# Fix Video Aspect Ratio Repair a video that plays stretched or squashed by retagging its display aspect ratio locally with ffmpeg stream copy — no upload, no re-encode, no quality loss. ## Run it - **CLI:** `gizza tool video-aspect-ratio-fix 'url=https://example.com/input' 'aspect=16:9' 'custom_aspect=1.85' 'container=keep' 'faststart=true'` - **Web:** https://gizza.ai/tools/video-aspect-ratio-fix/ - **Agents:** machine-readable descriptor (parameters JSON Schema) at https://gizza.ai/tools/video-aspect-ratio-fix/tool.json ## Inputs - `file` — Video _(file; accept: video/*)_ - `aspect` — Display aspect ratio _(field)_ - `custom_aspect` — Custom ratio (only used when aspect = custom) _(field)_ - `container` — Output container _(field)_ - `faststart` — Web fast-start (MP4/MOV: move index to front) _(field)_ ## Output - Aspect-fixed video (video) ## Query parameters Open the tool pre-filled and auto-run via URL: - `aspect` — Display aspect ratio - `custom_aspect` — Custom ratio (only used when aspect = custom) - `container` — Output container - `faststart` — Web fast-start (MP4/MOV: move index to front) - `url` — fetch the input file from a public URL (CORS-permitting) Example: `https://gizza.ai/tools/video-aspect-ratio-fix/?url=https://example.com/input&aspect=16%3A9&custom_aspect=1.85&container=keep&faststart=true` --- ## About this tool Some videos play **stretched, squashed, or anamorphic**: circles look like ovals, faces look narrow or fat, or a widescreen recording shows up letterboxed inside a 4:3 box. Almost always the encoded pixels are perfectly fine — the container is simply telling the player the wrong shape. Cameras that record a squeezed ("anamorphic") frame, DVD/DV rips, screen recorders, and badly configured transcoders all produce files like this. This tool rewrites the **display aspect ratio (DAR) tag** in the container with ffmpeg stream copy (`-map 0 -c copy -aspect W:H`). The file is remuxed with a fresh header carrying the correct ratio, and every audio and video packet is copied bit-for-bit — nothing is decoded, nothing is re-encoded, and the output is normally the same size as the input, byte for byte. Everything runs locally in your browser; the file is never uploaded. Three terms are worth knowing: - **DAR** — display aspect ratio: the shape you actually see, e.g. 16:9. - **SAR/PAR** — sample (pixel) aspect ratio: the shape of a single pixel. Square pixels are 1:1. - The stored frame size in pixels. `DAR = frame width ÷ frame height × SAR`. Because this tool only sets DAR, ffmpeg derives the new SAR for you and the stored frame size never changes. ## Worked example A camera records a 16:9 scene into a 640×480 frame and forgets to flag it, so it plays squashed. Upload `squeezed.mp4`, leave **Display aspect ratio** on `16:9`, keep **Output container** on `keep`, and download `squeezed-aspect-fixed.mp4`. `ffprobe` before and after: ``` before: width=640 height=480 sample_aspect_ratio=1:1 display_aspect_ratio=4:3 after: width=640 height=480 sample_aspect_ratio=4:3 display_aspect_ratio=16:9 ``` The frame is still 640×480 and the file is still 15,349 bytes — identical packets, correct shape. ## Options - **Display aspect ratio** — the DAR to stamp on the file. Presets cover 16:9, 9:16, 4:3, 3:4, 1:1, 21:9, 2.39:1, 1.85:1, 5:4, 4:5, 3:2, and 2:3. Choose `custom` to type your own. - **Custom ratio** — used only when the ratio is set to `custom`. It accepts `16:9`, `16/9`, a plain decimal (`1.85` means width ÷ height), or display dimensions (`1920x1080`). Anything that works out between 0.05 and 20 is allowed; the value is reduced to an exact integer ratio before it is written, so `2.39` becomes `239:100`. - **Output container** — `keep` rebuilds the same container as your input and is the safest lossless choice. `mp4`, `mkv`, `mov`, and `webm` remux into that container instead. This is stream copy, so the codecs must fit the container (H.264/H.265/AAC → mp4/mov/mkv; VP8/VP9/Opus → webm/mkv/mp4). `mkv` is the most tolerant target. - **Web fast-start** — MP4/MOV only. Moves the index (`moov` atom) to the front of the file (`-movflags +faststart`) so players read the new ratio immediately and the clip streams progressively. Ignored for mkv/webm. ## Limits and edge cases - Maximum input size is 64 MB, since everything is processed in your browser. - Accepted inputs are the usual containers ffmpeg can remux: MP4, MOV, M4V, MKV, and WebM. Very exotic containers may need to be remuxed to `mkv` first. - **This never changes pixels.** It cannot add letterbox bars, crop, or rescale. If you want bars or a real resize, use the aspect-pad, crop, or resize video tools instead — those re-encode by necessity. - Quality settings do not apply here: there is no encoder in the pipeline, so there is no bitrate or CRF to choose. - Matroska and WebM store display dimensions as whole pixels, so a ratio like 16:9 can read back as an equivalent-but-not-identical fraction (for example `427:240`, which is 1.7792 against 16:9's 1.7778). MP4/MOV store the exact reduced ratio. - Remuxing into a container the codec does not support (H.264 into `webm`, say) fails by design — use `keep` or `mkv`. - Some old hardware players ignore the container tag entirely and always assume square pixels. Nothing done without re-encoding can fix those; you would have to rescale the pixels. ## FAQ
Does retagging the aspect ratio lose any quality? No. The tool uses stream copy (`-c copy`), so the encoded audio and video packets are copied unchanged into a new container. Nothing is decoded or re-encoded, and the result is bit-for-bit identical apart from the corrected header. That is also why it finishes almost instantly, even on long clips.
What is the difference between this and adding black bars or cropping? This tool changes only **metadata** — the ratio the player is told to display. The stored frame keeps exactly the same pixels and the same width and height. Adding letterbox bars, cropping to a ratio, or stretching to new dimensions all rewrite the actual picture, which means decoding and re-encoding. Use the aspect-pad, crop, or resize video tools for those.
How do I reset a file to square pixels? Choose `custom` and enter the video's **stored pixel size** as the ratio — for a 640×480 file, type `640x480`. That sets DAR equal to the frame's own shape, which is the same as SAR 1:1. There is no ratio-free "make pixels square" switch, because ffmpeg needs a concrete ratio when it is only copying streams.
My video shows 1.7792:1 instead of exactly 16:9 afterwards. Is that wrong? No — that is a Matroska/WebM storage detail. Those containers record display width and height as whole pixels rather than as a fraction, so the ratio is rounded to the nearest whole-pixel pair. The visible difference is under a tenth of a percent. Choose `mp4` or `mov` output if you need the exact reduced ratio in the header.
Which ratio should I pick for social video? `9:16` for vertical full-screen formats such as Reels, Shorts, and TikTok; `1:1` for square feed posts; `4:5` for taller portrait feed posts; `16:9` for YouTube and general landscape playback. If your source was shot anamorphic for cinema, `2.39:1` (scope) and `1.85:1` (flat) are the standard theatrical ratios.
Can I multiply the current ratio instead of setting an absolute one? Not directly — the tool builds its ffmpeg command without first probing the file, so it has no way to read the existing ratio. If you know it, do the arithmetic yourself and enter the result as a custom ratio: an existing 4:3 file that should be 33% wider becomes `4:3 × 1.33`, i.e. roughly `1.78`, which you can enter as `1.78` or `16/9`.
## Related tools - [Fix Video Duration](https://gizza.ai/tools/video-duration-fix-remux/): Repair a video whose duration shows as Infinity, 0:00, or the wrong length by remuxing it locally with ffmpeg stream copy — no upload, no re-encode, no quality loss. - [Remove Embedded Subtitles From Video](https://gizza.ai/tools/video-remove-embedded-subs/): Strip all soft subtitle and caption streams from a video locally with ffmpeg stream copy — keep the video and audio intact, no upload and no re-encode. - [Strip Video Metadata](https://gizza.ai/tools/video-strip-metadata/): Remove GPS, device, timestamp, title, comment, stream, and chapter metadata from a video locally with ffmpeg stream copy — no upload and no re-encode. - [Change Video Speed](https://gizza.ai/tools/change-speed/): Speed up or slow down a video right in your browser, keeping audio in sync — pick a speed factor (0.25x–4x). Re-encodes locally with ffmpeg, nothing is uploaded, free. - [Extract Video Frames to a Contact Sheet](https://gizza.ai/tools/extract-frames/): Sample frames from a video by interval, fps or scene change and tile them into one contact-sheet image. Free, in your browser with ffmpeg, nothing is uploaded.