bit or 64-bit)
Speech Engine: At least one SAPI 4 or SAPI 5 voice installed on the system
Storage: Less than 10 MB for the program itself; extra space for installed voices and audio output files
Internet connection: Not required