About our samples
Sample credits
Sources, licenses and changes for the two real meeting recordings you can try in the editor.
AMI Meeting Corpus
Recording credit: The AMI Consortium / University of Edinburgh. Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). Both samples are excerpts from real meetings. Their use does not imply affiliation with or endorsement by the creators.
Single-speaker sample
1:02One speaker discussing how to synchronize recording equipment.
A continuous 62.01-second excerpt from the original headset mix, saved in a rebuilt WAV container. The editor uses the complete existing excerpt and its original WhisperX transcript, with no text, speaker-label or timestamp corrections.
Source: IN1014.Mix-Headset.wav · 1050.22–1112.23 seconds in the original meeting.
Two-speaker sample
0:23Two speakers discussing equipment booking and recording.
A continuous 23.28-second crop from an existing meeting excerpt, saved in a rebuilt WAV container. Original model segments, words and speaker labels are retained; timestamps are shifted to start at the crop. No text or speaker-label corrections were made.
Source: EN2009d.Mix-Headset.wav · 4941.96–4965.24 seconds in the original meeting.
How the samples are presented
The audio has no stitching, speech synthesis, voice isolation, denoising, silence removal or speed changes. The MP4 versions are waveform visualizations of the same audio, not footage of the meetings. Automatic transcripts, word times and speaker labels may contain errors; these examples are demonstrations, not an accuracy benchmark.