SHACHI - Language Resource Metadata Database

Desktop/Microphone
The Persian Speech Database Farsdat comprises the recordings of 300 Iranian speakers, who differ from each other with regards to age, sex, education level, and dialect (10 dialect regions of Iran were represented: Tehrani, Torki, Esfahani, Jonubi, Shomali, Khorassani, Baluchi, Kordi, Lori, and Yazdi). Each speaker uttered 20 sentences in two sessions, and 100 of these speakers uttered 110 isolated words. 6000 utterances were segmented and labelled phonetically and phonemically manually, including 386 phonetically balanced sentences, using IPA characters. The acoustic signal has been stored with a Wave file standard, so that it can be used by any other application software. The used sampling frequency reaches 22.5 KHz, and the signal-to-noise ratio 34 dB. The ambiguities in segmentation have been solved by reference to the corresponding spectrograms extracted from DSP sona-Graph KAY 5500.

Telephone
The Finnish SpeechDat(II) FDB-4000 comprises 4000 Finnish speakers (1830 males, 2170 females) recorded over the Finnish fixed telephone network. The FDB-4000 database is partitioned into 14 CDs, 13 CDs comprise 300 speakers sessions, the 14th comprises 100 speakers. The speech databases made within the SpeechDat(II) project were validated by SPEX, the Netherlands, to assess their compliance with the SpeechDat format and content specifications.
Speech samples are stored as sequences of 8-bit 8 kHz A-law. Each prompted utterance is stored in a separate file. Each signal file is accompanied by an ASCII SAM label file which contains the relevant descriptive information.
Each speaker uttered the following items:
? 1 isolated digit
? 1 sequence of 10 isolated digits
? 4 numbers: 1 sheet number (5 digits), 1 telephone number (9-10 digits), 1 credit card number (16 digits), 1 PIN code (6 digits)
? 1 currency money amount
? 1 natural number
? 3 dates: 1 spontaneous date (birthdate), 1 prompted date, 1 relative or general date expression
? 2 time phrases: 1 time of day (spontaneous), 1 time phrase
? 3 spelled words: 1 spontaneous own forename, 1 city name, 1 phonetically rich word
? 5 directory assistance names: 1 spontaneous own forename, 1 spontaneous city of growing up, 1 frequent city name, 1 frequent company name, 1 common forename surname
? 2 yes/no questions: 1 predominantly ?yes? question, 1 predominantly ?no? question
? 3 application words
? 1 word spotting phrase using an embedded application word
? 4 phonetically rich words
? 9 phonetically rich sentences
The following age distribution has been obtained: 545 speakers are below 16 years old, 1773 speakers are between 16 and 30, 980 speakers are between 31 and 45, 606 speakers are between 46 and 60, and 96 speakers are over 60.
A pronunciation lexicon with a phonemic transcription in SAMPA is also included.

Desktop/Microphone
The Finnish Speecon database is divided into 2 sets:
1) The first set comprises the recordings of 550 adult Finnish speakers (273 males, 277 females), recorded over 4 microphone channels in 4 recording environments (office, entertainment, car, public place).
2) The second set comprises the recordings of 50 child Finnish speakers (25 boys, 25 girls), recorded over 4 microphone channels in 1 recording environment (children room).

This database is partitioned into 22 DVDs (first set) and 3 DVDs (second set).

The speech databases made within the Speecon project were validated by SPEX, the Netherlands, to assess their compliance with the Speecon format and content specifications.
Each of the four speech channels is recorded at 16 kHz, 16 bit, uncompressed unsigned integers in Intel format (lo-hi byte order). To each signal file corresponds an ASCII SAM label file which contains the relevant descriptive information.

Each speaker uttered the following items:
Calibration data:
6 noise recordings
The “silence word” recording
Free spontaneous items (adults only):
3 minutes (session time) of free spontaneous, rich context items (story telling) (an open number of spontaneous topics out of a set of 30 topics)
17 Elicited spontaneous items (adults only):
3 dates, 2 times, 3 proper names, 2 city name, 1 letter sequence, 2 answers to questions, 3 telephone numbers, 1 language
Read speech:
30 phonetically rich sentences uttered by adults and 60 uttered by children
5 phonetically rich words (adults only)
4 isolated digits
1 isolated digit sequence
4 connected digit sequences
1 telephone number
3 natural numbers
1 money amount
2 time phrases (T1 : analogue, T2 : digital)
3 dates (D1 : analogue, D2 : relative and general date, D3 : digital)
3 letter sequences
1 proper name
2 city or street names
2 questions
2 special keyboard characters
1 Web address
1 email address
221 application specific words and phrases per session (adults)
74 toy commands and 48 general commands (children)

The following age distribution has been obtained:
Adults: 255 speakers are between 15 and 30, 213 speakers are between 31 and 45, 76 speakers are between 46 and 60, and 6 speakers are over 60.
Children: 25 speakers are between 8 and 10, 25 speakers are between 11 and 14.

A pronunciation lexicon with a phonemic transcription in SAMPA is also included.

Telephone
The Finnish-Swedish SpeechDat(II) FDB-1000 contains the recordings of 1,000 Finnish speakers (455 males, 545 females) recorded over the Finnish fixed telephone network. The FDB-1000 database is partitioned into 4 CDs, the first 3 CDs comprise 300 speakers sessions, the 4th comprises 100 speakers.

Speech samples are stored as sequences of 8-bit 8 kHz A-law. Each prompted utterance is stored in a separate file. Each signal file is accompanied by an ASCII SAM label file which contains the relevant descriptive information.

This speech database was validated by SPEX (the Netherlands) to assess its compliance with the SpeechDat format and content specifications.

Each speaker uttered the following items, in the variant of Swedish spoken in Finland:

* 1 isolated digit
* 1 sequence of 10 isolated digits
* 4 connected digits (1 sheet number -5 digits, 1 telephone number –9/10 digits, 1 credit card number -16 digits, 1 PIN code -6 digits)
* 1 currency money amount
* 1 natural number
* 3 dates (1 spontaneous date e.g. birthday, 1 prompted date, 1 relative or general date expression)
* 2 time phrases (1 spontaneous time of day, 1 time phrase)
* 3 spelled words (1 spontaneous e.g. own forename, 1 city name, 1 phonetically rich word)
* 5 directory assistance names (1 spontaneous e.g. own forename, 1 spontaneous city of growing up, 1 frequent city name, 1 frequent company name, 1 common forename surname)
* 2 yes/no questions (1 predominantly "yes" question, 1 predominantly "no" question)
* 6 application words
* 1 word spotting phrase using an embedded application word
* 4 phonetically rich words
* 9 phonetically rich sentences

The following age distribution has been obtained: 178 speakers are under 16, 412 speakers are between 16 and 30, 216 speakers are between 31 and 45, 160 speakers are between 46 and 60, and 34 speakers are over 60.

A pronunciation lexicon with a phonemic transcription in SAMPA is also included.

Speech Related
The Fixed1frDesign includes all database specifications including the full list of a designed corpus: a set of phonetically rich sentences and a set of application oriented utterances used within the French SpeechDat(II) database (ref. ELRA-S0076, produced in the framework of SpeechDat(II)). The SpeechDat common specification totals 40 utterances per call, comprising a mixture of spontaneous and read speech. The purpose of each telephone call was to record the basic structure of the utterances mentioned below. All utterances are read speech unless marked as spontaneous.

The list of utterances is as follows:

* 5 application words
* 1 sequence of 10 isolated digits
* 4 connected digits (1 sheet number -5+ digits, 1 telephone number -9/11 digits, 1 credit card number -14/16 digits, 1 PIN code -6 digits)
* 3 dates (1 spontaneous date e.g. birthday, 1 word style prompted date, 1 relative and general date expression)
* 2 word spotting phrases using an embedded application word
* 1 isolated digit
* 3 spelled words (1 spontaneous e.g. own forename, 1 spelling of directory assistance city name, 1 real/artificial name for coverage)
* 1 currency money amount
* 1 natural number
* 5 directory assistance names and 1 spelled name (1 spontaneous e.g. own forename, 1 city of birth/hometown, 1 most frequent city out of a set of 500, 1 most frequent company/agency out of a set of 500, 1 "forename surname", 1 spelled-out city of birth)
* 2 yes/no questions (1 predominantly "yes" question, 1 predominantly "no" question)
* 9 phonetically rich sentences
* 2 time phrases (1 spontaneous time of day, 1 word style time phrase)
* 8 phonetically rich words.

Statistics are supplied for each corpus, which are computed on the repetition of digits, letters or phonemes (diphones and triphones) depending on the corpus type. These statistics are reported in a separate file for each corpus.

A documentation file aiming at describing the entire corpus design is included on the CD-ROM. It also covers the motivations that lead to that particular design. Finally, a complete lexicon file (in SpeechDat format) is supplied. The CD-ROM does not contain any recordings.

Speech Related
With a view of supplying SpeechDat family projects with the textual material used in the Italian speech database, CSELT has decided to produce a CD-ROM with all database specifications including the full list of a designed corpus: a set of phonetically rich sentences and a set of application oriented utterances.
The Italian SpeechDat databases (produced in the framework of SpeechDat(M) and SpeechDat(II)) used this textual material.
The SpeechDat common specification totals 40 utterances per call, comprising a mixture of spontaneous and read speech. The purpose of each telephone call was to record the basic structure of the utterances mentioned below. All utterances are read speech unless marked as spontaneous.
The list of utterances is as follows:
- 3 application words;
- 1 sequence of 10 isolated digits;
- 4 connected digits: 1 sheet number (5 digits), 1 telephone number (9-11 digits), 1 credit card number (14-16 digits), 1 PIN code (6 digits);
- 3 dates: 1 spontaneous date (e.g. birthday), 1 prompted date (word style), 1 relative and general date expression;
- 1 word spotting phrase using an application word (embedded);
- 1 isolated digit;
- 3 spelled-out words (letter sequences): 1 spontaneous (e.g. own forename), 1 spelling of directory city name, 1 real/artificial name for coverage;
- 1 money amount in Lire;
- 1 natural number;
- 5 directory assistance names: 1 own forename (spontaneous), 1 city of birth/home town (spontaneous), 1 most frequent city, 1 most frequent company/agency, 1 "forename surname";
- 2 questions, including "Fuzzy" yes/no: 1 predominantly "yes" question, 1 predominantly "no" question;
- 9 phonetically rich sentences;
- 2 time phrases: 1 time of day (spontaneous), 1 time phrase (word style);
- 4 phonetically rich words.
In the case of the Italian fixed network database, four additional items were added to the one designed in the project:
- 1 telephone area code
- 1 money amount in EURO
- 2 "yes/no" questions
For the Italian SpeechDat corpus, the full list of items is supplied in two different files: the first contains the prompted text read by speakers in the supplied sheet and the second file contains the orthographic transcription.
Statistics are supplied for each corpus, which are computed on the repetition of digits, letters or phonemes (diphones and triphones) depending on the corpus type. These statistics are reported in a separate file for each corpus.
A documentation file aiming at describing the entire corpus design is included on the CD-ROM. It also covers the motivations that lead to that particular design.
Finally, a complete lexicon file (in SpeechDat format) is supplied.
The CD-ROM does not contain any recordings.

KEY FEATURES:

Type of resource : Textual material
Domain/Source: Textual material used within the Italian SpeechDat(M) and SpeechDat(II) databases
File format: ASCII; A lexicon in SAMPA is also provided
Distribution media: 1 CD-ROM
Related resources: Italian SpeechDat(M) database (ELRA-S0052 and S0053)

Telephone
This French SpeechDat(II) FDB-1000 database contains the recordings of 1,017 French speakers recorded over the fixed telephone network. This speech database was sponsored by the European Commission (CEC DGXIII), under the project LE2-4001. The database is partitioned into CD-ROMs

Speech samples are stored as sequence of 8-bit, 8kHz A-law and are uncompressed. Each prompt utterance is stored within a separate file (file extension FRA) and has an accompanying ASCII SAM label file (file extension FRO).

It contains 48 utterances (40 mandatory and 8 optional items) for 1,017 different speakers, 17 speakers have been added to the original 1,000 speakers to fit the requirements of the database. The main content of the database is speech and orthographic transcription files.

The database was validated by SPEX (the Netherlands) to assess its compliance with the SpeechDat format and content specifications.

It is designed for development and assessment of French speech recognizers.

Each speaker uttered the following items:

* 5 application words
* 1 sequence of 10 isolated digits
* 4 connected digits: 1 sheet number (5+ digits), 1 telephone number (9-11 digits), 1 credit card number (14-16 digits), 1 PIN code (6 digits)
* 3 dates: 1 spontaneous date (e.g. birthday), 1 prompted date (word style), 1 relative and general date exp.
* 2 word spotting phrases using an application word (embedded)
* 1 isolated digit
* 3 spelled words (letter sequences): 1 spontaneous, e.g. own forename, 1 spelling of direct. city name, 1 real/artificial for coverage
* 1 currency money amount
* 1 natural number
* 5 directory assistance names + 1 spelled name: 1 spontaneous, e.g. own forename, 1 city of birth / growing up (spont), 1 most frequent cities (set of 500), 1 most frequent company/agency (set of 500), 1 "forename surname", 1 spelled city of birth
* 2 questions, including "fuzzy" yes/no: 1 predominantly "yes" question, 1 predominantly "no" question
* 9 phonetically rich sentences
* 2 time phrases: 1 time of day (spontaneous), 1 time phrase (word style)
* 8 phonetically rich words

Desktop/Microphone
The French Speecon database is divided into 2 sets:

1. The first set comprises the recordings of 550 adult French speakers (275 males, 275 females), recorded over 4 microphone channels in 4 recording environments (office, entertainment, car, public place).
2. The second set comprises the recordings of 50 child French speakers (20 boys, 30 girls), recorded over 4 microphone channels in 1 recording environment (children room).

This database is partitioned into 23 DVDs (first set) and 3 DVDs (second set).
The speech databases made within the Speecon project were validated by SPEX, the Netherlands, to assess their compliance with the Speecon format and content specifications. Each of the four speech channels is recorded at 16 kHz, 16 bit, uncompressed unsigned integers in Intel format (lo-hi byte order). To each signal file corresponds an ASCII SAM label file which contains the relevant descriptive information.

Each speaker uttered the following items:

* Calibration data:
o 6 noise recordings
o The “silence word” recording
* Free spontaneous items (adults only):
o 5 minutes (session time) of free spontaneous, rich context items (story telling) (an open number of spontaneous topics out of a set of 30 topics)
* 17 Elicited spontaneous items (adults only):
o 3 dates, 2 times, 3 proper names, 2 city name, 1 letter sequence, 2 answers to questions, 3 telephone numbers, 1 language
* Read speech:
o 30 phonetically rich sentences uttered by adults and 60 uttered by children
o 5 phonetically rich words (adults only)
o 4 isolated digits
o 1 isolated digit sequence
o 4 connected digit sequences
o 1 telephone number
o 3 natural numbers
o 1 money amount
o 2 time phrases (T1 : analogue, T2 : digital)
o 3 dates (D1 : analogue, D2 : relative and general date, D3 : digital)
o 3 letter sequences
o 1 proper name
o 2 city or street names
o 2 questions
o 2 special keyboard characters
o 1 Web address
o 1 email address
o 208 application specific words and phrases per session (adults)
o 73 toy commands and 46 general commands (children)

The following age distribution has been obtained:

* Adults: 245 speakers are between 15 and 30, 210 speakers are between 31 and 45, 95 speakers are between 46 and 60.
* Children: 17 speakers are between 8 and 10, 33 speakers are between 11 and 14.

A pronunciation lexicon with a phonemic transcription in SAMPA is also included.

Written Corpora
The German-French Reciprocal Parallel Corpus (GeFRePaC) was produced by the Multilinguale Forschung/Multilingual Research Abteilung Lexik, Institut für Deutsche Sprache (Germany) through a funding from ELRA in the framework of the European Commission project LRsPThe German-French Reciprocal Parallel Corpus (GeFRePaC) was produced by the Multilinguale Forschung/Multilingual Research Abteilung Lexik, Institut für Deutsche Sprache (Germany) through a funding from ELRA in the framework of the European Commission project LRsP&P (Language Resources Production & Packaging - LE4-8335).
The German-French Reciprocal Parallel Corpus (GeFRePaC) is a 30 million word corpus (15 million for each language) for the purpose of developing, enhancing and improving translation aids (dictionaries, lexicons, platforms) for French-German and German-French translation.
The database consists of the following parallel corpora:
European Union CELEX Database: Treaties, Foreign relations, Law, Complementar Law and all the published documents of the "European Parliament".
Celex-Database: 22,000,000 words (German+French) (http://www.outlaw-web.com)
Europarl: 8,320,000 words (German+French) (http://www.europarl.eu.int)
It covers natural general language as used in public socio-political discourse and it has a focus on multilingual administration and commercial and legal documentation. GeFRePaC comprises a large variety of text types for which there is a rapidly growing need for translation but which currently defy successful machine translation. The corpus is encoded according to the PAROLE guidelines, it was aligned on the sentence level and also for single word translation units on the lexical level, POS-tagged in conformity with EAGLES recommendations and validated according to the most current version of the ELRA guidelines. The parallel German-French texts were aligned using a program developed at the Equipe Langue et Dialogue, Laboratoire Loria, Nancy. The text files containing markup for paragraphs and sentences were processed by the Tree Tagger developed at the IMS Stuttgart. The text files are automatically converted into TEI-conformant SGML format.

Desktop/Microphone
The German SpeechDat-Car database comprises 338 German speakers recorded over the mobile telephone network. This database is partitioned into 17 DVDs and 1 CD. The speech databases made within the SpeechDat-Car project were validated by SPEX, the Netherlands, to assess their compliance with the SpeechDat-Car format and content specifications.

The speech data files are in two formats. The signal data format for the in-car mobile platform recordings is 16 kHz, 16 bit, uncompressed unsigned integers in Intel format (lo-hi byte order); the channels are multiplexed in a single file, with the channel sequence being 0-1-2-3. The format of the fixed platform audio files is 8 kHz, 8 bit alaw encoding. Each signal file is accompanied by an ASCII SAM label file which contains the relevant descriptive information.

Each speaker uttered the following items:
- 2 voice activation keywords
- 1 sequence of 10 isolated digits
- 7 connected digits : 1 sheet number (4+ digits), 1 spontaneous telephone number (9-11 digits), 3 read telephone numbers, 1 credit card number (16 digits), 1 PIN code (6 digits)
- 3 dates : 1 spontaneous date (e.g. birthday), 1 prompted date, 1 relative or general date expression
- 2 word spotting phrases using an application word (embedded)
- German data phrases
- 4 isolated digits
- 7 spelled words : 1 spontaneous (own forename or surname), 1 spelling of directory city name, 4 real word/name, 1 artificial name for coverage
- 1 money amount
- 1 natural number
- 7 directory assistance names : 1 spontaneous (own forename or surname), 1 city of birth / growing up (spontaneous), 2 most frequent cities, 2 most frequent company/agency, 1 "forename surname"
- 9 phonetically rich sentences
- 2 time phrases : 1 time of day (spontaneous), 1 time phrase (word style)
- 4 phonetically rich words
- 69 application words: 13 mobile phone application words, 22 IVR function keywords, 32 car products keywords, 2 additional common application words
- 2 additional language dependent keywords
- spontaneous sentences

The following age distribution has been obtained: 187 speakers are between 16 and 30, 72 speakers are between 31 and 45, 70 speakers are between 46 and 60, and 9 speakers are over 60.

A pronunciation lexicon with a phonemic transcription in SAMPA is also included.