Learn To: Use Capped CRF with SVT-AV1 for Live Streaming

Related Articles

Deep Render: An AI Codec That Encodes in FFmpeg, Plays in VLC, and Outperforms SVT-AV1

May 8, 2025

Evaluating DCVC-RT: A Real-Time Neural Video Codec That Delivers on Speed and Compression

May 2, 2025

Google’s Cookie Reversal: While Advertisers Rejoice, Legal Trouble Lurks in the Weeds

April 24, 2025

HEVC Licensing: Misunderstood, Maligned, and Surprisingly Successful

April 22, 2025

Learn to use Capped CRF with SVT-AV1 and FFmpeg for live transcoding, including how capped CRF compares to VBR and CBR. All tests and results are available in downloadable PDF.

Capped CRF is an alternative to VBR and CBR that saves bandwidth on easy-to-encode sequences and preserves quality on hard-to-encode sequences. Our tests with SVT-AV1 and FFmpeg show:

• Performance: Capped CRF delivered greater throughput than VBR, though this varied by content type and logical processors deployed.
• Bitrate savings: Bitrate savings averaged about 44% over our five categories. For reference, the cap and VBR target for 60fps sports was 6 Mbps; otherwise, it was 4.5 Mbps.
• Average quality: The average VMAF score dropped 2.1% but still averaged 94.41, which is in the relevant target zone for top-rung quality for premium content producers.
• Low-frame quality: Low-frame quality is a predictor of transient quality problems. Capped CRF decreased the low-frame quality by 13.5%, though most of this was in sports-related footage.

Overall, all live-streaming producers should consider capped CRF as an alternative to VBR. Here’s the report that details our findings.