<![CDATA[Parallel computing has become an effective approach to improving data processing efficiency, particularly for CPU-bound applications. This study aims to implement parallel computing using the Ray framework for audio feature extraction in a multi-core single-node environment. The dataset consists of 1000 .wav audio files from the RAVDESS dataset, which are processed through feature extraction in the time, frequency, and time–frequency domains using the Librosa library. The extracted features include zero-crossing rate, spectral features, and Mel-Frequency Cepstral Coefficients (MFCC), which provide a comprehensive representation of audio signal characteristics. Each audio file is treated as an independent task and distributed across multiple workers for parallel processing. This approach allows data processing to be performed independently without inter-task dependencies. Experiments are conducted by varying the number of workers from 1 to 4 to observe their impact on execution time. The primary parameter observed is the total execution time required to complete the feature extraction process. The results show that the implementation of parallel computing reduces execution time compared to serial processing. However, the performance improvement is not strictly linear due to system overhead and hardware resource limitations. These findings indicate that a task parallelism approach using Ray can serve as a practical solution to improve audio data processing efficiency in resource-constrained environments without requiring complex distributed computing infrastructure.]]>
