Why videos look worse after you post them

You post a crisp clip and it comes back soft, with mushy detail wherever anything moves. The file was not damaged in transit. It was re-encoded on arrival, and understanding how gives you a fair amount of control over the result.

Every upload is re-encoded

X never serves the file you sent. On arrival, the video is decoded and re-compressed into the platform's standard format at several resolutions.

H.264 is lossy: each encode throws away detail judged least perceptible. Your original was probably already encoded once by your camera or editor. The upload adds a second pass, discarding more, and that second pass has no access to what the first one removed.

This is why quality never improves on upload and cannot be recovered afterwards. Every generation costs something.

Bitrate is what you actually notice

People blame resolution. Bitrate is usually the culprit.

Resolution counts pixels; bitrate decides how much data describes them per second. A 1080p file at 1,500 kbps looks worse than a 720p file at 4,000 kbps, because the 1080p version lacks the data to describe its own pixels.

Platform encoders target bandwidth, not fidelity. If your source was 20,000 kbps and the rendition is 2,500 kbps, seven-eighths of the data is gone. Detail survives in still areas; it collapses wherever there is motion.

What compresses badly

H.264 encodes the difference between frames. Anything defeating that prediction eats budget:

• Fast motion, sports, handheld pans, action. Every frame differs substantially from the last.
• Fine repeating texture, foliage, grass, crowds, confetti, rain.
• Grain and noise, random by definition, so unpredictable and expensive to encode. Low-light footage suffers most.
• Hard cuts, each demands a full keyframe.
• Gradients, sunsets and studio backdrops band into visible steps.

A static talking head at 720p often looks fine. A handheld shot of trees at the same settings falls apart.

Getting a better result

You cannot change the encoder, but you control what it receives.

Match the target resolution

Export at 1920×1080 or 1280×720 exactly. Handing over a 4K file means the encoder downscales and compresses in one pass, and its scaler is tuned for speed. Downscale yourself in an editor first, that scaler is better, and a clean downscale also averages away noise.

Export at a generous bitrate

Do not pre-compress. Give the encoder a high-bitrate source with detail intact. Around 10-15 Mbps for 1080p, 5-8 for 720p. It will compress hard regardless; the question is what it has to work from.

Denoise before exporting

The highest-leverage step for low-light footage. Grain consumes an enormous share of the bitrate budget for no perceptual gain after compression. Mild denoising frees that budget for detail people actually see.

Use H.264 with AAC audio

Matching the platform's own format avoids a needless transcode. Sending H.265 or ProRes forces a conversion before the compression pass even starts.

Keep it under the limit

Exceed the duration or size cap and the file is trimmed or compressed harder to fit. Confirm the current limit for your account tier before exporting.

What this means for downloads

The download is a copy of a rendition, the same file the player uses. Its quality was fixed at upload, so the download neither adds nor removes anything.

Which is why "enhance" or "4K unlock" claims on downloader sites do not hold up. If the best rendition is 720p, that is the ceiling. Taking the highest option, as the HD page does by default, is the whole of the available strategy.

Two useful implications:

• Keep your original. The uploaded version is strictly worse. Archive the export.
• Avoid round trips. Downloading from one platform and re-uploading to another adds another lossy generation. Upload your master to each.