Di-phone database for text-to-speech conversion

75 Last view: 2023-10-25

Di-phone database for text-to-speech conversion

Di-phone-hu

ID:

211 The Di-phone set (labelled wave form items) for Hungarian contains combinations of 38 sounds for TTS conversion. Besides the Di-phone set can be used for educational purposes and in speech research.

You don’t have the permission to edit this resource.

DistributionAvailability

Available - Restricted Use

Start date: 15/07/2012

Licence

MS - C - No ReD - FF

Restrictions: Academic - Non Commercial Use

Fee: 1,000 EURO

Download location: hidden

Distribution Access/Medium: Downloadable

Licensors:

Henk Tamás

Distribution rights holders:

Gábor Olaszy

IPR Holder

Gábor Olaszy

Contact Person

Géza Németh

text
audio

Monolingual text corpusLanguages

Hungarian (1,455 Words)

Linguality

Linguality type: Monolingual

Size

1,455 Words

Character encoding

ISO - 8859 - 2 (9,825 Bytes)

Creation

Creation mode: Manual

Original Sources

research

Monolingual audio corpusLanguages

Hungarian (1,646 Seconds)

Linguality

Linguality type: Monolingual

Size

1,646 Seconds

Effective speech duration

1,646 Seconds

Audio duration

1,646 Seconds

AnnotationSegmentation

Tagset: http://praat.org

Annotated elements: Other

Segmentation level: Phoneme

Format: Praat TextGrid

Annotation Mode: Mixed (TTS for the text to sound conversion, and manual alignment of sound boundaries)

Annotation Tools:

Profivox development tool

Start date: 01/01/2007

End date: 30/06/2012

Size: 10,693 Phonemes

Content

Speech items: Isolated Words

Noise Level: Low

Setting

Naturality: Read Speech

Conversational type: Monologue

Scenario: Other

Audience: No

Interactivity: Non Interactive

Audio Formatswave/audio (1,646 Seconds)

Compression: False

Recording quality: High

Quantization: 16

Number of tracks: 1

Sampling rate: 44100

Signal encoding: LinearPCM

RecordingRecorders

Gábor Olaszy

Recording platform software: soundforge_sony

Recording environment: Studio

Recording device type details: RME Fireface800

Recording device type: Other

Capture

Capturing device type details: AKG C-414 B-ULS

Capturing device type: Studio Equipment

Person SourceSet

Origin of persons: Native

Age of persons: Adult

Sex of persons: Female

Number of persons: 1

Dialect accent of persons: no dialect

Age range end: 24

Hearing impairment of persons: No

Number of trained speakers: 0

Age range start: 24

Speaking impairment of persons: No

CreationOriginal Sources

corpora

Resource Creation

Resource Creator

Gábor Olaszy

Creation lasted: 01/01/2002 - 30/06/2012

Metadata

Created: 09/07/2012

Last Updated: 09/07/2012

Source: METANET4U

Revision: 2012-07-09

Metadata Creator

Mátyás Bartalis

Version

Version: 1.0

Revision: Waves, texts, sound boundaries

Last Updated: 09/07/2012

ValidationValidated

Type of Validation: Content

Validation Mode: Manual

Mode Details: Manually checked the sound boundaries in the corpora

Validator

Bálint Tóth

Usage

Foreseen UseNlp Applications

Use NLP Specific: Speech Analysis, Speech Synthesis, Talking Head Synthesis

Human UseActual Use - Nlp Applications

Use NLP Specific: Speech Synthesis, Talking Head Synthesis

Details: The Profivox hungarian di-phone TTS uses a database based on this resource

Usage Report

International Journal of Speech Technology, Kluwer, 2000

Derived Resources

Hungarian di-phone speech synthesizer

Documentation

Samples Location: http://speechlab.tmi...

http://speechlab.tmi...

Olaszy G. - Gépi beszédkeltés információs rendszerekhez Magyarországon. AKUSZTIKAI SZEMLE III:(1-3) pp. 4-13. (1999)

http://speechlab.tmi...

People who looked at this resource also viewed the following:

Resources from the same creators