"glowing telegram" serverless with pulumi and python - Episode 132
3In this video, we explore the process of implementing silence detection for audio extraction using FFMPEG, and discuss how to efficiently structure our tasks to achieve this. We aim to optimize our application by running tasks in parallel and ultimately pushing the collected data into DynamoDB.
Throughout the video, I take you through the coding process, considering the parallelization of tasks that include metadata collection, keyframe extraction, and audio processing. We focus on adding a new task for silence detection, examining how to leverage existing tools without complicating the workflow.
I showcase how to refactor the existing code base by reorganizing it into a more modular structure, primarily focusing on the audio track extraction and silence detection code. This includes creating new modules and ensuring configurations suit the intended design.
Moreover, we discuss performance tuning, considering AWS Batch configurations, memory allocations, and CPU units to ensure efficient processing. As we explore different settings for batch jobs, I outline the potential for further performance enhancements and describe strategies for testing these adjustments.
Join me as we troubleshoot, optimize, and enhance our web application infrastructure using modern technology practices.
🔗 Check out my Twitch channel for more streams: https://www.twitch.tv/saebyn
GitHub: https://github.com/saebyn/glowing-telegram
Discord: https://discord.gg/N7xfy7PyHs