Nuance Vocalizer Expressive TTS 5.4

Nuance Vocalizer Expressive TTS 5.4

Website
vocalizer expressive

System Requirements
OS: Windows Vista - Windows 11 (Vocalizer Expressive 5.2.3 does support Windows XP)
Reading application: any application with SAPI-5 support

Description
Vocalizer Expressive is a text-to-speech engine used by Reading applications to produce speech from electronic text. The voices sound very close to the natural human speech. Some of them, like Spanish Jorge, were first developed by Loquendo and then improved by Nuance.

Incompatibility
The Voices from this release may not function properly if installed on the same computer with the following products: Vocalizer Expressive 5.2.3 (Previous release), kobaspeech.com, harposoftware.com. It's recommended to remove these products before installing Vocalizer Expressive 5.4.

Installation
To install the voices, extract the archive(s) to any folder and run Setup.exe
After installation is complete, Windows Installer will delete files in unpacked folder because it assumes them been temporary. So, unpacked folder will not have any further use and it must be deleted.
The list of voices can be seen in the Speech Properties, which is located under Control Panel, Speech Recognition, Text to Speech.

Uninstallation
To uninstall particular voice go to Start Menu, All Programs, VE 5.4 [Name of Voice], Uninstall.

Download from Google Drive

American English - Allison, Ava, Evan, Joelle, Nathan, Noelle, Samantha, Susan, Tom, Zoe
Arabic - Laila, Maged, Mariam, Tarik
Argentinian Spanish - Diego, Isabela
Australian English - Karen, Lee
Basque - Miren
Belgian Dutch - Ellen
Bhojpuri India - Jaya
Brazilian Portuguese - Felipe, Fernanda, Luciana
British English - Daniel, Kate, Malcolm, Oliver, Serena, Stephanie
Bulgarian - Daria
Canadian French - Amelie, Chantal, Nicolas
Catalan - Jordi, Montserrat
Chilean Spanish - Francisca
Chinese Mandarin - Bin-Bin, Li-Li, Tian-Tian
Colombian Spanish - Carlos, Soledad, Ximena
Croatian - Lana
Czech - Iveta, Zuzana
Danish - Magnus, Sara
Dutch - Claire, Xander
Farsi - Dariush
Finnish - Onni, Satu
French - Audrey, Aurelie, Thomas
Galician - Carmela
German - Anna, Markus, Petra, Viktor, Yannick
Greek - Melina, Nikos
Hebrew - Carmit
Hindi - Lekha, Neel
Hong Kong Cantonese - Sin-Ji
Hungarian - Mariska
Indian Bengali - Paya
Indian English - Rishi, Sangeeta, Veena
Indonesian - Damayanti
Irish English - Moira
Italian - Alice, Federica, Luca, Paola
Japanese - Ayane, Daisuke, Ichiro, Kyoko, Mizuki, Otoya, Sakura
Kannada - Alpana
Korean - Sora, Yuna
Malay - Amira
Marathi - Ananya
Mexican Spanish - Angelica, Juan, Paulina
Norwegian - Henrik, Nora
Polish - Ewa, Krzysztof, Zosia
Portuguese - Catarina, Joana, Joaquim
Romanian - Ioana
Russian - Katya, Milena, Yuri
Scottish English - Fiona
Sichuanese - Fangfang
Slovak - Laura
South African English - Tessa
Spanish - Jorge, Marisol, Monica
Swedish - Alva, Klara, Oskar
Taiwanese Mandarin - Mei-Jia
Tamil - Vani
Telugu - Geeta
Thai - Kanya, Narisa
Turkish - Cem, Yelda
Ukrainian - Lesya
Valencian - Empar
Vietnamese - Linh

Download from Yandex Disk

American English - Allison, Ava, Evan, Joelle, Nathan, Noelle, Samantha, Susan, Tom, Zoe
Arabic - Laila, Maged, Mariam, Tarik
Argentinian Spanish - Diego, Isabela
Australian English - Karen, Lee
Basque - Miren
Belgian Dutch - Ellen
Bhojpuri India - Jaya
Brazilian Portuguese - Felipe, Fernanda, Luciana
British English - Daniel, Kate, Malcolm, Oliver, Serena, Stephanie
Bulgarian - Daria
Canadian French - Amelie, Chantal, Nicolas
Catalan - Jordi, Montserrat
Chilean Spanish - Francisca
Chinese Mandarin - Bin-Bin, Li-Li, Tian-Tian
Colombian Spanish - Carlos, Soledad, Ximena
Croatian - Lana
Czech - Iveta, Zuzana
Danish - Magnus, Sara
Dutch - Claire, Xander
Farsi - Dariush
Finnish - Onni, Satu
French - Audrey, Aurelie, Thomas
Galician - Carmela
German - Anna, Markus, Petra, Viktor, Yannick
Greek - Melina, Nikos
Hebrew - Carmit
Hindi - Lekha, Neel
Hong Kong Cantonese - Sin-Ji
Hungarian - Mariska
Indian Bengali - Paya
Indian English - Rishi, Sangeeta, Veena
Indonesian - Damayanti
Irish English - Moira
Italian - Alice, Federica, Luca, Paola
Japanese - Ayane, Daisuke, Ichiro, Kyoko, Mizuki, Otoya, Sakura
Kannada - Alpana
Korean - Sora, Yuna
Malay - Amira
Marathi - Ananya
Mexican Spanish - Angelica, Juan, Paulina
Norwegian - Henrik, Nora
Polish - Ewa, Krzysztof, Zosia
Portuguese - Catarina, Joana, Joaquim
Romanian - Ioana
Russian - Katya, Milena, Yuri
Scottish English - Fiona
Sichuanese - Fangfang
Slovak - Laura
South African English - Tessa
Spanish - Jorge, Marisol, Monica
Swedish - Alva, Klara, Oskar
Taiwanese Mandarin - Mei-Jia
Tamil - Vani
Telugu - Geeta
Thai - Kanya, Narisa
Turkish - Cem, Yelda
Ukrainian - Lesya
Valencian - Empar
Vietnamese - Linh

Additional information
On some websites "Vocalizer Expressive TTS 5.4" can be called as "Vocalizer Expressive Version 2", while "Vocalizer Expressive 5.2.3" can be called as "Vocalizer Expressive Version 1". These are essentially the same products, only names are different. All voices have the quality "Premium High", except Bin-Bin, Li-Li, Sin-Ji, Tian-Tian, Fang-Fang which have the quality "Embedded". Here is small file TTSApp.exe, which can be used to check out installed voices. Inside this program the setting "Format" must be set to "22kHz 16 Bit Stereo" for better audio output.

Free software to read text aloud
Claude
NaturalReader
Balabolka
Zabaware TTS Reader
KMPlayer - KMPlayer is a video player with ability to speak subtitles of movies. However, the recent versions of this program have a bug: subtitles are repeated several times. The last version that speak subtitles properly is KMPlayer v3.8.0.120
Daum PotPlayer - PotPlayer is another video player which can speak subtitles of movies. In contrast to KMPlayer the subtitles are pronounced correctly

Screenshots



Edit Pronunciation
The pronunciation of the Voices can be modified by creating "User dictionary". It's binary (unreadable for human) file with name "userdct_xxx.dat", where xxx is abbreviation of a language. Where to get this abbreviation? This abbreviation can be found inside of HDR file of the Voice. The *.HDR files are located inside the "ve" folder of Vocalizer 5.4 installation path. For example, C:\Program Files\VE\languages\enu\speech\ve\ve_pipeline_enu_allison_22_premium-high_1-0-2.hdr. The content of HDR file looks like XML:

<?xml version="1.0"?>
<NUANCE>
<VERSION>NUAN_1.0</VERSION>
<HEADER>
  <BROKERSTRING>pipeline/American English/allison/22/premium-high/text/pcm</BROKERSTRING>
  <PRIORITY>5343</PRIORITY>
  <PARAMETERS>
    <language>American English</language>
    <langcode>ENU</langcode>
    <langid>10</langid>
    <langversion>6.0.9</langversion>
    <nativetypeofchar>utf-16</nativetypeofchar>
    <voice>Allison</voice>

In the above example the abbreviation of language is langcode ENU.
At a present time the file "userdct_xxx.dat" can be created in two ways: with Python and with "Vocalizer Expressive Studio". When using Python, it's necessary to know phonetic transcriptions of words, while adding them to dictionary. In contrast, Vocalizer Expressive Studio automatically generates phonetic transcriptions. However, Vocalizer Expressive Studio works only with "Vocalizer Expressive 5.2.3". Created dictionaries can be used in older and newer versions of Vocalizer (5.2.3 and 5.4). Below I will describe both methods.

Creating user dictionary with Python 2.5.4 and Vocon 3200

This method doesn't require "Vocalizer Expressive" been installed. The voices may not be present on a PC and still, it's possible to create user dictionary for them.
As an example, I will create user dictionary for the Finnish language. In this case the file name of user dictionary would be "userdct_fif.dat". First, create usual text file with similar content:

[Header]
Name=userdct_fif.dct
Description=userdct_fif
Language=FIF
Content=EDCT_CONTENT_BROAD_NARROWS
Representation=EDCT_REPR_SZZ_STRING

[Data]
tuhatvuotta // #te.x:At.vu.ottA#
ikuisuus // #@:nKu&iso:S#

This file must be encoded in UTF-8 without BOM. Here, the Name is the name of text file; Description can be any description; Language is uppercase of abbreviation; tuhatvuotta is the source string with 40 characters maximum length; space plus // plus space is separator; #te.x:At.vu.ottA# is a destination string – it's how the word should be pronounced. Note, the # sign here is the pause (silence). Destination string can only contain phonetic symbols in so called L&H+ (Lernout & Hauspie) notation.
All necessary symbols can be found in four tables of this PDF document. These tables contain all Finnish symbols of L&H+. First is called "Special characters" on page 13, another is "Punctuation Marks" on page 14, third is "The Finnish L&H+ and UNIPA Phonetic Alphabets" on pages 15-16 and the fourth is "SAPI5 Symbols" on pages 18-19. From tables 3 and 4 you'll need only first column.
You can add as many words as you like, but they must be separated by newline (CR+LF). After creating text file, you can save it as userdct_fif.dct.
Now, download python-2.5.4 and VoCon 3200. Install them to some path without spaces such as C:\Python25 and C:\PF\Nuance\vocon3200\EDS3. Create *.bat file with content:

@set in=C:\userdct_fif.dct
@set out=C:\userdct_fif.dat
@set VOCON3200_INSTALL_PATH=C:\PF\Nuance\vocon3200\EDS3
@set MODELLANG=fif
@C:\Python25\python.exe "%VOCON3200_INSTALL_PATH%\tools\dictcpl.pyc" --dictionaryFilepath=%in% --outputFilepath=%out%

Here, C:\userdct_fif.dct is a path to text file, created earlier and C:\userdct_fif.dat is path for output binary file. When you run bat file Python will generate binary dictionary "userdct_fif.dat"

Creating user dictionary with Vocalizer Expressive Studio 1.4.0

This method requires installation of Vocalizer Expressive 5.2.3 voices, for which you intend to create a dictionaries. These dictionaries can also be used by Vocalizer 5.4 voices of corresponding language. If you are going to use this method and you also have Vocalizer 5.4 installed, I recommend to uninstall Vocalizer 5.4 first, or use this method on another PC or in virtual machine.
As an example, I will create user dictionary for the German language. In this case the file name of user dictionary would be "userdct_ged.dat". Download archive VEX_GED_Markus.zip from the link above, unpack archive to any folder and run Setup.exe. It will install the voice "Markus". Download and install Java 32-bit if it is not already installed. To check if you have Java, open CMD and type in "java -version". Download archive VEStudio.zip and unpack it to any folder. Inside unpacked folder open "studio" folder and run "VocalizerExpressiveStudio.exe". After launching Studio go to the tab "TTS" and click button "Setup". Select folder "C:\Program Files (x86)\VEX". It's location, where Vocalizer 5.2.3 installed. Next, you'll be asked to select language and voice:

Click OK. After selecting language and voice the tab "TTS" will look like this:

Now go to the tab "User Dictionary Editor" and press "New" button in upper-right corner. Click "Add" button to add new words. Inside textbox "Sounds Like:" enter text - modified pronunciation. Click button "Generate Transcription". It will produce phonetics in L&H+ format. Click button "Copy To Dictionary Entry":

After dictionary is full, press "Save" button in upper-right corner:

Select option to save dictionary as binary DCB file. Rename DCB file into "userdct_ged.dat" (also change extension).

Now, we have the file "userdct_xxx.dat", created by Python or by Studio. It's time to install this dictionary. Installation is just copying this file into the "components" folder of the installed Voice. For example, in case of Finnish voice, copy "userdct_fif.dat" to the C:\Program Files\VE\languages\fif\speech\components. In case of German voice, copy "userdct_ged.dat" to the C:\Program Files\VE\languages\ged\speech\components. Now, if you open TTS application, the pronunciation of words, defined in dictionary will change.

The file userdct_fif.dat is a language specific dictionary – it can be used by either voice of the same language. If you want it to be used only by particular voice, such as Satu, you must rename it (or create new) to userdct_fif_satu.dat.
For the particular voice you can also add another dictionaries – these must be referenced in *.HDR file of the Voice. If you've created new dictionary named userdct_fif_satu1.dat, then you must add the following code to the end of *.HDR file:

<RESOURCE content-type="application/edct-bin-dictionary;loader=broker">
  userdct/fif/satu1
</RESOURCE>

Here, the slash corresponds to the underscore in the filename. So, the end of *.HDR should look like this:

<RESOURCES>
<!-- The next entries specify default broker strings for the user dictionary resources, once language specific and once voice specific -->
<RESOURCE content-type="application/edct-bin-dictionary;loader=broker">
  userdct/fif
</RESOURCE>
<RESOURCE content-type="application/edct-bin-dictionary;loader=broker">
  userdct/fif/satu
</RESOURCE>
<RESOURCE content-type="application/edct-bin-dictionary;loader=broker">
  userdct/fif/satu1
</RESOURCE>
<!-- The next entries specify default broker strings tuning resources if available -->
</RESOURCES>

All user dictionaries must be located inside "components" folder of the Voice/Language.

Vocalizer Expressive 5.2.3 (Previous Release)
Here are 87 voices from previous version of Vocalizer - Vocalizer 5.2.3. All these voices, including Sin-Ji, Mei-Jia and Tian-Tian have the Premium High quality and can be installed on Windows XP. However, this release is not compatible with Vocalizer 5.4. Therefore, before installing Vocalizer 5.2.3, the Vocalizer 5.4 must be uninstalled if it exist. Similarly, before installing Vocalizer 5.4, the Vocalizer 5.2.3 must be uninstalled.

Download from Google Drive

American English - Allison, Ava, Samantha, Susan, Tom
Arabic - Laila, Tarik
Argentinian Spanish - Diego
Australian English - Karen, Lee
Basque - Miren
Belgian Dutch - Ellen
Brazilian Portuguese - Felipe, Luciana
British English - Daniel, Kate, Oliver, Serena
Bulgarian - Daria
Canadian French - Amelie, Chantal, Nicolas
Catalan - Jordi, Montserrat
Chinese Mandarin - Tian-Tian
Colombian Spanish - Carlos, Soledad
Czech - Iveta, Zuzana
Danish - Magnus, Sara
Dutch - Claire, Xander
Finnish - Onni, Satu
French - Audrey, Aurelie, Thomas
Galician - Carmela
German - Anna, Markus, Petra, Yannick
Greek - Melina, Nikos
Hebrew - Carmit
Hindi - Lekha
Hong Kong Cantonese - Sin-Ji
Hungarian - Mariska
Indian English - Veena
Indonesian - Damayanti
Irish English - Moira
Italian - Alice, Federica, Luca, Paola
Japanese - Kyoko, Otoya
Korean - Sora
Mexican Spanish - Angelica, Juan, Paulina
Norwegian - Henrik, Nora
Polish - Ewa, Krzysztof, Zosia
Portuguese - Catarina, Joana, Joaquim
Romanian - Ioana
Russian - Katya, Milena, Yuri
Scottish English - Fiona
Slovak - Laura
South African English - Tessa
Spanish - Jorge, Monica
Swedish - Alva, Klara, Oskar
Taiwanese Mandarin - Mei-Jia
Thai - Kanya
Turkish - Cem, Yelda
Valencian - Empar

Claude v1.0
Warning: some Characters contain nudity
Supported OS: Windows Xp - Windows 11 64bit
Other requirements: None
Installation: run ClaudeInstaller.exe as Administrator
Description: Claude is Text-To-Speech program with built-in support of MSAgent and SSML 1.0. The interface is intuitive and simple. It doesn't have many settings because its purpose is a preservation of MS Agent technology which is only present in older OS'es preceding Windows 7. Claude automatically fixes the absence of MSAgent in recent versions of Windows and puts 40 Characters to their default location.
This program can open, edit and speak the following formats: HTML, RTF, DOC, DOCX, TXT, XML. It can operate in two modes - with MSAgent enabled or disabled.
To switch into second mode uncheck "Use MSAgent".
To install your own MSAgent Characters copy *.acs files to folder "C:\Windows\MSAgent\chars".
Additional notes: HTML files, containing Unicode, must have tag <meta http-equiv="content-type" content="text/html; charset=UTF-8"> inside their <head> tag in order to correctly render in Claude.
Screenshots:



Download: ClaudeInstaller.exe 125Mb (MD5: 963ff3692804ceb0cb155ce8c49fbe36)

×