The new Enhanced Voice Services (EVS) codec was standardized by the 3GPP in December 2014 to offer improved performance. Among its many audio quality enhancement features, it provides an error concealment mechanism to neutralize the effects of transmission errors.
EVS codec error resilience
A frequent issue for audio services is deterioration of quality due to radio channel characteristics and network error conditions, leading to the loss, corruption, or unacceptable latency of audio packets. To combat this issue, the EVS codec performs frame error concealment (FEC) on a limited number of frames in order to produce a neutral signal and replace what otherwise sounds strange or missing. Previous AMR-WB audio codecs can manage approximately 3% of lost frames without critical artifacts, but there is noticeable distortion. In comparison, the EVS codec can perform FEC on more than 15% of lost frames without critical artifacts and still maintain relatively clear voice signals.
Packet loss is seen by users to be the most noticeable disturbance in voice communications and is likely to occur for a variety of reasons. This is especially true with unmanaged networks, where Quality of Service (QoS) is not present, due to congestion and misrouted packets. Managed networks have more capability than unmanaged networks, but they require a skilled administrator or engineer to achieve the best performance. Still, packet loss causes degradation in voice quality. Packet loss can be bursty in nature, with periods of several seconds during which the loss may be as high as 10% to 15%. The average packet loss rate for a call may be low; however, these periods of high loss rate can cause noticeable degradation in call quality.
The EVS channel aware mode offers significantly improved error concealment in voice communications as compared to AMR-WB and AMR-NB codecs over Voice over LTE (VoLTE). The error resilience is achieved using a form of in-band forward error correction. Source-controlled coding techniques are used to identify candidate speech frames for bit rate reduction, leaving spare bits for transmission of partial copies of prior frames such that a constant bit rate is maintained. The self-contained partial copies are used to improve the error robustness, in case the original primary frame is lost or discarded due to late arrival. Subjective evaluation results from ITU-T P.800 Mean Opinion Score (MOS) tests are provided below, showing improved quality under channel impairments as well as negligible impact to clean channel performance. EVS channel aware mode is specifically created for reliable speech transmission under especially bad network conditions, such as high packet loss, jitter, etc. Channel aware is available in two configuration modes: EVS WB 13.2 and EVS S-WB 13.2.
As the graph indicates, an audio quality comparison between EVS and AMR reveals that EVS delivers superior quality at similar bit rates. The robustness of EVS using channel aware mode (as shown with 13.2k) delivers superior performance through all packet loss scenarios, with excellent audio quality results.
More on audio quality measurement for voice
Perceptual Objective Listening Quality Assessment (POLQA) is based on ITU-T Recommendation P.863. POLQA provides voice quality analysis of HD Voice and Full HD Voice. It offers an advanced level of benchmarking accuracy, and adds significant new capabilities for WB, S-WB and EVS voice signals along with support for the most recent voice coding and VoIP transmission technologies. POLQA is one of the best choices for evaluating, optimizing and measuring the voice quality of next-generation networks and delivers accurate, high-resolution analysis results. POLQA is measured as a MOS. MOS is expressed as a single number in the range 1 to 5, where 1 is lowest perceived audio quality and 5 is the highest perceived audio quality measurement. The standard deviation of the mean distribution needs to be very small.
It is important to test the uplink and downlink audio MOS measurement of devices. When tested in parallel, the MOS needs to remain stable. For more information on EVS and the challenges associated with testing devices, please refer to Spirent’s white paper: