Speech recordings can be played back and annotated. It’s a very handy tool for scoring speech from students or clients that need to be analysed for reading mistakes. Specific error labels can be assigned to parts of the audio. The results can be used for further investigation by a more senior annotator, or simply be archived for a longer time.
It’s useful as a standalone tool for individual linguists and speech therapists. It also serves as a hub for annotating massive amounts of speech recordings, for assignment to multiple annotators. Here, the goal is teaching the machine learning algorithms how to detect reading errors automatically. The tool is previously used to train our speech models for English and Dutch.
The tool can be tailored to suit the needs of our customers (different target languages, different error labels) to optimize trained models for your target audience.
Demo and pricing information available upon request. Usually part of a R&D arrangement.