Perception of emotions in compressed speech by cochlear implant users
Introduction Cochlear implants (CIs) can restore speech perception in individuals with severe cochlear hearing loss. However, access to paraverbal cues, such as emotional prosody, remains limited, particularly in electronic communication, where speech is often digitally compressed. Methods This study compares CI users and normal-hearing (NH) listeners in their ability to recognize basic emotions from neutral sentences presented either as uncompressed audio or audio processed using common audio codecs (MP3, SPEEX, AMR-WB), measuring accuracy, response time, and confidence. Results CI users showed significantly lower recognition accuracy, longer response times, and greater confusion between several specific emotions than NH listeners. An additional acoustic analysis confirmed that the investigated codecs measurably altered several acoustic properties of the stimuli, particularly spectral and formant-related measures. Nevertheless, under the investigated compression settings, no statistically significant effect of codec condition on emotion-recognition performance was detected.
Authors
- Ingo Siegert (ORCID: https://orcid.org/0000-0001-7447-7141)
- Martin Böckmann‐Barthel
- Jesko L Verhey
- Ece Usseli
Institutions
- Otto-von-Guericke-Universität Magdeburg (DE)
Publication Details
- Journal
- Frontiers in Audiology and Otology
- Published
- 2026-11-16
- DOI
- https://doi.org/10.3389/fauot.2026.1934163
- Primary Topic
- Hearing Loss and Rehabilitation
- Type
- article
- Field-Weighted Citation Impact
- 0.00