Florian Schiel, Katerina Louka
BAS Bayerisches Archiv für Sprachsignale
Florian Schiel: The Validation of Speech Corpora, Bastard Verlag, 2003, www.bas.uni-muenchen.de/Forschung/BITS/TP2/Cookbook
The speech corpus SC2 has been validated against general principles of good practise. The validation covered completeness, formal checks and manual checks of the selected sub samples. It must be considered that missing data reduces the corpus in its usability and there could occur problems in using the corpus for further applications.
the results of an in-house validation of the speech corpus SC2 made
year 2004 within the project 'BITS' by the
General Documentation directory contains the following documentation
the SC2 corpus which can be found under /doc:
||file describing the database structure|
|| session information
||text prompt corpus (100) in
||text prompt corpus (100) in LaTeX
||text prompts in order of
prompting (400) in ISO88591
||text prompts in order of
prompting (400) in LaTeX
||words of all spoken words (LaTeX)
|| SAMP-PA dictionary of
canonical pronunciation (LaTeX)
||Extended German SAM-PA
|| documentation of the BAS
guidelines for proper pronunciation coding in German SAM-PA
· Administrative Information:
Validating person: n. a.
Date of validation: n. a..
Contact for requests regarding the corpus: ok
Number and type of media: 1 CD-ROM ok
Content of each medium: total size 520MB
statement and intellectual property rights (IPR): ok
· Technical information:
about file system type and directory structure:
1 CDROM with ISO 9660 format
of used codes (no white space in file names!):
data/<session>/<speaker><hall>< mic><utt_id>.nis ok
Formats of signals and annotation files: If non standard formats are used it is common to give a full description or to convert into a standard format: ok, raw signal file
Compression: n. a.
Sampling rate: 8kHz ok
Valid bits per sample: (others than 8, 16, 24, should be reported): 16 bit, ok.
Used bytes per sample: 2 ok
· Database contents:
Clearly stated purpose of the recordings: ok. (README)
Speech type(s): (multi-party conversations, human-human dialogues, read sentences, connected and/or isolated digits, isolated words etc.) ok
Instruction to speakers in full copy: ok
· Linguistic contents of prompted speech:
Specifications of the individual text items: ok (sc2_pro_*.txt)
Specification for the prompt sheet design or specification of the design of the speech prompts: n.a.
Example prompt sheet or example sound file from the speech prompting: n.a.
· Linguistic contents of non-prompted speech:
Multi-party:(number of speakers, topics discussed, type of setting - formal/informal) n.a.
Human-human dialogues: (type of dialogues, e.g. problem solving, information seeking, chat etc., relation between speakers, topic(s) discussed, type of setting, scenarios) n.a.
Human-machine dialogues: (domain(s), topic(s), dialogues strategy followed by the machine, e.g. system driven, mixed initiative, type of system, e.g. test, operational service, Wizard-of-Oz) n.a.
· Speaker information:
Speaker recruitment strategies: indirect information in the README file
Number of speakers: 10 ok
Distribution of speakers over sex, age, dialect regions: ok,
(in each data directory is a speaker file: vp.txt)
Description/definition of dialect regions: indirectly given through the place of origin (vp.txt)
· Recording platform and recording conditions:
Recording platform: ok
and type of microphones: The data was transmitted via a wireless phone
(Siemens Gigaset) to an analog phone set connected to the base station
of the wireless. To this analog phone set a laptop was connected via
standard I/O technique.
- Company name and type id: ok
- Electret, dynamic, condenser: n.a
- Directional properties: n.a.
- Mounting: n.a.
Position of speakers: (distance to microphone) ok
Bandwidth: (if other than zero to half of sampling rate) no information
Number of channels and channel separation: 1 channel
environment: a hall
of car diagnosis maintenance
· Annotation (BAS Partitur Format (BPF) files):
Unambiguous spelling standard used in annotations: ok
Labeling symbols: ok
List of non-standard spellings (dialectal variation, names etc.): given
Distinction of homographs which are no homophones: n.a.
Character set used in annotations: ok
Any other language dependent information as abbreviations etc: given
Annotation manual, guidelines, instructions: given
Description of quality assurance procedures: not given
Selection of annotators: not given
Training of annotators: not given
Annotation tools used: n.a.
Text-to-phoneme procedure: ok
Explanation or reference to the phoneme set: ok. (SAMPA.TXT)
Phonological or higher order phenomena accounted in the phonemic transcriptions: ok
· Statistical information:
Frequency of sub-word units: phonemes (diphones, triphones, syllables,...): n.a.
Word frequency table: n.a.
Any other essential language-dependent information or convention: n.a.
Indication of how many files were double-checked by the producer together with percentage of detected errors: not given
Status of documentation: acceptable
The following list contains all validation steps with the methodology and results.
Completeness of signal
Completeness of meta data files: ok
Completeness of annotation files: ok.
Correctness of file names: ok.
Empty files: none
Status of signal, annotation and meta data files: ok
Cross checks of meta information: ok
Cross checks of summary listings: ok
Annotation and lexicon
5% of the 'usable' data, the BAS files and the annotation SAM files were checked in comparison. 10,50% of the data contained errors (42 errors out of 400). In some par files the various noise classes were not marked.
The revalidation was able to repair some data (par files, lexicon). The results of the manual validation couldn't be repaired.
The corpus is ok. The corpus is well documented and no data or documentation files are missing.