In today's fast-paced world of communication, the importance of voice call quality cannot be ignored. As technology advances, so do people’s expectations for clear and high-quality calls. In order to meet this demand, ITU-T developed a new standard in 2011 - POLQA (Perceptual Objective Listening Quality Assessment). This technology is widely used in various voice applications and aims to predict call quality by analyzing digital voice signals to provide users with a more favorable call experience.
POLQA is an objective model for measuring speech quality, which aims to make automated measurement results close to subjective listening scores.
The core of POLQA lies in its testing scope and technical capabilities. According to the standard, POLQA is not only suitable for traditional voice calls, but also extends to the processing capabilities of high-bandwidth audio signals, which makes it adaptable to the current rich and diverse communication environment. POLQA allows the evaluation of speech signals no longer to be limited to a narrow frequency band (300-3400 Hz), but also supports a wider speech range (50-14000 Hz) and sound quality evaluation of recordings, which further enhances its application potential.
As the successor of PEAQ (Perceptual Evaluation of Audio Quality), POLQA's technical superiority lies not only in avoiding the shortcomings of PESQ, but also in its ability to cope with a wider range of audio signals. Unlike PESQ, the POLQA algorithm is better trained to measure problems caused by latency variations and time-sensitive signals. This flexibility enables it to meet the increasingly complex challenges of modern communications technology.
POLQA’s testing process uses real speech as test stimuli to conduct a comprehensive subjective hearing evaluation.
The development of POLQA began in 2006. After gathering the ideas of many companies, the standard was further launched and released in 2011. Today, POLQA is not only a standard, but also an evaluation tool relied on by countless technical experts. It represents a qualitative change in voice quality assessment technology. Its structure consists of multiple complex modules, including delay alignment algorithm, sample rate estimation and core model operation.
According to relevant definitions, POLQA is a Full Reference algorithm. This means that it compares the degraded or processed speech signal to the original signal. The difference in each sample is scored as a difference in sound quality, which helps provide an objective perspective on clarity and naturalness in the context of the call.
POLQA's algorithm focuses on analyzing the time and frequency relationships between samples and converting them into mean opinion scores (MOS). The rating system ranges from 1 (poor) to 5 (excellent) to help communication service providers understand users' perception of call quality.
POLQA's results simulate users' subjective ratings, making it an ideal voice quality assessment tool.
In the core model of POLQA, the perceptual model is an important component. Its purpose is to simulate the human auditory experience based on subjective scoring criteria. This perceptual model is designed to take into account multiple distortion types and adjust scores based on the user's auditory response. Compared with normal speech signals, POLQA can idealize various audio features to make them more realistic during evaluation.
With the development of technology, POLQA has also shown many applications in academic research. Researchers have used this tool to explore the relationship between speech quality and multilingual or non-native listeners, further expanding its application areas.
POLQA is not only an advanced technical standard, it also provides a reliable solution for improving the quality of telephone calls. Through accurate evaluation and continuous improvement, POLQA makes our voice communication experience clearer. Against this background, are you also thinking about what innovative technologies will be available in the future that can further improve the quality of our voice calls?