Configure FFmpeg Floodfill Coordinates and Tolerance

The FFmpeg floodfill filter allows you to replace a region of connected pixels of a similar color with a new color, starting from a specific seed point. This guide provides a straightforward explanation of how to configure the starting coordinates and control the color tolerance using the filter’s parameters, complete with practical command examples.

Starting Coordinates

To set the starting point of the flood fill, use the x and y parameters. These parameters define the pixel coordinates on the input video where the filling operation begins.

• x: The horizontal coordinate of the starting pixel (default is 0).
• y: The vertical coordinate of the starting pixel (default is 0).

For example, to start the flood fill at the center of a 1920x1080 video, you would set x=960 and y=540.

Color Tolerance

Color tolerance determines how closely adjacent pixels must match the starting pixel’s color to be included in the fill area. FFmpeg controls this using the d0, d1, d2, and d3 parameters, which correspond to the color channels of your video’s pixel format.

• RGB Video: d0 represents Red, d1 represents Green, d2 represents Blue, and d3 represents the Alpha (transparency) channel.
• YUV Video: d0 represents Luma (Y), d1 and d2 represent Chroma (U and V), and d3 represents Alpha.

The tolerance values range from 0 to 255 (default is 0): * A value of 0 means adjacent pixels must be an exact color match to be filled. * Higher values allow the fill to spread over pixels with greater color variation.

Command Example

Below is a practical FFmpeg command that applies the floodfill filter. It sets the starting coordinates to x=300 and y=400, specifies a red fill color, and sets a moderate color tolerance of 15 across the first three color channels to allow the fill to spread across slightly varying shades.

ffmpeg -i input.mp4 -vf "floodfill=x=300:y=400:fillcolor=red:d0=15:d1=15:d2=15" output.mp4

Parameter Reference Table