SVOX Corporate SVOX desktop German,British English, Czech, Danish, Dutch, European French, European Portuguese, European Spanish, Italian, Norwegian, Polish, Russian, Swedish, Turkish, American English, Canadian French, Mexican Spanish, Cantonese Chinese, Japanese, Korean and Mandarin Chinese. Nicole, Petra, Marlene, Markus non-uniform unit selection http://ttssamples.syntheticspeech.de/ttsSamples/svox_petra_22kHz_s1.mp3 http://www.svox.com/ 15/1/09 SVOX Mobile (pro), Automotive SVOX mobile see desktop Nicole, Petra, Marlene, Markus non-uniform unit selection 7 MB of RAM and 12 MB of ROM http://ttssamples.syntheticspeech.de/ttsSamples/svox_petra_22kHz_s1.mp3 http://www.svox.com/ 15/1/09 SVOX Pico SVOX embedded see desktop, Nicole, Petra, Marlene, Markus HMM-based non-uniform unit selection, 8, 16 kHz Symbian, WinCE, WM, WMfA, proprietary operating systems engine ~400 kB, ~500 kB per language, ~500 kB per additional voice, 50 MIPS on ARM9 processor load X-SAMPA, NT-SAMPA http://ttssamples.syntheticspeech.de/ttsSamples/svox_petra_22kHz_s1.mp3 http://www.svox.com/ 15/1/09 RealSpeak Telecom Nuance server German, Basque, Catalan, Czech, Dutch, Portugese, Canadian, Mandarin, Spanish, Danish, French, Italian, Japanese, Korean, Latin-American, Norwegian, Polish, Russian, Swedish, Indian, Irish, Scottish, South African, US- and UK-English, Finnish, Greek, Hindi, Turkish Yannick, Monika, Steffi, Vera non-uniform unit selection, 8 kHz windows, linux redhat, solaris 500 GB disk, 1GB RAM Tts-tuning enabled by "PromptSculptor" graphical tool. http://ttssamples.syntheticspeech.de/ttsSamples/realSpeak_yannik1.mp3 http://www.nuance.com/realspeak/ 15/1/09 Reimund Schmald, reimund.schmald@nuance.com RealSpeak Solo Nuance desktop see telecom version (might be less) see telecom version non-uniform unit selection, 11-16 kHz win32, winCE depending on sample rate and reduction: 8-60 MB disk space and 4-8 MB RAM 4.0 To speech-enable automotive, desktop and mobile applications. http://www.nuance.com/realspeak/ 15/1/09 see telecom version RealSpeak Mobile Nuance embedded see telecom version (might be less) see telecom version non-uniform unit selection RealSpeak Mobile SF at around 2 MB total footprint and RealSpeak Mobile HQ at around 5 MB total footprint (between 400 KB and 1.5 MB RAM). A scalable solution for high quality SMS and email reading on mobile handsets. http://www.nuance.com/realspeak/ 15/1/09 see telecom version Loquendo TTS telephony Loquendo server German, American English, British English, Castilian Spanish, Catalan, Valencian, Galician, French, Italian, Greek, Portuguese, Swedish, Dutch, Polish, Brazilian Portuguese, Mandarin Chinese, Mexican Spanish, Chilean, Argentinean, American Spanish, Canadian French, Turkish, Finnish, Russian, Danish (2008) Stefan, Katrin, Ulrike non-uniform unit selection, 8 kHz Windows 2000, 2003 Server, XP, Vista; Linux Red Hat 7.3, 9.0; Red Hat Enterprise 3; Suse 7.3, 8.1 10 MB RAM for the engine, 80 MB per voice 3 MB per channel C/C++ Loquendo API SAPI 4 and 5 (Win 32) W3C SSML 1.0 for enhancing TTS rendering Loquendo TTS ActiveX (Win 32) Microsoft .net* Announced features:Loquendo TTS director for prompt customization, Expressive TTS, Mixed Language Capability: Phonetic Mapping, Language guesser, Dynamic switching between multiple voices, Flexible voice control - for creating special effects, modifying speech rate, pitch and timbre SMS and E-mail Preprocessing Customized voices - for extending corporate image and branding through unique voices (TeleAtlas® and Navteq™ extensions included) http://ttssamples.syntheticspeech.de/ttsSamples/loquendo_stefan_s1.mp3 http://www.loquendo.com/de/ 15/1/09 Klaus Dietz, Inloq Voice Solutions, klaus.dietz@inloq.de Loquendo TTS multimedia Loquendo desktop see server version Stefan, Katrin, Ulrike non-uniform unit selection Windows 2000, 2003 Server, XP, Vista; Linux Red Hat 7.3, 9.0; Red Hat Enterprise 3; Suse 7.3, 8.1 Depends on 8, 11, 16, 22, 32 or 44 kHz sampling rate, various compression bit rates and different sized databases are available C/C++ Loquendo API SAPI 4 and 5 (Win 32) W3C SSML 1.0 for enhancing TTS rendering Loquendo TTS ActiveX (Win 32) Microsoft .net* http://www.loquendo.com/de/ 15/1/09 see server version Loquendo TTS automotive Loquendo pc see server version Stefan, Katrin, Ulrike non-uniform unit selection Windows Mobile 6.0 (all editions), Windows Mobile 5.0 (all editions), Pocket PC 2003, CE 5.0, CE.NET 4.2, Windows XP Embedded and TabletPC ed., VXWorks, Linux, Symbian OS S60 (2nd/3rd edition). Depends on 8, 11, 16, 22, 32 or 44 kHz sampling rate, various compression bit rates and different sized databases are available Loquendo API, JSAPI and Microsoft SAPI, SSML, IPA and SAMPA. special dictionaries for automative usage, e.g. navigation, SMS reading etc. http://www.loquendo.com/de/ 15/1/09 see server version Loquendo TTS embedded Loquendo embedded see server version Stefan, Katrin, Ulrike non-uniform unit selection, 8/11/16/22/32/44 KHz Windows Mobile 6.0 (all editions), Windows Mobile 5.0 (all editions), Pocket PC 2003, CE 5.0, CE.NET 4.2, Windows XP Embedded and TabletPC ed., VXWorks, Linux, Symbian OS S60 (2nd/3rd edition). from 2.5 MB RAM from 3.5 MB storage per voice http://www.loquendo.com/de/ 15/1/09 see server version Natural Voices Server edition ATT server German, US-English, UK-English, Latin American Spanish, French, French Canadian Reiner, Klara non-uniform unit selection Windows NT, 2000, XP Linux GCC 2.91, 2.96, 3.2, 3.32, 3.33, 3.44 SAPI 4, 5, 5.1, SSML, JSML, JSAPI, C++ SDK, SNMP 1.4 http://ttssamples.syntheticspeech.de/ttsSamples/att_reiner_s1.mp3 http://www.naturalvoices.att.com/ 15/1/09 http://wizzardsoftware.com Natural Voices desktop edition ATT pc see server version see server version see server version Windows NT, 2000, XP Windows 98, ME Linux GCC 2.91, 2.96, 3.2, 3.32, 3.33, 3.44 SAPI 4, 5, 5.1, SSML, JSML, JSAPI, C++ SDK, 1.4 http://ttssamples.syntheticspeech.de/ttsSamples/att_reiner_s1.mp3 http://www.naturalvoices.att.com/ 15/1/09 http://wizzardsoftware.com Mbrola Univ. of Mons pc German, Afrikaans, American, Arabic, Brazilian, Breton, British, Canadian, Croation, Czech, Dutch, Estonian, European, French, German-Bavarian, Greek, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Iranian, Italian, Japanese, Korean, Lithuanian, Malay, Maori, Polish, Romanian, Small, Spanish, Swedish, Telugu, Turkish, Venezuelan 10 voices diphone concatenation PC/Windows, PC/CYGWIN, PC/DOS, MAC OS 9, MAC OS 9 SharedLib!, MacOSX, LINUX i386 / ppc / alpha / ultra1, LINUX / Pocket PC, LINUX / Ircha / Mbrola / ES1 DBA / Z80 , NetBSD (i386 and sparc) + OpenBSD (i386), FreeBSD (i386), QNX Os (i386) , SCO OpenServer 32v5, BeBox, BeOs(x86/ppc), OS/2, PC/Solaris2.4, SUN/SunOS5.4, SUN/SunOS4.xx, HPUX9.0 and HPUX10.0, VAX/VMS V6.2, DECALPHA(AXP)/VMS 6.2, NEXT/NEXTSTEP (Intel and Motorola), IBM RS6000 AIX 4.1, SGI INDY/IRIX 5.3 and 6.2, Symbian | Symbian OS 7, AMD64/Linux about 10 MB RAM, 20 MB disk only signal generation, free natural language processing tools available http://ttssamples.syntheticspeech.de/ttsSamples/de1_s1.mp3 http://tcts.fpms.ac.be/synthesis/mbrola.html 15/1/09 http://www.acapela-group.com/ IBM WebSphere CTTS IBM server German, US English, UK English, French, Canadian French, Brazilian Portuguese, Cantonese (Hong Kong), Norwegian, Swedish, Italian, Japanese, Simplified Chinese (Mandarin/China), Traditional Chinese(Mandarin/Taiwan), Danish, Finnish, Mexican Spanish, Castilian Spanish male, female non-uniform unit selection 512 MB Ram, 2G disk VoiceXML 2.1, MRCP WebSphere Voice Server 4.2 http://ttssamples.syntheticspeech.de/ttsSamples/ctts_female_8kHz.mp3 http://www-01.ibm.com/software/pervasive/voice_server 15/1/09 part of Web Sphere Voice Server (2006) IBM Concatenative IBM desktop see server version see server version non-uniform unit selection win 98, 2000, ME, XP, NT 4.0, Linux 48MB RAM http://ttssamples.syntheticspeech.de/ttsSamples/ctts_female_8kHz.mp3 http://www-01.ibm.com/software/pervasive/tech/demos/tts.shtml 15/1/09 http://wizzardsoftware.com IBM Formant IBM desktop Formant formerly viaVoice outloud http://ttssamples.syntheticspeech.de/ttsSamples/eloqdemo.mp3 http://www-01.ibm.com/software/pervasive/tech/demos/tts.shtml 15/1/09 http://wizzardsoftware.com IBM embedded ViaVoice IBM embedded see server version see server version non-uniform unit selection, 11/16/22kHz sampling rates Windows XP, Windows 2000, Windows CE / Windows Mobile, QNX, Linux, Embedded Linux, T-Engine, MicroItron, VxWorks, RTXC The modular Embedded ViaVoice architecture provides fully integrated, automatic speech recognition, speech synthesis through text-to-speech (TTS) and other technology engines supporting the full-feature requirements of an application with minimal processor utilization and memory requirements. http://ttssamples.syntheticspeech.de/ttsSamples/ctts_female_8kHz.mp3 http://www-01.ibm.com/software/pervasive/embedded_viavoice/ 15/1/09 Acapela Telecom Acapela server German, Arabic, Czech, Danish, Dutch (B), Dutch (NL), English (UK), English (US), Finnish, French, French (Canadian), Greek, Icelandic, Italian, Norwegian, Polish, Portuguese, Portuguese (Brazil), Spanish, Spanish (North America), Russian, Swedish, Turkish Sarah, Klaus non-uniform unit selection, 8kHz NT 4SP6, Windows 2000SP3, Windows XP. Linux RedHat ES 3.0., 4.0, 5.0 A single processor PIV 2.4 Ghz is able to run up to 50 simultaneous TTS sessions with any combination of languages and voices MRCP, proprietary API, Microsoft SAPI 4, Microsoft SAPI 5.1, Nuance 8.5 and NVP, API, Native Java Client supporting JSAPI/JSML, Acapela Voice Factory offers companies to achieve voice branding http://ttssamples.syntheticspeech.de/ttsSamples/brightSpeech_klaus_16kHz_s1.mp3 http://www.acapela-group.com/ 15/1/09 antoine.kauffeisen@acapela-group.com Acapela Virtual Speaker and Multimedia Acapela desktop see server version Gerhard, Helga non-uniform unit selection (HQ), diphon verkettung (MBrola) (HD), 8kHz, 11kHz, 16 kHz, 22kHz, - PCM, A-law, μ-Law, vox, mp3. Windows 2000, XP, Vista, Mac depends on HQ (high quality) tts or HD (high density) Windows Edition: SAPI 4.0a, SAPI 5.1b / 5.3, ActiveX and proprietary APIs (BabTTS API and NSC API), Mac OSX Edition: Apple Speech Synthesis Manager http://www.acapela-group.com/ 15/1/09 see server version Acapela Mobility Acapela embedded see server version Gerhard, Helga 2 version: non-uniform unit selection (HQ), OR diphon concatenation (MBrola) (HD),8, 11,16, 22 kHz PPC2002 , Windows Mobile 2003, WM5, WM6.0, Linux, Symbian Series 60, Palm OS. disk 600 Kb to 6 Mb, RAM : from 112 Kb to 2 Mb. C, C++ , VB through Active X , C#. For Automotive: Other CPUs / OS / voice formats are available via Acapela OnBoard http://ttssamples.syntheticspeech.de/ttsSamples/babil_gerhard_s1.mp3 http://www.acapela-group.com/ 15/1/09 see server version Dress PC VoiceINTERConnect desktop German, Chinese Mandarin, Italian, US-English, UK-English, Russian, Czech male, female diphone concatenation Windows, linux, Unix, Mac RAM/disk less than 10 MB http://ttssamples.syntheticspeech.de/ttsSamples/voiceInterConnectDress_m_s1.mp3 http://www.voiceinterconnect.de 15/1/09 http://www.voiceinterconnect.de Dress mobile MDRESS VoiceINTERConnect embedded see pc version see pc version diphone concatenation ADSP 21065, 219x, Blackfin, TI TMS3, Ressourcenbedarf 205x, TMS3206x, Motorola, ARM less than 1 MB platform hardware module http://ttssamples.syntheticspeech.de/ttsSamples/voiceInterConnectDress_m_s1.mp3 http://www.voiceinterconnect.de 15/1/09 http://www.voiceinterconnect.de