_/_/_/_/ _/_/ _/_/_/_/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/_/_/_/ _/_/_/_/ _/_/_/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/ _/_/_/_/ _/ _/ _/_/_/_/ BAVARIAN ARCHIVE FOR SPEECH SIGNALS F. Schiel - 07/13/04 The Bavarian Archive for Speech Signals (BAS) was founded in Jan 1995 as an initiative of the Institute of Phonetics at the University of Munich, Germany. The BAS will develop, validate, administrate and disseminate corpora of spoken German to the speech community as well as to speech engineering industry. GENERAL The BAS is located at the University of Munich, Germany. Address: BAS c/o Institut fuer Phonetik Schellingstr. 3/II 80799 Muenchen Germany Phone: +49-89-21802758 Fax: +49-89-21805790 email: bas@bas.uni-muenchen.de Detailed and up-to-date information about the work of BAS and the available speech resources can be found in the following URL: http://www.bas.uni-muenchen.de/Bas/ CORPORA Presently the following corpora are available on CD-R/DVD-R/Harddisc. Note that a subset of these corpora is also online accessible for university members and licensees of BAS resources in the BAS CLARIN Repository (tagged with (*) in the following list). Siemens 1000 - SI1000 (SC7) 10 speakers - 10000 utterances - dictation - orthography Siemens 100 - SI100 100 speakers - 10000 utterances - dictation - orthography PhonDat 1 - PD1 (*) (2nd edition) 201 speakers - 21681 utterances - read speech - orthography, canonical transcription, automatic segmentation PhonDat 2 - PD2 (*) (2nd edition) 16 speakers - 3200 utterances - read speech - orthography, canonical transcription, automatic segmentation, prosodic labeling Verbmobil I (*) (Recording 1993 - 1996) Verbmobil II (*) (Recording 1997 - 2000) Strange Corpus 1 - SC1 ('Accents') (*) 88 speakers - 1 story - read speech - orthography, canonical transcription Strange Corpus 2 - SC2 ('Noises') (*) 8 speakers - 8 repetitions of 100 utterances - field recordings with real background noise - noise annotated - orthography, canonical pronunciation, noises Strange Corpus 10 - SC10 ('Accents II') 70 speakers (67 non-native, 3 native German speakers) - 100 phonetically balanced sentences, numbers from 1 to 100, 1 story, 1 dialogue, 1 re-telling of a German story - transliteration, orthography, canonical transcription Erlanger Bahnansage - ERBA 106 speakers - 11100 utterances - read speech - orthography SPINA ('Robot Comands') (*) (new edition) 22 speakers - robot control - 10810 utterances - read speech - phoneme and word segmentation Regional Variants of German 1 - RVG1 (*) Recordings of all German spoken regions - 498 speaker - 32 CD-ROMs Siemens Synthesis Corpus - SI1000P 2 professional speakers - laryngographic signal - prosodical labelling - 4 CD-ROMs Taxi Corpus - TAXI (*) 94 dialogues of a German speaking cab dispatcher and an English speaking client - recorded via real phone connections (fixed and GSM) - orthography, canonical pronunciation, translation Hempel's Sofa - HEMPEL (*) 3909 recordings of spontaneous speech (monologues) via public phone lines - SpeechDat transcription FORMTASK (*) 17293 recordings of 4366 speakers answering 4 questions via public phone (land)lines - SpeechDat transcription Regional Variants of German J - RVG-J (*) recordings of read and spontaneous speech by adolescents age 13-20 - SpeechDat annotation Siemens Webcommand - WEBCOMMAND 15600 recording of commands addressed to a web pad - British English and Frensh - 49 speakers - office environment - SpeechDat annotation Ziptel (SpeechDat(M)) - ZIPTEL (*) 7746 recordings of street names, ZIP codes, city names and phone numbers - 1957 speakers - all environments - SpeechDat annotation BITS Logatome Synthesis Corpus - BITS-LG 11036 logatom recordings covering all German diphones - 4 professional speakers - studio, 2 mics, laryngo - manual segmentation, BAS Partitur Format BITS Unit Selection Synthesis Corpus - BITS-US 6732 sentence recordings covering all German diphones in different prosodic contexts - 4 professional speakers - studio, 2 mics, laryngo - manual phonetic segmentation and prosodic annotation, BAS Partitur Format SmartKom Audio - SKAUDIO 1.0 (*) Special edition containing all audio channel recording of the SmartKom corpora - 224 speakers - 448 sessions - Scenarios: Public, Home, Mobil SmartWeb Handheld Corpus - SHC (*) 10966 human-machine queries using a smart phone - 156 speakers - natural environment, 2 mics (collar, Bluetooth), UMTS + high quality channel - transliteration (Verbmobil compatible), BAS Partitur Format SmartWeb Motorbike Corpus - SMC 2315 human-machine queries on running motorcycle - 36 speakers - natural environment, 2 mics (Bluetooth helmet, neck micro), UMTS + high quality channel - transliteration (Verbmobil compatible), BAS Partitur Format SmartWeb Video Corpus - SVC (*) 2218 human-machine queries in a human-human-machine situation using a smart phone, video face-capture of asking person - 99 speakers - natural environment, 2 mics (collar, Bluetooth), UMTS + high quality channel - transliteration (Verbmobil compatible), manual turn segmentation, BAS Partitur Format Ph@ttSessionz - PHATTSESSIONZ (*) 1019 sessions of up to 138 items each (read, spontaneous) of adolescent speakers of age 12-20 - 1019 speakers - natural environment (school), 2 microphones (headset, desktop), demoscopic distribution within Germany - transliteration according SpeechDat standard, manual segmentation start/end utterance, BAS Partitur Format files, MAUS segmentation Alcohol Language Corpus - ALC (*) Recordings of intoxicated and sober speakers of age 22-75 - 150 speakers (estimate for the final corpus) - automotive environment, 2 microphones (headset, mouse micro) - transliteration according extended SpeechDat standard, manual segmentation start/end utterance, BAS Partitur Format files, MAUS segmentation VeriDat Speaker Verification Corpus - VERIF1DE (*) Corpus for speaker verification via telephone - 150 speakers - 20 recording sessions per speaker - transliteration SpeechDat standard, SpeechDat Database Format Age and Gender Speech Corpus - aGender (*) Paralinguistic speaker classification via phone lines - 945 speakers - 1-7 recording sessions (calls) per speaker - transliteration according to SpeechDat standard, SpeechDat database format Corpora e Lessici dell'Italiano Parlato e Scritto, map task recordings - CLIPS_MT_MANUAL (*) Italian Maptask Recordings from CLIPS - 30 Speakers - 2 Recording Sessions per Speaker - Transliteration, Segmentation according CLIPS Standard - BPF, TextGrid, Emu BAS Siemens Hoergeraete Corpus - HOESI (*) Lombard Dialogue speech - 24 Speakers - 12 Recording Sessions per Speaker Pair - Segmentation speech - non-speech - BPF Atlante sintattico della Calabria - AsiCa Recordings of Calabrese (Italy) - 68 Speakers - 331 Recording Sessions - Orth.-Phon. Transcription - TextGrid Audioatlas Siebenbuergisch-Saechsischer Dialekte - ASD Historic recordings of Saxonian German spoken in Romania in 1970 - 1805 Speakers - 2264 Recording Sessions - Orth. and phon. transcriptions - TextGrid Speech of brother pairs (Dissertation Feiser) - BROTHERS Speech data of dissertation Feiser (2016) German - 20 (10 brother pairs) - 7240 recording sessions - orth. and autom. phon. transcription - TextGrid, emuDB German Pronunciation Rules Set - PHONRUL 9.0 This set of computer-readable rules describes the most common known effects in German pronunciation if deviating from the so called canonic or citation form of words. The knowledge of this rule set was derived from empirical analysis of speech corpora as well as from a multitude of publications about German phonetics. German Pronunciation Lexicon PHONOLEX 3.1 PHONOLEX is the result of a close cooperation between the DFKI Saarbruecken, Computational Linguistics Lab, the University of Leipzig and the Bavarian Archive for Speech Signals in Munich. It comprises a simple list of word forms (inflected words) with the entries orthography, linguistic information, canonical pronunciation and empiric pronunciation. See the BAS Web pages for more information. Version 3.1 comes with over 1.600.000 entries. We are now in the process to verify a core of words in PHONOLEX according the newly developed BAS Standard for German Pronunciation Coding (www.bas.uni-muenchen.de/Bas/BasGermanPronunciation) German Geo Names Database GEO1 1.0 BAS GEO1 is a simple database about the most important location names of Germany, Austria and Switzerland together with their canonical pronuciation coded in SAM-PA. BAS GEO1 may be used as a basis for automatic speech recognition of German postal addresses or to feed a speech synthesis algorithmn. Although we aim at a complete set of names the database in its current is far from complete. Future updates will be distributed to all users automatically, if you provide us with a valid email address. See www.bas.uni-muenchen.de/Bas/BasGEO1eng.html for details. SOFTWARE The BAS develops the following major software packages for free usage: MAUS (Munich Automatic Segmentation) MAUS is a package of tools that may be used together with the HTK of University of Cambridge to automatically segment speech into phonemic segments. See www.bas.uni-muenchen.de/Bas/software for the download package and some documentation. SpeechRecorder The SpeechRecorder is a Java based application that allows the easy and flexible recording of large corpora on a laptop or desktop computer. See www.bas.uni-muenchen.de/Bas/software for details WebTranscribe WebTranscribe is a JavaWebstart based transcriber tool that allows - in connection with a WWW server and a database - the distributed annotation of large speech corpora. Contact Chr. Draxler (draxler@bas.uni-muenchen.de) if you are interested using this tool for your own work. MISCS In www.bas.uni-muenchen.de/Bas/BasGermanPronunciation you will find a rule set that describes the proper coding of canonical German pronunciation in SAM-PA. Based on these rules manually transcribed pronunciation dictionaries at BAS are produced. In www.bas.uni-muenchen.de/Forschung/BITS/TP1/Cookbook you will find a html version of the cookbook 'The Production of Speech Corpora' In www.bas.uni-muenchen.de/Forschung/BITS/TP2/Cookbook you will find a html version of the cookbook 'The Validation of Speech Corpora' COOPERATIONS The BAS seeks to cooperate with public institutions as well as with industrial partners to further develop new German speech databases. BAS can be a platform to re-distribute existing German speech corpora or to unify forces for the development of new resources. MORE INFORMATION Please refer to the BAS WWW HomePage (see above) or contact us via email: bas@bas.uni-muenchen.de Florian Schiel, Institut fuer Phonetik und sprachliche Kommunikation Universitaet Muenchen, Schellingstr. 3, 80799 Muenchen, GERMANY tel.: +49-(0)89-2180-2807 fax: +49-(0)89-21805790 email: schiel@bas.uni-muenchen.de www: http://www.bas.uni-muenchen.de/