What is HEVC? H.265 Video Codec Explained

HEVC (High Efficiency Video Coding), also known as H.265 or MPEG-H Part 2, is a video codec that delivers roughly 50% better compression than H.264 at equivalent quality. Finalized in 2013 by the Joint Collaborative Team on Video Coding (JCT-VC)—a partnership between ITU-T and ISO/IEC MPEG—it was designed for the 4K and HDR era, and it anchors premium video delivery across the Apple ecosystem, smart TVs, and set-top boxes today.

For our platform, HEVC is the codec of the premium tier: it is how we deliver 4K and HDR brand films and executive broadcasts at bitrates real networks can carry, aimed squarely at the Apple-heavy audiences leadership content most often reaches.

How HEVC works

HEVC keeps H.264's hybrid pipeline—partition, predict, transform, quantize, entropy-code—but scales every stage up.

Coding tree units (CTUs). Instead of fixed 16×16 macroblocks, HEVC partitions each frame into CTUs of 64×64, 32×32, or 16×16 samples. Each CTU splits recursively via a quadtree into coding units (CUs) down to 8×8, and CUs divide further into prediction units and transform units. Large blocks are the main reason HEVC wins on high-resolution content: a 64×64 flat region of sky is described once instead of sixteen times.

Intra prediction. HEVC expands directional intra prediction to 35 modes (33 angular directions plus DC and planar), versus H.264's nine. Finer angular granularity means more accurate prediction of edges and textures, and smaller residuals to transmit.

Inter prediction. Motion compensation adds merge and skip modes that inherit motion data from neighbors at almost no cost, advanced motion vector prediction, and asymmetric motion partitions—useful for objects crossing block boundaries.

Transform and quantization. Residuals use transform units from 4×4 up to 32×32 (H.264 maxed out at 8×8), which preserves energy compaction on large smooth areas. Quantization works on the same principle as H.264—coefficients are divided by a step size and rounded—but the larger transforms make the tradeoff visible at bigger scales: push the quantizer too hard and 32×32 blocks fail together.

Loop filtering. HEVC retains the deblocking filter and adds Sample Adaptive Offset (SAO), which classifies reconstructed samples and applies targeted offsets to reduce ringing and banding artifacts.

Entropy coding. CABAC is the only entropy coder—CAVLC was dropped—simplifying the standard while keeping its ~10–15% efficiency advantage.

Parallelism. HEVC was designed for multi-core hardware: tiles split a frame into independent rectangles, and wavefront parallel processing lets CTU rows be encoded concurrently. This is why real-time 4K HEVC encoding is practical in silicon.

Key features

• ~50% bitrate savings vs H.264 at the same perceptual quality—the headline claim, borne out across independent studies, with the largest gains at 4K and above.
• 10-bit and HDR as first-class citizens. The Main 10 profile carries 10-bit 4:2:0, which underpins HDR10 and Dolby Vision delivery and eliminates most color banding.
• Scalable to 8K, with main profiles supporting resolutions up to 8192×4320.
• Native support in Apple's ecosystem. Every iPhone, iPad, Mac, and Apple TV since ~2015 decodes HEVC in hardware, and iPhones record HEVC by default.
• HLS integration. Apple added HEVC to HLS in 2017 via fragmented MP4, making it the standard upgrade path for premium streaming tiers.

How Videospan uses HEVC

HEVC is our efficiency play at the highest-quality end of what customers ship. When a brand film, product launch, or leadership broadcast needs to go out in 4K or HDR, HEVC delivers that quality at bitrates a corporate network—or a viewer on hotel Wi-Fi—can actually sustain, and it does so natively on the Apple hardware where executive audiences tend to live.

• 4K and HDR delivery. For flagship content, HEVC's Main 10 profile carries 10-bit HDR that holds up on large displays without the bandwidth bill H.264 would impose at 4K.
• Apple-heavy audiences. Executive communications skew toward iPhones, iPads, and Macs, all of which have decoded HEVC in hardware for years. HEVC renditions mean better quality per bit for exactly the viewers a CEO message most needs to reach.
• Tiered, not exclusive. HEVC variants sit alongside our universal H.264 exports, so every video still plays everywhere—newer devices simply get a sharper, lighter stream.

HEVC vs H.264, AV1, and VP9

Advantages

• Roughly half of H.264's bandwidth for the same quality—immediate CDN cost savings at scale
• The safest path to 4K and HDR on TVs and Apple devices, with battle-tested hardware decode everywhere in that segment
• Mature encoders: x265, plus fast hardware encoders in every modern GPU
• Main 10 profile makes 10-bit HDR delivery routine

Limitations

• Licensing friction. HEVC patent rights are spread across multiple pools (MPEG LA, Access Advance/HEVC Advance, Velos) with overlapping claims and content-fee structures—uncertainty that pushed much of the web toward AV1
• Browser playback is inconsistent outside Safari and Edge; Chrome and Firefox typically rely on hardware decoders when present
• Encoding is several times more compute-intensive than H.264 at comparable settings
• AV1 now matches or beats its compression without royalties

When to use HEVC

Use HEVC when your audience lives on Apple devices, smart TVs, or set-top boxes, and especially for 4K HDR content where H.264's bitrate would be prohibitive. In a practical adaptive ladder, HEVC renditions serve the premium tier while H.264 covers the compatibility floor and AV1 covers modern browsers and Android. Avoid HEVC-only delivery to desktop web audiences unless you have verified decode support.

What to ask a platform vendor: if your team produces 4K or HDR content, ask whether the platform can deliver it as HEVC with the hvc1 tag so Apple devices and HLS players accept it—and whether it still ships a universal H.264 fallback, so nobody in the organization is handed a file that won't play.

A standard HEVC encode with ffmpeg:

ffmpeg -i in.mov -c:v libx265 -preset medium -crf 24 -tag:v hvc1 -pix_fmt yuv420p10le -c:a aac -b:a 160k out.mp4

Two details matter here: -tag:v hvc1 marks the stream so Apple devices and HLS players accept it from an MP4 container, and -pix_fmt yuv420p10le produces 10-bit output for HDR or banding-sensitive content (drop it for plain 8-bit SDR).

Glossary

CTU— Coding Tree Unit

HEVC's fundamental processing unit, up to 64×64 samples, recursively split into coding units by a quadtree.

Quadtree partitioning

The recursive subdivision of a CTU into four equal quadrants, letting the encoder match block size to picture content.

Intra prediction mode

One of 35 directions (plus DC and planar) used to predict a block from its already-decoded neighbors within the same frame.

SAO— Sample Adaptive Offset

An in-loop filter that classifies reconstructed samples and applies corrective offsets, reducing ringing and banding after deblocking.

Main 10

The HEVC profile supporting 10-bit 4:2:0 video—the required format for HDR10 and Dolby Vision delivery.

Wavefront parallel processing

A technique that encodes offset CTU rows concurrently, enabling multi-core speedups without breaking prediction dependencies.

Tile

An independently coded rectangular region of a frame, allowing decoders to parallelize or to decode only part of a very large picture.

Patent pool

A licensing body that bundles patents from many holders into a single license. HEVC is unusual in having several competing pools with overlapping claims.

hvc1

The codec tag indicating HEVC parameters are stored in the container (rather than inline), required for HEVC playback on Apple platforms and in HLS.

Frequently asked questions

Related

• Video Codecs Explained: H.264, HEVC, AV1 & VP9
• What is H.264? AVC Video Codec Explained
• What is AV1? The Royalty-Free Video Codec Explained
• What is VP9? Google's Open Video Codec Explained
• MP4 Container Format
• HLS Streaming