How to Scale Video in FFmpeg Without Upscaling

This guide explains how to resize videos using the FFmpeg scale filter while ensuring that smaller source videos are not stretched or upscaled. You will learn the specific FFmpeg commands and expressions needed to scale down large videos to a target resolution while preserving the original dimensions of smaller files to maintain their visual quality.

The Challenge with Standard Scaling

By default, if you use a standard FFmpeg scaling command like -vf scale=1280:720, FFmpeg will stretch any input video to match those exact dimensions. If your source video is smaller than 1280x720 (for example, 640x360), FFmpeg will upscale it, resulting in pixelation, blurriness, and unnecessary file size inflation.

The Solution: Using the `min()` Expression

To prevent upscaling, you must use FFmpeg’s internal math evaluation tools. By using the min() function inside the scale filter, you can instruct FFmpeg to choose the smaller value between your target resolution and the input video’s actual resolution.

Here is the basic command structure:

ffmpeg -i input.mp4 -vf "scale='min(1280,iw)':'min(720,ih)'" output.mp4

How This Command Works:

iw and ih: These variables represent the Input Width and Input Height of the source video.
min(1280,iw): FFmpeg compares 1280 and the input width, then uses the smaller of the two values. If the input is 1920 pixels wide, it scales down to 1280. If the input is 640 pixels wide, it remains 640.
min(720,ih): This does the same for the height, choosing the smaller value between 720 and the input height.

Scaling While Preserving Aspect Ratio

Using the command above might distort the aspect ratio if the input video is not of the same proportions (e.g., a square video). To scale the video to a maximum width of 1280 pixels, preserve the original aspect ratio, and prevent upscaling, use the following command:

ffmpeg -i input.mp4 -vf "scale='min(1280,iw)':-2" output.mp4

Why Use `-2`?

Using -2 for the height tells FFmpeg to automatically calculate the correct height based on the width while ensuring the resulting dimension is divisible by 2. This is crucial because many modern video codecs (such as H.264/libx264) require even pixel dimensions to encode successfully.

If the input video is larger than 1280 pixels wide, it will scale down to 1280 pixels wide with a proportionally scaled height. If the input video is smaller than 1280 pixels wide, it will bypass scaling entirely and retain its original resolution.