The paper describes the current state of development of a multi-purpose software tool for speech research. This is composed by a "visualization front-end", for displaying and editing the speech signal with associated annotations and acoustic features, and a "batch-processing interface" for applying speechprocessing algorithms to a whole database of signals. The software is mostly written in JAVA, but an extension mechanism is provided in order to integrate the interface with processing techniques implemented in different programming languages. The presented tool includes an original phone segmentation algorithm, for which some new experimental results are reported that prove its robustness to telephone bandwidth distortions.