Video Encoding & Transcoding — FFmpeg, AWS MediaConvert & Multi-Profile Delivery

H.264 has the widest device compatibility and lowest encode/decode cost. H.265/HEVC offers better compression but has licensing and Android fragmentation considerations. AV1 compresses best but costs more compute to encode and has more limited older-device support. Most production ladders mix H.264 as a compatibility baseline with H.265 or AV1 for capable devices.

Per-title (content-adaptive) encoding tunes bitrate to each piece of content's actual complexity — a static talking-head video needs less bitrate than a fast-action scene at the same visual quality. For large VOD libraries, the storage and CDN cost savings usually justify the added encoding pipeline complexity; for small libraries it may not be worth it.

FFmpeg gives full control and no per-minute cost, but you own the compute infrastructure and scaling. AWS MediaConvert is serverless and scales automatically but has a usage-based cost. Teams with existing AWS infrastructure and variable encoding volume usually prefer MediaConvert; teams with steady, high-volume encoding often find self-managed FFmpeg cheaper at scale.

Yes — WebVTT, SRT and CEA-608/708 closed captions are embedded directly into the HLS/DASH manifests during packaging, with multi-language track support, so this doesn't require a separate post-processing step.