Wavefront parallel processing (WPP) is a clever technique for parallelising video encoding and decoding within a single frame — without splitting it into independent tiles or slices that would lose compression efficiency. WPP lets multiple CPU cores work on consecutive rows of the same frame in parallel, with each row "lagging" the row above it by a fixed offset (typically 2 CTU blocks). That offset is just enough that each row's predictions can still reference content from the row above, which keeps compression tight, but small enough that all rows are working concurrently.
Visually, if you watched WPP encode a 4K frame in slow motion, you'd see a diagonal "wavefront" sweeping down the frame: row 1 starts at column 0, row 2 starts when row 1 reaches column 2, row 3 starts when row 2 reaches column 2, and so on. By the time all rows are active, dozens of CPU cores are processing the frame in lockstep. HEVC was the first major codec to standardise WPP as a syntax-level tool; AV1 implements equivalent parallelism by a different mechanism (no bitstream changes, encoder-side only).
For a product team, WPP is invisible plumbing that makes hardware-accelerated encoding and decoding of high-resolution video practical. The compression cost is essentially zero — under 1 % bitrate penalty versus non-parallel encoding — because the prediction dependencies are preserved. The speed benefit is huge: 4K and 8K real-time decode on consumer chips relies on WPP-like techniques. You won't configure WPP directly (modern encoders and decoders use it automatically when worth it), but it's part of the architectural reason 4K HDR streaming actually plays smoothly on phones and TVs in 2026 rather than burning their batteries flat.
