New official version
On Wednesday, 05 June 2024, we have released a new version of the EXMARaLDA tools!
This version contains new functionality, usability enhancements, and a few bug fixes for all tools. We are happy to say that a part of this work has been made possible by users supporting EXMARaLDA development and maintenance with an EXMARaLDA support licence. If you are using EXMARaLDA, please consider purchasing a support licence for your project, research group or institute.
A full list of changes for the new version can be found here. Feel free to post or comment on issues on GitHub if you discover a bug or have an idea for a feature. Here is an overview of the most important changes in the new official version:
Automatic speech recognition (ASR) via OpenAI Whisper web service
Besides importing the results of several ASR services (Amberscript, aTrain, noScribe, Adobe Premiere) into the Partitur-Editor, it is now also possible to call OpenAI’s Whisper web service directly for a selection of a media file in the editor. All it takes is a registration with and an API key from https://platform.openai.com/, and the service can be called as a “drop-in” replacement to manual transcription. Watch the little screencast on the left to see how it works.
Improved ergonomics for basic transcription actions
Buttons used for selection, playback, and creation of transcription and annotation events in the Partitur-Editor have been rearranged to make the transcription process more ergonomic. This is an optional change which you can ignore if you are happy with the “traditional” arrangement.
New editing operations
Functionality for copying / pasting entire “structures” (i.e. an arrangement of more than one event on more than one tier) and for bulk operations on groups of events (move left/right) have been added to the Partitur Editor’s Edit and Event menus. This should make it easier to work with data that is annotated on several layers. A new menu item “Tabulate events…” in the Tier menu of the Partitur-Editor arranges events of selected in a tabular view, enabling a more focused inspection of dependencies between tiers.
Further improvements
- Collect tier categories and types across a whole corpus in COMA
- Collect annotation types across a whole corpus in COMA (with an option for corpus-wide Search & Replace)
- Navigate from the recording into the transcription (Partitur-Editor)
- Open the Partitur-Editor for a given search result from EXAKT
- Case-insensitive sorting in EXAKT