Video for the Web 2020 and Beyond

Episode Summary

Executive Summary: The episode explores the evolving web video stack in 2020 and beyond, focusing on why “just use an MP4” is insufficient for serious streaming. Wes and Scott compare formats (MP4, WebM, HLS, DASH), players, and hosting/encoding services, then explain Scott’s real-world migration from YouTube/Vimeo toward Mux + Video.js for better control, quality, and UX.

Main Topics: Why web video is more complex than simple embedding: The hosts explain that modern video delivery involves ingest, transcoding, adaptive streaming, and playback—far beyond dropping an MP4 into a video tag. Video formats and adaptive streaming: They break down container formats (MP4, WebM) versus adaptive streaming protocols (HLS, DASH), including browser support, segmenting, and bitrate switching. Players and browser compatibility: The discussion compares players such as Video.js, JW Player, Shaka Player, HLS.js, and Theoplayer, emphasizing that players bridge format limitations and add features like Chromecast and AirPlay. Hosting and encoding service roundup: A long comparison of YouTube, Vimeo, Wistia, Cloudflare, Mux, Cloudinary, Brightcove, Bitmovin, Azure, AWS, and JW Player, judged by cost, developer effort, and feature depth. Scott’s production stack choice: Scott explains why he moved to Mux for hosting/encoding and Video.js for playback, citing easier development, customization, and successful launch of a premium series. Tradeoffs: ease of use vs control: The episode repeatedly contrasts fully managed platforms with highly technical, API-first stacks, highlighting the pain of building custom players and workflows from scratch.

Key Arguments: Serious video products need adaptive streaming and a proper player; a simple MP4 in a video tag is fine for small assets but not for a premium streaming experience. HLS and DASH enable bitrate switching and segmented delivery, which improves playback quality based on network conditions. Browser support is fragmented, so JavaScript libraries and custom players are often required to make streaming work reliably across devices. Many services appear simple on the surface but hide complexity in ingest, transcoding, permissions, and player integration. Vimeo can be cheaper than raw DIY infrastructure for some creators, especially if public hosting fits the plan. Wistia is excellent for marketing video but becomes expensive for large tutorial libraries. Cloudflare’s custom stream element created enough friction that Scott abandoned it, showing how nonstandard abstractions can create long-term maintenance problems. Mux offered the best balance of developer experience and control for Scott, while Video.js provided the ecosystem and flexibility he wanted.

Data Points: Flash Player retirement: Occurred the other day before recording - Used as a marker that the web video landscape has shifted away from legacy technology Vimeo plan: $200/year - Mentioned as the old low-cost Vimeo option for unlimited hosting Vimeo private tutorial pricing: $6,000–$7,000/year - Scott says Vimeo required a custom plan for private tutorials Mux encoding cost: 5 cents/minute - Per minute of video encoded Mux storage cost: 0.3 cents/minute/month - Storage pricing for hosted video Mux streaming cost: 0.13 cents/minute - Per minute of streaming Cloudflare setup time: Close to launch, but ultimately abandoned - Scott got very far before deciding their custom stream element was too limiting JW Player starter plan: $10/month - For 150 GB hosting and 500 GB streaming Wistia overage cost: 33 cents/GB after 200 GB - Used to illustrate how quickly Wistia becomes expensive Cloudinary video issue: Certain video sizes have limitations - Support said some uploads can only be modified in specific ways depending on upload method AWS video workflow components: S3, MediaConvert, Lambda, SNS, CloudWatch, DynamoDB, CloudFront - Scott describes the VOD-on-demand architecture he tried to use Final stack: Mux + Video.js - Scott’s chosen production setup after testing multiple services Supported service accounts tested: 7+ accounts - Scott signed up for Vimeo, Mux, Wistia, Cloudflare, Cloudinary, Bitmovin, AWS, and Azure

Pivotal Quotes: "There's so much more to it, specifically around transcoding and encoding and live streaming and all that good stuff." — Wes Bos: Opening framing for why the episode matters "I ended up going with Mux." — Scott Tolinsky: Scott states his final hosting/encoding decision after testing many services "You do want a custom player because if you just use the browser's built-in one, you don't get access to Chromecast, which is number one feature I wanted, or AirPlay." — Wes Bos: Explaining why custom players are necessary for premium video experiences

Implications: For creators and SaaS teams, video infrastructure is now a strategic choice: managed platforms save time, but API-first stacks offer better control, branding, and features. Expect more use of adaptive streaming, specialized players, and hybrid architectures.