Service quality assessment

Ensuring quality and consistency across your entire network

Measuring subjective service quality is a tricky business that requires a lot of specialized knowledge and expertise. SwissQual, which has been working in this area for more than ten years, has developed three major algorithms for detailed analysis of subjective quality, including two new ITU-T standards for HD video and HD audio quality assessment.

POLQA

POLQA – which stands for “Perceptual Objective Listening Quality Analysis” – is the standard, next-generation voice quality testing technology for fixed, mobile and IP based networks. It offers a new level of benchmarking capability to determine the voice quality of HD Voice, 3G and 4G/LTE mobile network services.

POLQA is also the only HD quality assurance algorithm that has official ITU-T approval. In 2010 POLQA was selected to form the new ITU-T voice quality testing standard, P.863, after rigorous testing and a four-year competition between the leading companies in speech quality testing.

And an enormous amount of effort has gone into ensuring POLQA works quickly and accurately. For example, more than 100,000 samples of mobile and VoIP phone calls were used to train and optimize the POLQA algorithm, meaning that POLQA is superior to existing standards, and has overcome all known issues and limitations of PESQ.

POLQA, which was developed in conjunction with OPTICOM and TNO, offers immediate, strong support for testing of new wideband 4G/LTE networks delivering HD-quality voice services.

SQUAD-HD

SQuad-HD is SwissQual’s proprietary full-reference quality assessment algorithm for HD voice channels. It makes use of pre-processed speech samples transmitted over the channel under test to determine the degradation from the channel itself. On the receiving end the degradation impact is determined by comparing the degraded sample to the original test sample.   SQuad-HD consists of three key elements:

• a pre-processing unit that adjusts the reference and coded samples in the time and level-domains;
• a psychoacoustic model is that used to reduce both samples to the perceptually relevant features; and
• an opinion unit which evaluates the perceptual difference between reference and degraded samples and outputs the result as a MOS objective value.

The output quality scale is mapped onto the metrics of the “overall listening quality” and “single talk conditions” by using the Absolute Category Rating (ACR) opinion scale per ITU-T’s recommendation P.800.1. For each speech clip evaluated, SQuad-HD provides detailed information such as:

• MOS rating
• Temporal clipping
• Interruptions / Impulsive noise
• Frequency shift
• Delay spread & deviation
• Active speech level
• Noise level
• Static SNR
• Channel gain pattern
• Channel frequency response