Create Video Contact Sheets with FFmpeg Tile Filter

Generating a video contact sheet or thumbnail grid is a highly efficient way to preview video content at a glance. This guide explains how to configure and use the tile video filter in FFmpeg to create custom contact sheets, detailing the essential parameters for frame selection, scaling, and grid layout.

The Basic Command Structure

To create a contact sheet, FFmpeg needs to extract frames from a video, optionally resize them, and then arrange them into a grid.

Here is the standard command template:

ffmpeg -i input.mp4 -vf "fps=1/10,scale=320:-1,tile=4x3" output.png

How the Filters Work

The key to creating a clean contact sheet lies in the video filter (-vf) chain. The filters are applied in order from left to right, separated by commas.

1. Frame Selection (fps or select)

By default, FFmpeg processes every single frame. If you pass every frame to the tile filter, your contact sheet will only represent a fraction of a second of video. You must limit the input frames: * Time-based (fps): Use fps=1/10 to capture one frame every 10 seconds, or fps=1/60 for one frame every minute. * Frame-based (select): Use select='not(mod(n\,100))' to select every 100th frame.

2. Thumbnail Scaling (scale)

If your source video is 1080p or 4K, placing those frames directly into a grid will result in an enormous image file. * scale=320:-1 resizes each thumbnail to a width of 320 pixels and automatically calculates the height to maintain the original aspect ratio.

3. Grid Layout (tile)

The tile filter defines the layout of the final contact sheet using the format columnsxrows. * tile=4x3 creates a grid of 4 columns and 3 rows (12 thumbnails total). * Once the grid is full, FFmpeg will output the image. If the video is long enough to fill the grid multiple times, FFmpeg will generate multiple output files (e.g., output_%03d.png).

Advanced Configuration Options

The tile filter accepts additional parameters to customize the look of the contact sheet. Parameters are separated by colons.

ffmpeg -i input.mp4 -vf "fps=1/10,scale=320:-1,tile=4x3:nb_frames=12:padding=10:color=white" output.png

• nb_frames: Limits the maximum number of frames rendered in the grid. Setting nb_frames=12 in a 4x3 grid ensures the process stops immediately once the grid is filled.
• padding: Adds a border (in pixels) around each thumbnail.
• margin: Adds an outer border (in pixels) around the entire contact sheet.
• color: Defines the color of the padding and margin area (e.g., white, black, red, or hex codes like 0xEEEEEE).

Real-World Example

If you want to generate a single 5x5 grid (25 thumbnails total) from a 5-minute video, you need a frame roughly every 12 seconds (\(300 \text{ seconds} / 25 = 12\)).

ffmpeg -i input.mp4 -vf "fps=1/12,scale=240:-1,tile=5x5:padding=5:color=black" contact_sheet.jpg

This command extracts a frame every 12 seconds, scales them to 240px wide, arranges them in a 5x5 grid with a 5-pixel black border, and saves the output as a JPG file.