SLPOtherTrends in hearing2021

Interleaved Acoustic Environments: Impact of an Auditory Scene Classification Procedure on Speech Perception in Cochlear Implant Users.

Anja Eichenauer, Uwe Baumann, Timo Stöver and 1 others

PMID 34028332

WHAT IT FOUND

Automated scene classification improved speech perception in quiet and steady noise, but offered little or no benefit in reverberant rooms or when multiple speakers were talking.

It can also misclassify complex scenes, so it is not a reliable solution for every noisy environment.

Key findings

01Enabling the automated scene classification improved the signal-to-noise ratio required for understanding speech by 2.4 dB for unilateral users and 1.3 dB for bilateral users on average.

02The benefit of the automated setting was significant in free-field continuous noise but was not significant in modulated noise, and was smaller in reverberant conditions.

03Pretesting showed the algorithm failed to consistently classify reverberant speech in quiet and produced asymmetric classifications in some noise conditions.

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.

Already have one?

What it does not show

The ASC classifications were not recorded during the actual test runs, so the specific program selected for each participant at each moment is unknown. The test used nonsense sentences, which do not reflect the cognitive demands of real-life conversation. The study used a highly complex 128-loudspeaker setup that cannot be replicated in most clinics.

Declared interests

The study was funded by Cochlear Deutschland GmbH & Co. KG and the Moessner Foundation Frankfurt am Main.

The easy way to misread this

Do not assume the automated scene classification will improve speech understanding in all noisy environments. The study found no significant benefit in modulated noise, and the algorithm frequently misclassified reverberant scenes, potentially leaving the user in a suboptimal program.

Read it on PubMed →