Study On The Sound Quality Evaluation System Of Recording Device

With the development of digital multimedia technology, sound recording products are widely popular, and their application fields include media, justice forensics, cultural diffusion, etc. The recording quality relates to the authenticity and reliability of the application, so it is very important to do something research on sound quality evaluation. Research on sound quality evaluation of recording products is mainly subjective method in our country, and it is rarely reported in other countries. This paper focuses on sound quality objective evaluation research. Combined with the dynamic time warping algorithm, we choose the Mel-frequency cepstral coefficient as the critical characteristic parameter of speech signals to realize the purpose of judging recording equipment quality. In this paper, a sound quality objective evaluation system is designed and it is composed with the playback, recording and evaluation function. The main contents are as follows:1. An improved dual-threshold endpoint detection algorithm is proposed. Through the Mel-frequency cepstral coefficient and the differential Mel-frequency cepstral coefficient calculation of the pretreatment speech signal, MEL values could be extracted. Simulation experiments validate that the improved method is feasible.2. A sound quality evaluation method is proposed through combining dynamic time warping algorithm, which can be used to search the optimal trajectories, and model comparison method that get the signal similarity.3. A sound quality objective evaluation system is designed and performed. The system includes the playback, recording and evaluation function. We tested three kinds of recording device using subjective and objective ways, respectively. The validation indicators include the evaluation scores, potentially influencing factors, etc. The tested results show that the evaluation system is accurate and efficient.