Quantifying Quality Degradation of the EVS Super-Wideband Speech Codec

摘要: Voice transmission networks are commonly planned with the help of computational quality models, which give an estimate expected that a user will experience. The most popular these tools is E-model. When certain parameters known, such as applied codec and its bitrate, model able to predict perceived communication system. Up now, E-model only available for narrowband telephony (300–3400 Hz) limited also wideband (100–7000 Hz). With extension voice super-wideband (50–14000 Hz), introduction EVS mobile state art smartphones, update has become necessary. To this end, we firstly examined improvement over results from mixed-band listening-only tests, where found improved by 15%. Then, calculated impairment factors analyzed robustness towards packet loss, using auditory instrumental methods.