Akademik Veri Yönetim Sistemi
International Conference on Electrical and Electronics Engineering (ICEEE), Muğla, Türkiye, 22 - 24 Nisan 2024, ss.426-433, (Tam Metin Bildiri)
Degradation in speech on VoIP (Voice over Internet Protocol) refers to any deterioration in the quality of the audio during a VoIP call. This degradation can manifest in various forms, including distortion, delay, jitter, packet loss, echo, background noise, and so on. Detecting degradation in VoIP calls is crucial for the following several reasons such as improving user experience and satisfaction, boosting the effective communication and collaboration, impacting productivity and efficiency, allowing providers to proactively address issues and maintain service quality, helping network administrators to identify the root causes of quality issues and implementing solutions to improve performance and reliability. Overall, detecting degradation in speech on VoIP is essential for ensuring high-quality communication, maintaining customer satisfaction, and optimizing the performance of VoIP systems and networks. In this work, focus on detecting degradations on voice over VoIP using deep learning algorithms, namely, Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Long Short-Term Memory Networks (LSTMs). To show the effectiveness of the models, TCD- VoIP dataset is carried out during the experiments. Experiment results demonstrate that the inclusion of deep learning methodologies exhibits remarkable experiment results.