Paper ID | SPE-7.1 | ||
Paper Title | Speaker embeddings for diarization of broadcast data in the ALLIES challenge | ||
Authors | Anthony Larcher, Ambuj Mehrish, Marie Tahon, Sylvain Meignier, LIUM - EA4023, Le Mans Université, France; Jean Carrive, David Doukhan, French National Institute of Audiovisual (INA), France; Olivier Galibert, Laboratoire National d’Essais (LNE), France; Nicholas Evans, EURECOM, France | ||
Session | SPE-7: Speaker Recognition 1: Benchmark Evaluation | ||
Location | Gather.Town | ||
Session Time: | Tuesday, 08 June, 14:00 - 14:45 | ||
Presentation Time: | Tuesday, 08 June, 14:00 - 14:45 | ||
Presentation | Poster | ||
Topic | Speech Processing: [SPE-SPKR] Speaker Recognition and Characterization | ||
IEEE Xplore Open Preview | Click here to view in IEEE Xplore | ||
Abstract | Diarization consists in the segmentation of speech signals and the clustering of homogeneous speaker segments. State-of-the-art systems typically operate upon speaker embeddings, such as i-vectors or neural x-vectors, extracted from mel cepstral coefficients (MFCCs) or spectrograms. The recent SincNet architecture extracts x-vectors directly from raw speech signals. The work reported in this paper compares the performance of different embeddings extracted from MFCCs or the raw signal for speaker diarization and broadcast media treated with compression and downsampling, operations which typically degrade performance. Experiments are performed with the new ALLIES database that was designed to complement existing, publicly available French corpora of broadcast radio and TV shows. Results show that, in adverse conditions, with compression and sampling mismatch, SincNet x-vectors outperform i-vectors and x-vectors by relative DERs of 43% and 73% respectively. Additionally we found that SincNet x-vectors are not the absolute best embeddings but are more robust to data mismatch than others. |