Configure VP9 Dynamic Scaling in FFmpeg

This article explains how to configure the dynamic scaling (spatial resampling) option in the libvpx-vp9 encoder using FFmpeg. You will learn the specific command-line parameters required to enable this feature, set target bitrate thresholds, and optimize your video streams for fluctuating bandwidth conditions.

Dynamic scaling in libvpx allows the encoder to automatically downscale the video resolution on the fly when the available bandwidth drops, and scale it back up when network conditions improve. This is particularly useful for real-time streaming and WebRTC applications where maintaining a consistent frame rate is more critical than maintaining full resolution.

Key FFmpeg Parameters for Dynamic Scaling

To configure dynamic scaling in the libvpx-vp9 encoder, you need to use the following flags:

• -resize-allowed 1: This parameter enables spatial resampling (dynamic scaling). Set it to 1 to turn the feature on, or 0 (the default) to disable it.
• -resize-down-threshold <percentage>: This defines the threshold for scaling down. It is represented as a percentage of the target bitrate. If the encoder’s buffer falls below this percentage, the encoder reduces the frame size. For example, a value of 30 means the encoder will downscale if the bandwidth drops below 30% of the target bitrate.
• -resize-up-threshold <percentage>: This defines the threshold for scaling back up. When the buffer capacity exceeds this percentage of the target bitrate, the encoder scales the video back up toward its original resolution.

Example FFmpeg Command

Below is a practical example of an FFmpeg command that encodes a video using VP9 with dynamic scaling enabled:

ffmpeg -i input.mp4 -c:v libvpx-vp9 -b:v 1M -resize-allowed 1 -resize-down-threshold 30 -resize-up-threshold 60 output.webm

Command Breakdown

• -c:v libvpx-vp9: Specifies the VP9 encoder.
• -b:v 1M: Sets the target video bitrate to 1 Megabit per second.
• -resize-allowed 1: Enables the dynamic resolution scaling feature.
• -resize-down-threshold 30: Instructs the encoder to scale down the resolution if the bitrate allocation efficiency falls below 30%.
• -resize-up-threshold 60: Instructs the encoder to scale the resolution back up once bandwidth recovery clears the 60% threshold.

By utilizing these settings, you ensure that your output stream adapts dynamically to network constraints, preventing frame drops and buffering at the cost of temporary resolution reductions.