When you upload a video to X, what comes back out is not the file you put in. X runs it through its own processing pipeline before anyone else can see it, and that pipeline is optimized for X's bandwidth bill, not your footage.
Two passes, not one
X transcodes video twice. The first pass happens right after upload, producing a fast, low-quality preview so the post is watchable within seconds. The second pass runs in the background over the next several minutes, producing the higher-bitrate versions that eventually replace the preview — the same 1080p/720p/360p ladder you see in this tool's quality picker. If you download immediately after posting, you may be catching the first-pass file; wait a few minutes and re-check if a post looks unexpectedly soft.
A hard ceiling on bitrate
Regardless of your source file's bitrate, X caps the output. A crisp 4K clip shot on a recent phone can carry 80–100 Mbps; X's top video tier tops out far below that. The compressor isn't being lazy — it's a deliberate tradeoff between visual quality and the cost of serving billions of video views a day. The result is the same for everyone: what you get back is your footage plus X's compression, not your footage as shot.
Why the audio sounds thinner too
Audio gets the same treatment — down-mixed and encoded at a fixed, modest bitrate regardless of what you uploaded. It's rarely the first thing people notice, but it's part of the same budget.
What this means for downloading
When you pull a video back down through a tool like this one, you're getting X's compressed version — the only version that exists on their servers. There's no way to recover the pre-upload quality after the fact; the original left your device the moment you hit post. The practical upshot: archive from the source (your camera roll) when quality matters, and treat X as a distribution copy, not a backup.