Compare VP9 and AV1 video codecs in detail. Analyze encoding speed, compression efficiency, quality at same bitrate, browser support, and hardware decoding to choose the right royalty-free codec.
Marcus Rivera·February 19, 2026·13 min read
The video codec landscape has been dominated by proprietary standards for decades. H.264 and H.265 power most of the video you watch, but both carry patent licensing costs that ripple through the entire chain — from hardware manufacturers to streaming services to the open-source developer building a video player in their spare time. That licensing reality drove Google and the broader Alliance for Open Media to build royalty-free alternatives.
VP9 was Google's first serious answer, launched in 2013. AV1 is the successor, developed collaboratively by a consortium that includes Google, Mozilla, Microsoft, Apple, Netflix, Amazon, and others. Both codecs are royalty-free, both target high efficiency, and both compete directly with their patent-encumbered counterparts. But they are not interchangeable. AV1 is technically superior in almost every measurable dimension, yet VP9 maintains practical advantages that keep it relevant.
This guide breaks down the VP9 vs AV1 comparison across every dimension that matters: compression efficiency, encoding and decoding performance, hardware support, browser compatibility, and real-world deployment scenarios. If you have already read our H.265 vs H.264 vs AV1 breakdown, this article goes deeper into the royalty-free side of the codec world.
VP9 and AV1 codec architecture comparison diagram
The Royalty-Free Revolution
Before diving into technical comparisons, it is worth understanding why royalty-free codecs matter. H.264 licensing is managed by MPEG LA, which collects fees from hardware manufacturers, software distributors, and content providers. H.265 licensing is even more complex, split across multiple patent pools (MPEG LA, HEVC Advance, Velos Media) with different terms, making the total cost unpredictable.
VP9 and AV1 eliminate this entirely. Any company or individual can implement, distribute, and use these codecs without paying licensing fees. This is why YouTube encodes every video in VP9 (and increasingly AV1), why Netflix chose AV1 for mobile streaming, and why Firefox and Chrome prioritize open codec support.
The practical impact is significant. Open-source projects like FFmpeg, VLC, and OBS can include VP9 and AV1 support without legal concerns. Hardware manufacturers can add decode chips without per-unit royalties. And streaming services can deploy at scale without escalating licensing costs tied to viewership.
VP9 uses a block-based hybrid coding approach with variable block sizes from 4x4 to 64x64 pixels (called superblocks). It supports 8-bit and 10-bit color depth, 4:2:0 and 4:2:2 chroma subsampling, and resolutions up to 8K. Key technical features include:
Superblock partitioning: Recursive splitting of 64x64 blocks into smaller units
10 intra prediction modes for within-frame prediction
Compound prediction: Combines two reference frames for inter-frame prediction
8-tap switchable interpolation filters for sub-pixel motion estimation
Segmentation map: Assigns different encoding parameters to regions of a frame
Tile-based parallelism: Divides frames into independent tiles for multi-threaded encoding
AV1
AV1 builds on VP9's foundation but introduces substantially more coding tools. It uses superblocks up to 128x128 pixels and adds numerous efficiency features:
56 intra prediction modes (vs VP9's 10) including directional and non-directional modes
Loop restoration filter: Wiener and self-guided filtering for quality improvement
Film grain synthesis: Strips film grain during encoding and re-synthesizes during decoding
Warped motion compensation: Handles rotation, zoom, and perspective distortion
Palette mode: Efficient encoding of screen content and graphics
Reference frame scaling: Uses reference frames at different resolutions
Extended block partitions: More flexible recursive partitioning options
These additional tools give AV1 more ways to represent video data efficiently, which translates to better compression at the cost of encoding complexity.
Compression Efficiency
This is where AV1 decisively outperforms VP9. Multiple independent studies have measured the difference:
Study / Source
AV1 Bitrate Savings vs VP9
Test Conditions
Moscow State University (2024)
25-35%
Objective metrics (VMAF, SSIM, PSNR)
Netflix (2023 tech blog)
~30%
720p mobile streaming content
YouTube Engineering (2024)
20-30%
Mixed content, 480p-4K
Facebook Video (2023)
25-40%
User-generated content, mobile delivery
Bitmovin encoding study
20-30%
Professional broadcast content
On average, AV1 delivers the same visual quality as VP9 at 25-30% lower bitrate. This means a video that requires 4 Mbps in VP9 would need only about 2.8-3 Mbps in AV1 for equivalent quality. For streaming services delivering billions of hours of video, that reduction translates to massive bandwidth savings.
To put this in context with the broader codec landscape:
Codec
Approximate Bitrate for "Good" 1080p
Relative to H.264
H.264
5-8 Mbps
Baseline
VP9
3-5 Mbps
~30-40% savings
H.265
3-5 Mbps
~30-40% savings
AV1
2-3.5 Mbps
~50-55% savings
For a deeper dive into how these codecs compare, our video codecs explained article covers the full landscape.
Pro Tip: If you are encoding for web delivery and your audience primarily uses modern browsers (Chrome, Firefox, Edge, Safari 17+), AV1 will give you the best quality-to-filesize ratio of any available codec. The bandwidth savings are especially impactful for mobile viewers on limited data plans.
Quality comparison between VP9 and AV1 at the same bitrate
Encoding Speed
AV1's superior compression comes at a steep cost: encoding time. AV1 encoders are dramatically slower than VP9 encoders, though the gap has been narrowing.
Encoder Options
VP9 encoders:
libvpx-vp9: Google's reference encoder, mature and well-optimized
Typical 1080p encoding: 5-15 fps (speed depends on quality setting)
AV1 encoders:
libaom: Reference encoder, extremely slow but maximum quality
SVT-AV1: Developed by Intel/Netflix, dramatically faster than libaom
rav1e: Written in Rust, focuses on safety and correctness
aomenc with AOM_USAGE_REALTIME: For live streaming applications
Real-World Encoding Speed Comparison
Using FFmpeg on a modern 8-core CPU (encoding 1080p30 source):
Encoder
Preset / Speed
Encoding Speed
Quality (VMAF)
libvpx-vp9 (speed 1)
High quality
~8 fps
93.2
libvpx-vp9 (speed 4)
Balanced
~25 fps
91.8
libaom-av1 (cpu-used 3)
High quality
~2 fps
94.5
libaom-av1 (cpu-used 6)
Balanced
~8 fps
93.1
SVT-AV1 (preset 4)
High quality
~12 fps
94.1
SVT-AV1 (preset 8)
Balanced
~40 fps
92.5
SVT-AV1 (preset 10)
Fast
~80 fps
90.8
SVT-AV1 has been a game-changer for AV1 adoption. At preset 8, it encodes faster than libvpx-vp9 at speed 1 while delivering better quality. This makes AV1 encoding practical for most workflows today.
Pro Tip: For most users, SVT-AV1 at preset 6-8 offers the best balance of encoding speed and output quality. The libaom encoder produces marginally better results but takes 5-10x longer. Unless you are encoding a final master for a streaming service, SVT-AV1 is the practical choice.
Hardware Decode Support
Encoding speed matters for content creators, but decode support matters for viewers. A codec is useless if the audience's devices cannot play it.
VP9 Hardware Decoding
VP9 hardware decode has been widespread since 2016:
Mobile SoCs: Qualcomm Snapdragon 835+, Samsung Exynos 9810+, MediaTek Dimensity series, Apple M1+ (via software on iOS, hardware on macOS)
Smart TVs: Most TVs manufactured since 2018
Streaming devices: Chromecast, Fire TV, Roku (2018+)
Game consoles: PS5, Xbox Series X/S
AV1 Hardware Decoding
AV1 hardware decode is newer but expanding rapidly:
Desktop GPUs: Intel Arc, AMD RDNA3+, NVIDIA RTX 30-series+ (Ada Lovelace has full decode+encode)
Mobile SoCs: Qualcomm Snapdragon 8 Gen 1+, Samsung Exynos 2200+, MediaTek Dimensity 9000+, Apple A17 Pro+
Smart TVs: Samsung (2022+), LG (2023+), most 2024+ models
Streaming devices: Chromecast with Google TV (2022+), Fire TV Stick 4K Max (2023+)
Game consoles: PS5 (via system update), Xbox Series X/S
Hardware Encode Support
Hardware encoding is where VP9 and AV1 diverge significantly:
VP9 hardware encode: Intel Quick Sync (Kaby Lake+), no NVIDIA/AMD support
AV1 hardware encode: Intel Arc, NVIDIA RTX 40-series (Ada), AMD RDNA3+ — all major GPU vendors now support AV1 encoding
This is a notable shift. AV1 has better hardware encode support than VP9 ever achieved, largely because GPU manufacturers committed to AV1 collectively through the Alliance for Open Media.
Browser and Platform Support
Browser / Platform
VP9
AV1
Google Chrome
Full support (2014+)
Full support (Chrome 70+)
Mozilla Firefox
Full support (2014+)
Full support (Firefox 67+)
Microsoft Edge
Full support
Full support (Chromium-based)
Apple Safari
macOS only (partial)
Safari 17+ (macOS/iOS)
Android
Full support (5.0+)
Android 12+ (with hardware)
iOS
Limited VP9 (via apps)
iOS 17+ (A17 Pro+ hardware)
YouTube
Primary codec
Used for 4K and efficient delivery
Netflix
Not used
Primary codec for mobile
Twitch
Supported
In rollout (2025+)
Safari's historically limited support for VP9 was a major pain point for web developers. Apple's adoption of AV1 in Safari 17 (September 2023) was a turning point — it means AV1 now has universal browser support across all major platforms.
Browser and device support timeline for VP9 and AV1
Container Format Compatibility
Both VP9 and AV1 can be stored in several container formats:
WebM: The primary container for both codecs on the web. WebM is a subset of MKV designed specifically for VP8/VP9/AV1 video with Vorbis/Opus audio.
MKV (Matroska): Full-featured container that supports both codecs alongside virtually any audio format.
MP4 (ISOBMFF): AV1 is officially supported in MP4 containers (via the av01 codec identifier). VP9 in MP4 is technically possible but rarely used.
For web delivery, WebM is the standard choice for both codecs. Use our video converter or WebM converter to produce WebM files from any source format.
When to Use VP9
Despite AV1's technical superiority, VP9 remains the right choice in several scenarios:
Live streaming and real-time encoding. VP9's encoding is well-optimized and predictable. Real-time VP9 encoding at 1080p60 is achievable on mid-range hardware. AV1 real-time encoding is still emerging and may not meet latency requirements.
Broad device compatibility. VP9 hardware decode is available on more devices, particularly older phones, TVs, and streaming boxes manufactured between 2016-2021. If your audience includes these devices, VP9 ensures smooth playback.
Encoding throughput requirements. If you process thousands of videos per day and encoding time is a bottleneck, VP9's faster encoding pipeline reduces compute costs. At scale, the 3-5x encoding speed difference between VP9 and AV1 translates to significant infrastructure savings.
OBS and screen recording. Many screen recording applications support VP9 output natively. For screen recording format guidance, see our best screen recording format guide.
When to Use AV1
AV1 is the better choice for most new projects and forward-looking workflows:
Maximum compression efficiency. When bandwidth costs matter or your viewers are on constrained connections, AV1's 25-30% bitrate advantage over VP9 is substantial. A 2-hour movie at the same quality might save 1-2 GB compared to VP9.
4K and HDR content. AV1's compression advantages are most pronounced at higher resolutions. For 4K video conversion, AV1 can deliver UHD content at bitrates that would be insufficient for VP9.
Mobile-first delivery. Newer phones with AV1 hardware decode benefit from both the quality improvement and the reduced data usage. Netflix's shift to AV1 for mobile was driven by this exact consideration.
Long-term archival. AV1 will have a longer relevance horizon than VP9. Investing in AV1 encoding now means your content library remains in a modern, efficient format for years.
Apple ecosystem. With Apple's AV1 support in Safari 17 and hardware decode on A17 Pro+ chips, AV1 finally reaches the Apple ecosystem where VP9 was always awkwardly supported.
Practical Encoding Recommendations
Based on real-world testing, here are target settings for common use cases:
VP9 is effectively in maintenance mode. Google has not announced any significant VP9 updates, and development resources are focused on AV1 and its eventual successor.
AV1 continues to gain momentum. Hardware encode support from all three major GPU vendors (Intel, AMD, NVIDIA) means that real-time AV1 encoding at high quality is now practical for streaming, screen recording, and video calls. Browser support is now universal. And the encoder ecosystem (especially SVT-AV1) continues to improve in both speed and quality.
The next generation — AV2 — is already in early development within the Alliance for Open Media, targeting another 30%+ efficiency gain over AV1. But AV2 is years from deployment, and AV1 will be the dominant open codec for the foreseeable future.
For most users producing video content today, AV1 is the recommended codec for new projects, with VP9 serving as a reliable fallback for older device support. Use ConvertIntoMP4's video converter to produce optimized output in either codec, or explore our video compressor to reduce file sizes for distribution. For a broader perspective on format selection, check our best video formats guide and video bitrate explained guide.
