为了解决这一问题，欧洲电信标准化协会（ETSI）的语音和多媒体传输质量委员会（TC STQ）最新发布了测试方法，能够客观评估超宽带和全频段终端在背景噪音存在下的语音通信性能。ETSI TS 103 281提供了测试方法和模型，用以评估电话会议和视听应用程序的对话服务。
为了分析影响语音质量的各种参数，新的ETSI规范将语音质量、背景噪音传输质量和整体质量都纳入考量。TS 103 281所描述的客观模型是基于对美式英语、德语和汉语（普通话）说话者的大量研究，因此其预测精度高于目前市场上的大多数模型。这也是在ETSI标准的验证结果
ETSISTQ技术委员会主席汉斯威廉吉尔利希（Hans Wilhelm Gierlich）说：“移动行业十分需要这些测试方法。”
ETSI releases test methods to improve QoS in the presence of background noise
Background noise is an issue for telecommunications and needs to be taken into account in terminal design. The higher the intended speech quality as e.g. in wideband or super-wideband telephony, the more critical the terminal optimization is.
To tackle this problem, ETSI’s technical committee on Speech and multimedia Transmission Quality (TC STQ) has just published testing methods to objectively evaluate the performance of super-wideband and fullband terminals for speech communication in the presence of background noise. ETSI TS 103 281 provides testing methods and models to evaluate conversational services for teleconferences and audio-visual applications. It includes state-of-the-art codecs, typical communication packet loss and jitter conditions and recordings in handset, headset, hands-free and car hands-free mode.
To analyse the various parameters influencing speech quality, the new ETSI specification considers the speech quality, the background noise transmission quality and the overall quality. The objective models described in TS 103 281 are based on a huge subjective study conducted with American English, German, and Chinese (Mandarin) speakers. The prediction accuracy of the models is higher than most models currently on the market which is proven by the validation results also provided in the ETSI standard.
“The mobile industry was in demand of these testing methods.” says Hans Wilhelm Gierlich, chairman of the ETSI STQ technical committee.
The STQ committee continues work to develop methodologies, test arrangements and requirements for evaluating the performance of wearable devices for speech communication. The performance evaluation of voice assistant devices e.g. for use in IoT as well as the simulation of room and background noise characteristics in laboratory environments is progressing, new specifications are expected in 2018. The group is also working on a specification aimed at improving listening quality for people with impaired hearing. The specification will define requirements and measurement methods to assess the impact of different transmission impairments on intelligibility as well as a model to assess speech intelligibility objectively. It is expected to be published later this year.
地址：北京海淀区马甸东路9号 邮编：100088 邮箱：email@example.com