Development and validation of the Brazilian voice bank for various emotions (EMOVOX-BR).
Héryka Maria Oliveira Lima, Larissa Nadjara Almeida, Alexandra Christine de Aguiar and 1 others
PMID 42018805WHAT IT FOUND
Ten experienced voice therapists correctly identified the intended emotion in more than 70% of the clips in a new Brazilian Portuguese audio bank of 39 recordings.
They relied on pitch for every emotion, and on loudness only for anger.
Key findings
01All 39 clips in the bank reached agreement above 0.7 between judges, and 24 of them (61%) reached almost perfect agreement, meaning the judges nearly always named the emotion the actor intended.
02The ten speech-language pathologist judges identified the intended emotion in more than 70% of the clips; anger and neutral had the highest success rates and fear the lowest.
03Pitch was the parameter the judges chose most often for every emotion and loudness mattered for anger, while articulation, speech speed, pneumofonoarticulatory incoordination, vocal quality and fluency were each named by fewer than half of them.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
Only ten judges took part, all of them speech-language pathologists already working in voice, so this says nothing about how well untrained listeners or clients would judge the same clips. The speakers were actors with no vocal changes, so the finding that articulation, speed, fluency and vocal quality did not matter for recognising emotion may not hold for the dysphonic clients you see. The emotions were acted rather than spontaneous and were recorded from a single sentence, so the clips may not represent emotion in everyday speech. Each judge rated 200 clips in around 40 minutes, which is quick work for listening judgments and may have affected their answers. The study reports no patient outcomes, so it does not show that using this bank improves anything in therapy.
Declared interests
Funded by the National Council for Scientific and Technological Development (CNPq) in Brazil, process number 434508/2018-7. The paper declares no conflicts of interest.
The easy way to misread this
Do not treat this bank as a test that measures a client's ability to recognise emotion, or as evidence that emotion recognition training works. Ten experienced voice specialists labelled recordings made by actors, which shows the clips are consistent, not that they are a clinical assessment or that anything changes for patients.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →