


                     _/_/_/_/         _/_/         _/_/_/_/
                    _/      _/       _/ _/        _/      _/
                   _/      _/       _/  _/       _/
                  _/      _/       _/   _/       _/
                 _/_/_/_/         _/_/_/_/        _/_/_/
                _/      _/       _/     _/             _/
               _/      _/       _/      _/             _/
              _/      _/       _/       _/    _/      _/
             _/_/_/_/         _/        _/     _/_/_/_/


                   BAVARIAN ARCHIVE FOR SPEECH SIGNALS 

               University of Munich, Institut of Phonetics
               Schellingstr. 3/II, 80799 Munich, Germany
                      bas@phonetik.uni-muenchen.de


         COPYRIGHT University of Munich 2001. All rights reserved.   
    This corpus and software may not be disseminated further - not even
      partly - without a written permission of the copyright holders.  

                      Additional Copyright Holders
		      University of Braunschweig

----------------------------------------------------------------------

VERBMOBIL II Dialog Database  (BAS Edition)

12.01.98 / 20.08.03

----------------------------------------------------------------------

BONUS MATERIAL  1.2

This volume of the Verbmobil Dialog Database collects all materials that
do not fit into the scheme of signal files and accompanying BAS Partitur
Files (BPF).

CONTENTS

README		: this file
doc		: generell VM documentation on CDROMs
softw		: generell examples for I/O and some useful scripts
BPS		: Prosodic LBG and LBP tiers that were produced by more
		  than one human labeler in parallel.
		  These data might be interesting to investigate the 
		  reliability of manual prosodic labelings.
		  See also the documentation about the BPF tier LBG and
		  LBP in the Dokument doc/formats.html
SAP		: Manual annotations in SAM-PA that were produced by more 
		  than one labeler in parallel.
		  These data might be used to investigate the 
		  inter-labeler-agreement of manual phonetic labelings.
AdditionalDok   : Collection of generell (internal) documentation, 
		  scripts, programs, publications, VM documents, etc.
		  Most of it in German only. This part might be useful,
		  if certain questions arise concerning the way data
		  were processed, labelled, annotated, etc. This section
		  is provided 'as is' and was not revised by the BAS
		  before edition.
English_Treebank	: This directory contains syntax trees of the 
                          english turns in the Negra-file format. Annotated 
			  by the University of Tuebingen.
German_Treebank		: This directory contains syntax trees of the
			  german turns in the Negra-file format. Annotated
			  by the University of Tuebingen.  Newer version
			  than partitur file.
Japanese_Treebank	: This directory contains syntax trees of the
			  japanese turns in the Negra-file format.
			  Annotated by the University of Tuebingen.  Not
			  included into partitur file, because of
			  problems with the word segmentation.
Translations	: Contains manually produced translations of parts of the 
		  corpus. These data may be used e.g. for training of 
		  statistical translation engines.
		  Note that these translations are also incorporated in to the
		  standard BAS Partitur Formats files (BPF) of the corpus
		  (tier TLN).
LanguageModel	: The official VM language model for German. This LM was 
		  used as the official language model in the VM acoustic
		  evaluations as well in the official speech recognizer
		  of the VM prototype. 

See also the READMEs in the resp. subdirs.		  


HISTORY 

01.01.01 : 1.0  First edition
20.03.03 : 1.1  Bug fixes in SAP tier files: minor typos
20.08.03 : 1.2  Bug fixes in Translations: leading blanks in turns removed
                                           typos in data of VM45.1 removed
					   superfluous translation files
					   .../ger-eng/cd24/g226ac.trl.fv
					   .../ger-jap/cd24/g226ac.trl.fv
					   .../ger-eng/cd24/g240ac.trl.fv
					   .../ger-jap/cd24/g240ac.trl.fv
					   deleted
