Lossless compression reduces file size without throwing away any information — the decoded video is bit-for-bit identical to the original. Every pixel can be reconstructed exactly. The opposite is lossy compression (lossy), which discards information to achieve dramatically smaller files. Where lossy hits compression ratios of 100:1 or 1000:1, lossless typically manages 2:1 to 4:1 — useful but not enough for streaming.
The home of lossless video is production and archive. ProRes 4444 XQ at its highest setting, FFV1, HuffYUV, Lagarith and uncompressed sequences all preserve every bit of the original. They're used during editing (so multiple passes of compositing and grading don't accumulate quality loss), for VFX work (where mathematical precision matters), and for long-term archival. The Library of Congress in 2023 made FFV1 in MKV containers its "Preferred Format" for video preservation, recognising that even small losses compound over decades of migration.
For a streaming product, lossless rarely shows up in delivery but always shows up in the ingest and master pipeline. The master file ideally arrives lossless (or visually lossless ProRes 422/422 HQ, which is good enough in practice). All processing — colour grading, captioning, watermarking, segmentation — happens at full quality. Only the final transcoding step introduces lossy compression for delivery. This "lossless master, lossy delivery" pattern is universal across professional video pipelines and explains why a studio holds 100× the storage of what any viewer ever sees.
