Balabolka Utilitate pentru servicii online de sinteză a vorbirii

The command line application allows to use online text-to-speech services: text files or subtitles can be converted to audio files. The utility can be used for testing purposes: it will help you to choose a cloud computing service that satisfies your needs. The separate application for Yandex SpeechKit is available for downloading, because Yandex is the Russian IT company with close government ties.

Online services with speech technologies: Google Cloud Text-to-Speech; Amazon Polly; Baidu Text-to-Speech; CereVoice Cloud; Iciba Text-to-Speech; iTranslate Text-to-Speech; Microsoft Azure; Naver Text-to-Speech; Youdao Text-to-Speech; Yandex SpeechKit.

Download Balabolka (Online TTS Utility)

The utility handles various command line parameters to be able to read text aloud or save as an audio file. The command line options use the syntax "bal4web [options ...]", all parameters must be separated by a space. Options can appear in any order on the command line so long as they are paired with their related parameters. Use the "bal4web -?" command line to get help on the command line syntax and parameters.

-s service_name

Sets the name of the online TTS service ("google" or "g", "amazon" or "a", "baidu" or "b", "cerevoice" or "c", "iciba" or "k", "itranslate" or "t", "microsoft" or "m", "naver" or "n", "youdao" or "y"). The default is "google".

-l language_name

Sets the language name for the online TTS service. The name is a combination of an ISO 639 two-letter lowercase culture code associated with a language and an ISO 3166 two-letter uppercase subculture code associated with a country or region.
For example: ro-RO, de-DE, fr-FR. The default is "en-US".

-g gender

Sets the gender for the online TTS service (if supported). The available values: "female" or "f", "male" or "m". The default value is not defined. This parameter is supported by services: Amazon Polly, CereProc TTS, Google TTS, iTranslate TTS, Microsoft Azure, Naver TTS.

-n voice_name

Sets the voice name for the online TTS service (if supported). The default value is not defined. This parameter is supported by services Amazon Polly, CereProc TTS, Google Cloud TTS, Microsoft Azure, Naver TTS.

-r speech_rate

Sets the rate of the synthesized speech (if supported). The default is "1.0" (average human speech rate).
For Amazon Polly the rate values range from "0.20" to "2.00".
For CereProc TTS the rate values range from "0.30" to "4.00".
For Google TTS and Microsoft Azure the rate values range from "0.10" to "3.00".
For iTranslate TTS the rate values range from "0.50" to "2.00".
For Naver TTS the rate values range from "0.50" to "1.50".

-p integer

Sets the speaking pitch in a range of -20 to 20 (if supported). The default is 0.
This option is supported by Amazon Polly, CereProc TTS, Google Cloud TTS, Microsoft Azure.

-v integer

Sets the volume in a range of 0 to 200 (the default is 100).

-m

Prints the list of supported languages (genders and voices' names, if available) for the online TTS service.

-f file_name

Sets the name of the input text file. The command line may contain few options -f.

-fl file_name

Sets the name of the text file with the list of input files (one file name per line). The command line may contain few options -fl.

-w file_name

Sets the name of the output file in WAV format.

-c

Gets the text input from the clipboard.

-t text

Gets the text input from the command line. The command line may contain few options -t.

-i

Gets the text input from STDIN.

-o

Writes sound data to STDOUT; if the option is specified, the option -w is ignored.

--encoding encoding sau -enc encoding

Sets the input text encoding ("ansi", "utf8" or "unicode"). The default is "ansi".

--silence-begin integer sau -sb integer

Sets the length of silence at the beginning of the audio file (in milliseconds). The default is 0.

--silence-end integer sau -se integer

Sets the length of silence at the end of the audio file (in milliseconds). The default is 0.

-ln integer

Selects a line from the text file by using of a line number. The line numbering starts at "1". The interval of numbers can be used for selecting of more than one line (for example, "26-34"). The command line may contain few options -ln.

-e integer

Sets the length of pauses between sentences (in milliseconds). The value should be set less than 5000. If the option is not specified, the service will use the default pauses between sentences. This parameter is supported by Microsoft Azure only.

-d file_name

Applies a dictionary for pronunciation correction (*.BXD, *.DIC or *.REX). The command line may contain few options -d.

-lrc

Creates the LRC file. Lyrics will be synchronized with the speech in the output audio file.

-srt

Creates the SRT file. Subtitles will be synchronized with the speech in the output audio file.

-sub

Input text will be processed as subtitles. The option may be useful, when the options -i or -c are specified.

-host host_name

Sets the hostname of the proxy server.

-port integer

Sets the port number of the proxy server.

-fr integer

Sets the output audio sampling frequency in kHz (8, 11, 16, 22, 24, 32, 44, 48). If the option is not specified, the default value for the selected serice will be used.

--ignore-square-brackets sau -isb

Ignore text in [square brackets].

--ignore-curly-brackets sau -icb

Ignore text in {curly brackets}.

--ignore-angle-brackets sau -iab

Ignore text in <angle brackets>.

--ignore-round-brackets sau -irb

Ignore text in (round brackets).

--ignore-url sau -iu

Ignore URLs.

--ignore-comments sau -ic

Ignore comments in text. Single-line comments start with // and continue until the end of the line. Multiline comments start with /* and end with */.

-dp

Display progress information in a console window.

-cfg file_name

Sets the name of the configuration file with the command line options (a text file where each line contains one option). If the option is not specified, the file bal4web.cfg in the same folder as the utility will be used.

-h

Prints the list of available command line options.

--lrc-length integer

Sets the maximal length of text lines for the LRC file (in characters).

--lrc-fname file_name

Sets the name of the LRC file. The option may be useful, when the option -o is specified.

--lrc-enc encoding

Sets the encoding for the LRC file ("ansi", "utf8" or "unicode"). The default is "ansi".

--lrc-offset integer

Sets the time shift for the LRC file (in milliseconds).

--lrc-artist text

Sets the ID tag for the LRC file: artist.

--lrc-album text

Sets the ID tag for the LRC file: album.

--lrc-title text

Sets the ID tag for the LRC file: title.

--lrc-author text

Sets the ID tag for the LRC file: author.

--lrc-creator text

Sets the ID tag for the LRC file: creator of the LRC file.

--lrc-sent

Inserts blank lines after sentences when creating the LRC file.

--lrc-para

Inserts blank lines after paragraphs when creating the LRC file.

--srt-length integer

Sets the maximal length of text lines for the SRT file (in characters).

--srt-fname file_name

Sets the name of the SRT file. The option may be useful, when the option -o is specified.

--srt-enc encoding

Sets the encoding for the SRT file ("ansi", "utf8" or "unicode"). The default is "ansi".

--raw

Output is raw PCM; audio data does not contain the WAV header. The option is used together with the option -o.

--ignore-length sau -il

Omits the length of data in the WAV header. The option is used together with the option -o.

--wss

Use the WebSocket protocol for Microsoft Azure. It allows to improve sound quality of audio files (24 KHz instead of 16 KHz). The option is ignored if the subscription key for the Microsoft Azure Cognitive Services is defined. Use the option -m to check if a voice supports the WebSocket protocol or not.

--sub-format text

Sets the format of input subtitles ("srt", "lrc", "ssa", "ass", "smi" or "vtt"). If the option is not specified, the format will be determined through the file extension.

--sub-fit sau -sf

Automatically increases the speech rate to fit time intervals (when the program converts subtitles to audio file). The SoundTouch library will be used for changing tempo.

--sub-max integer sau -sm integer

Sets the maximal rate of speech in a range of 110% to 200% (when the program converts subtitles to audio file). The program will automatically increase the speech rate without exceeding the set rate value.

--aws-keyid text sau -ak text

Sets AWS access key ID for the Amazon Polly. It is recommended to apply such key if you have it.

--aws-secret text sau -as text

Sets AWS secret access key for the Amazon Polly.

--aws-region text sau -ar text

Sets AWS region for the Amazon Polly.

--crv-email text sau -ce text

Sets the email address used when registering on the CereProc website. This information is necessary for CereVoice Cloud API authorization. It is recommended to apply such email if you have it.

--crv-pwd text sau -cp text

Sets the password used when registering on the CereProc website. This information is necessary for CereVoice Cloud API authorization. It is recommended to apply such password if you have it.

--gc-apikey text sau -gk text

Sets API key ID for the Google Cloud. It is recommended to apply such key if you have it.

--ms-apikey text sau -mk text

Sets the subscription key for the Microsoft Azure Cognitive Services. It is recommended to apply such key if you have it.

--ms-region text sau -mr text

Sets the subscription region for the Microsoft Azure Cognitive Services.

Create the text file LANGUAGE.TXT with the list of all supported languages and genders for the Google TTS service:

bal4web -s Google -m > language.txt

Convert text from BOOK.TXT to speech and save as BOOK.WAV:

bal4web -f "d:\Text\book.txt" -w "d:\Sound\book.wav" -s Google -l en-US -g female

Convert subtitles to speech and save as MOVIE.WAV:

bal4web -f "d:\Subtitles\movie.srt" -w "d:\Sound\movie.wav" -s m -l de-DE -n Conrad -r 1.1

bal4web -f "d:\Subtitles\movie.srt" -w "d:\Sound\movie.wav" -s m -l de-DE -n Conrad -sub-fit-lib

The example of use together with LAME.EXE:

bal4web -f d:\book.txt -s Baidu -l en-US -o --raw | lame -r -s 16 -m m -h - d:\book.mp3

The example of use together with OGGENC2.EXE:

bal4web -f d:\book.txt -s Baidu -l en-US -o -il | oggenc2 --ignorelength - -o d:\book.ogg

The command line options can be stored as a configuration file "bal4web.cfg" in the same folder as the utility.

The sample configuration file:

The utility may combine options from the configuration file and the command line.

You are free to use and distribute software for noncommercial purposes. For commercial use or distribution, you need to get permission from the copyright holder.