August 2023 Summaries

Multi-track audio is an emerging technology that allows creators to provide multiple audio tracks for their on-demand videos, enhancing the viewer's experience by enabling them to choose the sound they prefer. This can include different languages, director's commentary, or varying music volumes. It also increases accessibility for visually impaired viewers with audio descriptions. Multi-track audio is currently available for all Mux Video accounts and can be easily implemented using their API.

Mux Data successfully monitored one of the largest live streaming events in history, with over 30 million concurrent viewers. The event shattered existing streaming records and generated massive amounts of data. To support this scale, Mux deployed a separate ingestion pipeline, scaled event collection and processing horizontally and vertically, load tested the system architecture, and made adjustments based on real traffic conditions. Flexibility in their microservice architecture allowed them to handle unprecedented scale during the event.

Mux's journey from its origins as Zencoder highlights the transformative impact of cloud-based video encoding and the evolving landscape of video technology. Founded by the same team behind Zencoder, an early innovator in cloud encoding, Mux reflects a decade of lessons learned from both successes and failures in the industry. Zencoder, acquired by Brightcove, disrupted traditional encoding with its cost-efficient API, but faced competition from Amazon's Elastic Transcoder, which ultimately couldn't match Zencoder's market momentum. The narrative emphasizes that the future of video encoding is moving towards integration with audience-specific data and just-in-time (JIT) encoding, which offer significant cost savings and improved quality of experience by encoding only what is watched and adapting to audience needs. This shift challenges standalone encoding APIs, which lack the means to fully leverage audience data or implement JIT encoding effectively. As the market evolves, integrated platforms like Mux are positioned to lead by offering advanced solutions that traditional services may struggle to provide.

Mux has released upload SDKs for iOS and Android in public beta to help developers upload videos directly from their mobile applications. These SDKs handle file chunking and retry failed chunks, ensuring successful uploads even on variable network connections. The SDKs also standardize video inputs before uploading, allowing for faster processing times and quicker readiness of videos. Additionally, the SDKs enable resolution-based pricing control, which can help lower costs by configuring the player to stream no higher than a specified resolution. Mux encourages developers to try out these SDKs and provide feedback during the public beta period.

Service workers, often seen as a means for caching in progressive web apps, are HTTP proxies that can do much more than just caching. They can manipulate requests, resize or convert images, add security layers to HLS streams, and stitch in ads. A service worker is first-party JavaScript that intercepts requests for all pages within its scope. Service workers run in their own context, allowing intensive work without interfering with web page interactivity. They are well supported across evergreen browsers, including mobile ones. One overlooked element of modern streaming video is that it's a lot of text files, making service workers an ideal tool for manipulating media by editing those text files.

React Server Actions, available as an alpha in Next.js 13.4, aim to simplify posting data by offering a React-official way of calling functions on the server. This feature is particularly useful for handling form submissions and uploading videos to Mux. By using React Server Actions, developers can avoid writing separate APIs and reduce boilerplate code. However, there are still some limitations in terms of success and error states, which may be addressed in future updates.

The text discusses various aspects of integrating video into applications, including different types of providers, cost considerations, and the components involved in a DIY approach. It also touches upon video modes such as on-demand and real-time video, engineering expertise required for building video pipelines, and subsystems that could be included along with streaming footprints. The text suggests using video APIs or services as a middle ground between DIY and off-the-shelf solutions, highlighting Mux's features in this regard. It concludes by emphasizing the importance of feature prioritization when choosing a service for integrating video into applications.