Article
Subjective Testing of the IVAS Codec (en)
* Presenting author
Abstract:
A new low-delay speech and audio codec for upcoming services with stereo or immersive audio communication is currently being developed in a work item within 3GPP. This so-called Immersive Voice and Audio Services (IVAS) codec is an extension to the Enhanced Voice Services (EVS) codec and will support different stereo and immersive audio formats at bit rates between 13.2 and 512 kbps. The capabilities and the performance of the codec candidate were recently tested by an extensive set of auditory experiments.The auditory tests comprised different test designs for naïve and experienced listeners as well as different playback configurations (headphones and multi-channel loudspeaker setups). The tests were designed to assess the performance of the IVAS codec candidate in relation to the EVS codec. This contribution provides a brief overview of the codec candidate itself but its main focus is on the auditory tests that were conducted and their results. It is illustrated that the IVAS codec candidate generally outperforms EVS, i.e., it is capable of achieving better quality at comparable bit rates or the same quality at lower bit rates, respectively.