Very Slow
ፍጥነት
Exceptional
ጥራት
አዎ
መተላለፊያ
1
ቋንቋዎች
ስለ ቱርቶይዝ TTS
Tortoise TTS is an autoregressive text-to-speech model that prioritizes audio quality above all else. Using a combination of autoregressive transformers and diffusion models, Tortoise generates extremely natural speech that captures subtle nuances of human voice. While slower than other models, Tortoise produces the most natural-sounding TTS output available.
የቁልፍ ባህሪያት
ከፍተኛ ጥራት
የTTS ውጤት
የድምፅ ቅጂ
ድምጾችን በልዩ ልዩ ግልጽነትና ልዩነት ክሎን አድርግ
የቋንቋ ችሎታ
የንግግር ቅርጾችን እና የጥቂት-አመለካከቶችን ይይዛል
የጥራት ዕቅድ
ከultra_fast እስከ high_quality ማቀናጀት ይመርጡ
የሕሊና ጥልቀት
እውነተኛ የስነልቦና ምላሽ ያለው ንግግር ይፈጥራል
የክፍል ፋይል
አፓቺ 2.0 የኮሜርሺያል ጥቅም መብቶች ጋር ፈቃድ
ጥቅም
የድምፅ መጽሐፍት
የፊልም ምርት
መዝገብ ቤት
የሙያ ድምፅ
መዝገብ ቤት
የከፍተኛ ደረጃ ይዘት
ቱርቶይዝ TTS Voices
View All 18Tortoise Angie
ENTortoise Deniro
ENTortoise Freeman
ENTortoise Geralt
ENTortoise Halle
ENTortoise Jlaw
ENTortoise Lj
ENTortoise Mol
ENTortoise Myself
ENTortoise Pat
ENTortoise Pat2
ENTortoise Snakes
ENብዙ ጊዜ የሚጠየቁ ጥያቄዎች
Tortoise TTS is an autoregressive text-to-speech model created by James Betker that prioritizes audio quality. It uses transformers and diffusion models to generate speech with unmatched naturalness and emotional depth.
Tortoise is open-source under Apache 2.0 license. On TextToSpeechAI, we charge 50 credits per 1000 characters (Ultra tier) due to extensive compute requirements and exceptional output quality.
Tortoise primarily supports English. It was trained on English speech datasets. For multilingual needs with similar quality, consider F5-TTS or use Tortoise in combination with other models.
Tortoise is the slowest TTS model due to its quality-first architecture. Generation can take 30 seconds to several minutes depending on text length and quality preset. Use "fast" preset for reasonable wait times.
Tortoise offers 4 presets: ultra_fast (testing), fast (production default), standard (balanced), and high_quality (maximum quality). Higher quality presets generate multiple candidates and select the best.
የድምፅ ቅጂዎችን ለመክተት ብዙ የድምፅ ቅጂዎችን (በተለይም 3-10 ክሊፖች፣ 5-10 ሰከንዶች) መስጠት አለብዎት። Tortoise እነዚህን የድምፅ ባህሪያትን ለመያዝ፣ የመናገር ንድፎችን እና ጥልቅ ቅጦችን ለመለየት ያጠናክራል።
Tortoise ልዩ የድምፅ ጥራት ያመጣል - በስፋት በጣም ተፈጥሯዊ-ድምፅ ያለው TTS ተደራሽ ነው ተደርጎ ይወሰዳል. እርሱ micro- expressions, breathing patterns, እና ሌሎች ሞዴሎች የሚጠፉት ስሜታዊ nuances ይይዛል.
Tortoise በጥራት እና በሞዴል መጠን ላይ የተመሠረተ የ VRAM 12-24GB ያስፈልጋል. እንደ RTX 3090, 4090, ወይም A100 ያሉ ከፍተኛ-መጨረሻ GPUs መከራከሪያዎች ናቸው. CPU ውጤት ሊሆን ይችላል ግን በጣም ዝቅተኛ ነው.
አዎ፣ Tortoise የአፓቺ 2.0 ፈቃድ ያለው ሲሆን ይህም የኮሜርሺያል ጥቅም በባለቤትነት የሚፈቀድ ነው። ለፕሪሚየም ይዘት ተስማሚ ነው፣ ምክንያቱም ውጤቱ ረጅም የጊዜ ሂደት ይኖረዋል፤
Select a Tortoise voice and optionally specify a quality preset in your API request. Note that generation times are longer than other models. We recommend the "fast" preset for most use cases.
Tortoise outputs high-quality WAV audio at 24kHz. Through TextToSpeechAI, you can request MP3, WAV, or OGG with quality-preserving encoding.
Tortoise produces the highest quality speech but is by far the slowest. Use it when quality is paramount and time is not a constraint. For faster results, StyleTTS 2 offers excellent quality. For real-time needs, use Piper.
Technical Specs
- Generation Speed Very Slow
- Output Quality Exceptional
- Voice Cloning Supported
- Languages 1
- GPU VRAM 12-24GB
- Credits/1000 chars 50