FFmpeg: Hide Banner for Specific Y-Position in MP4 with Quality, Size, and Frame Rate

To hide a banner at a specific Y-position in an MP4 video while maintaining custom quality, size, and frame rate settings, use FFmpeg’s drawtext or overlay filters combined with geometric transformations. The most robust method involves cropping the video area below the banner, scaling to the desired resolution, and re-encoding with precise parameters for bitrate and frame rate.

Understanding the Geometry of Banner Removal

Before applying any filters, you must determine the exact pixel coordinates where the banner resides. Most digital videos have a standard aspect ratio, but banners often span the full width at either the top (positive Y) or bottom (negative Y relative to the end). If your goal is to remove a static graphic overlay that was burned into the video file, simple cropping is the most efficient approach because it does not require complex content-aware fill algorithms which are computationally expensive and often produce artifacts. The key variable here is the height of the banner. You need to know how many pixels from the top or bottom edge constitute the "banner zone." For example, if your video is 1080p high and the banner occupies the top 50 pixels, you will crop the top 50 pixels off. This reduces the total height of the video frame. Understanding this geometry is crucial before moving to scaling, as scaling happens after cropping to ensure the final output dimensions match your target size exactly. For more complex scenarios where you might want to manipulate frames differently, such as removing specific intervals to smooth out playback, you can refer to guides on How to Delete Every 25th Frame in FFMPEG for Smoother Video Playback. While this technique removes temporal redundancy rather than spatial elements, it highlights FFmpeg's flexibility in frame manipulation.

Cropping the Video Stream

The first step in the pipeline is to remove the unwanted visual data using the crop filter. This filter takes four arguments: width, height, x-offset, and y-offset. To remove a banner from the top of the video, you keep the original width but reduce the height by the banner's pixel count, setting the y-offset to the height of the banner so that FFmpeg starts capturing frames from below the banner. Assume your source video is 1920x1080 and the banner is 100 pixels tall at the top. The command would look like this:

-vf "crop=1920:980:0:100"

Here, 1920 is the new width, 980 is the new height (1080 - 100), 0 is the horizontal offset (starting from the left edge), and 100 is the vertical offset (skipping the top 100 pixels). If the banner were at the bottom, the y-offset would be 0, and you would subtract the banner height from the total height again, effectively keeping the top portion of the video. This operation is lossless regarding the remaining pixels; no data is lost from the visible area, preserving maximum quality for the subsequent encoding stage. If you need to adjust the timing or synchronization of these operations, understanding how FFmpeg handles timestamps is vital. Incorrect PTS handling can lead to audio desync or stuttering, so consider reading about Incorrect pts filename with ffmpeg's -ss and -frame_pts true to avoid common timestamp pitfalls during complex filter chains.

Scaling to Target Size and Quality

After cropping, the video dimensions have changed. If your requirement specifies a particular output size (e.g., 1280x720), you must apply the scale filter immediately after the crop filter. It is important to maintain the aspect ratio unless the target format requires a specific non-standard dimension. To scale the cropped video to 1280x720, you append the scale filter to your chain:

-vf "crop=1920:980:0:100,scale=1280:720"

Quality control is primarily managed through the video codec and bitrate settings rather than the filter itself. When using H.264 or H.265 codecs, you should specify a constant rate factor (CRF) or a target bitrate. A lower CRF value (e.g., 18-23 for H.264) results in higher quality but larger file sizes. For strict size constraints, calculate the bitrate based on the duration and target file size.

-c:v libx264 -crf 20 -preset medium

This ensures that the visual fidelity remains high even after the geometric transformations. The medium preset offers a balanced trade-off between encoding speed and compression efficiency.

Setting Frame Rate and Final Output

Frame rate consistency is critical for smooth playback. If the source video has a variable frame rate (VFR), it is best practice to convert it to a constant frame rate (CFR) using the fps filter in the video filter chain.

-vf "crop=1920:980:0:100,scale=1280:720,fps=30"

This forces the output to exactly 30 frames per second. Combining this with the previous steps creates a complete processing pipeline. You also need to ensure the audio stream is handled correctly, typically by copying it without re-encoding to save time and preserve audio quality:

-c:a copy

Finally, specify the output format as MP4. Note that some users attempt to concatenate multiple processed clips afterward. If your workflow involves merging several banner-free segments, ensure they share the same codec and parameters to avoid compatibility issues, as detailed in How to Concatenate Videos with FFmpeg. The complete command structure looks like this:

ffmpeg -i input.mp4 -vf "crop=1920:980:0:100,scale=1280:720,fps=30" -c:v libx264 -crf 20 -c:a copy output.mp4

This command crops the top banner, scales to HD, enforces 30 FPS, encodes with high-quality H.264, and copies the audio, resulting in a clean, properly sized MP4 file.

Frequently Asked Questions

Can I blur the banner instead of cropping it?

Yes, you can use the boxblur filter applied to a specific region defined by the select filter or by using the drawrect filter to cover the area. However, blurring leaves black or blurred artifacts where the banner was, whereas cropping physically removes the pixels and allows scaling to fill the space seamlessly.

What if the banner is semi-transparent?

Cropping will still work perfectly fine because it ignores the alpha channel and simply cuts the pixel data. Semi-transparency only matters if you are trying to composite a new background behind the banner, which requires advanced inpainting tools not natively available in standard FFmpeg filters.

Will changing the frame rate cause audio desync?

No, the fps filter only affects the video stream. As long as you use -c:a copy to pass through the original audio stream without modification, the audio timeline remains independent of the video frame rate changes.

How do I find the exact pixel height of my banner?

You can open the video in a media player like VLC and pause on a clear frame. Use the player's screenshot feature or a dedicated image editor to measure the banner's height in pixels. Alternatively, use FFmpeg's vhook or external scripts to analyze the average color change across the vertical axis to detect the banner boundaries automatically.

Is it better to crop or use a mask overlay?

Cropping is superior for reducing file size and simplifying the pipeline because it reduces the number of pixels that need to be encoded. Mask overlays require additional layers and often result in larger files due to the extra data required to encode the transparent or masked regions.