Parts of the research efforts of the IPS involve the development of software to aid in the scientific work with speech signals and corpora. The software is used both for basic research in speech science and for the creation of speech technology.
Data collection is costly and sometimes tedious in experiments on speech production and perception. But most steps in the workflow can be drastically sped up by using and combining computer tools. Some tools have been designed specifically for single steps along the workflow, others are apt for multiple tasks. For some tasks there is a variety of tools to choose from.
Some programs can be downloaded, some are offered as a web service. Web services can be used directly in the browser (more information here).
Manages copora of speech recordings (both acoustic and physiogical) and allows for segmentation, annotation, hierarchical structuring and statistical evaluation.
Performs forced alignment (phonetic segmentation) and labeling based on audio files and the corresponding orthographical/phonological transcription in several languages.
Platform-independent speech recorder designed specifically for phonetic experiments.
Online perception experiment for mobile devices and computers.
Distributed orthographical transcription of spoken language.
Platform-independent web-based speech recorder.
Converts metadata tables to CMDI files.
Converts a written text into a canonical phonetic transcription. Several languages are supported.
Generates BAS partitur files with a speech chunk tier TRN from various input formats. Used to enhance the performance of our MAUS segmentation services for sound files that contain long pauses or dialog data.
Automatic phonetic segmentation and labelling without text input for several languages.
Syllabification of phonological and phonetic transcripts in several languages.
Audio recorder and audio editor.
Aligns text sequence pairs by minimizing their edit distance.
Web interface for MaryTTS, a text to speech system developed by phoneticians of Uni Saarland and DFKI.