Reducing Listening Effort of Audio Podcasts by Applying Equalization and Dynamic Processing at Playback
Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
When broadcasting audio for radio and television an important attribute of the program material is speech intelligibility. Podcast creators without knowledge of broadcasting audio may not take this into account when creating audio material for podcasts. The playback systems used for listening to podcasts are often mobile and may often have a limited frequency- and dynamic range. In noisy environments listening effort can often be high. The aim of this paper was to evaluate processing techniques that can be used at playback to decrease listening effort for a wide range of podcasts without affecting speech quality negatively. Four different processing techniques implementing equalization and dynamic compression was evaluated in a listening test with 17 naive listeners. The stimuli were played through circumaural headphones and multitalker babble noise was played through four loudspeakers in a quad setup to simulate a noisy playback environment. Two different EQs, Formant EQ (attenuates low frequency vocals and emphasizes high frequency consonants) and Modified formant EQ (same curve as Formant EQ but with less gain/attenuation) were combined with two different soft knee dynamic compressors (fast attack/release-times and slower attack/release-times). All four processing techniques significantly reduced listening effort and the processing technique with a Formant EQ and a compressor with slower attack/release times was significantly better at lowering listening effort than the two techniques using a Modified formant EQ. None of the processing techniques significantly affected perceived speech quality.
Place, publisher, year, edition, pages
2015. , 40 p.
Teknik, speech intelligibility, listening effort, speech quality, podcast, podcasts, ljudteknik, sound engineering
IdentifiersURN: urn:nbn:se:ltu:diva-49633Local ID: 6f3b716e-b0b5-4755-b14c-f046c231121bOAI: oai:DiVA.org:ltu-49633DiVA: diva2:1022980
Subject / course
Student thesis, at least 15 credits
Audio Technology, bachelor's level
Validerat; 20150524 (global_studentproject_submitter)2016-10-042016-10-04Bibliographically approved