Fara í leit

Tæknilegar upplýsingar um Ivona talgervla

Byggt á gæðum

Talgervlarnir frá Ivona hafa unnið til margra alþjóðlegra verðlauna og er Ivona fyrirtækið í dag viðurkennt sem leiðandi framleiðandi á þessum markaði þegar kemur að því að bjóða upp á nákvæmasta og náttúrulegasta upplesturinn og bestu hlustunargæðin.

Verðlaunin sem Ivona talgervlarnir hafa fengið eru:

  • 2011 IVONA – ASR Study: the most accurate commercial TTS engine in the world
  • 3 times in a row: Nr. 1 quality TTS at Blizzard Challenge.
  • 2009 Edinburgh,
  • 2007 Bonn,
  • 2006 Pittsburgh,


Suppocrted OSs

  • Linux
  • Windows
  • Mac OS X
  • Solaris
  • Android
  • FreeBSD, OpenBSD
  • Windows Mobile, Windows CE
  • iOS
  • MeeGo, ST Linux Windows Phone 7*

Standards compliance

  • SSML, PLS
  • SAPI
  • MRCP, MRCP v2
  • IPA, X-SAMPA, Navteq™, TeleAtlas Lip-sync / Viseme

Supported HW platforms

  • x86 (32/64 bit)
  • ARM 7, 8, 9, 11
  • Strong ARM
  • X-Scale
  • Sparc (32/64 bit)
  • PowerPC*
  • MIPS*


Supported API

  • IVONA C/C++ API
  • IVONA Java API
  • Android API (apk)
  • TCP/IP
  • Unix socket
  • SAPI 4, SAPI 5
  • Web Services (SOAP/REST)

General IVONA TTS architecture



Skýringarmynd fyrir Ivona talgervilinn


Core engine:

Core speech synthesis, speech coding

Languages and models:

Language models
Voice talent specific models
Voice acoustic data

SDK Speach cloud

HW / OS adaptation layer;
Common TTS interfaces:
SSML
PLS
IPA, X-SAMPA, Navteq
• Lip-sync

Platform speach API

Platform level compliance:

  SAPI, OSX Voice Over
  Android Speech
MRCP

System spec

IVONA

BrightVoice

(SDK)

Statistical parametric**

(SDK)

Speech Cloud

Storage

memory

100-250 MB

< 50 MB*

< 20 MB**

< 5MB** 0 MB
Runtime memory 5 – 13 MB < 5MB** 0 MB
CPU Scalable Under 50 MIPS
BrightVoice™ output
Low response time
Sampling rate 8 kHz, 16kHz, 22 kHz (up to 48 kHz)
Audio formats

PCM 16 bit mono, A-law, µ-law, AMR,

mp3, vorbis (ogg)

Prosody control





Þetta vefsvæði byggir á Eplica