Fundamentals
What Happens to Your Video When You Upload It
Every platform re-encodes what you send. Knowing what that process does explains why your upload looks softer than your export, and what actually helps.
3 min read · Published 23 August 2026
You export something that looks excellent, upload it, and the version on the platform looks softer. Detail in the shadows has gone, fast movement smears, gradients have banded. Nothing was done to your file in transit. It was simply replaced.
Your file is not what gets served
No large platform serves the file you uploaded. Yours is a source, and from it the platform produces its own set of versions: several resolutions, often several codecs, each at a bitrate it chose.
It has to. A single file cannot serve a phone on a weak connection and a television on fibre, and the platform is paying for every byte it sends to millions of people. Your upload is raw material.
Why the result is softer
Two reasons, and only one of them is about bitrate.
The first is generational loss. Your export was already compressed, which means it already contains the compromises your encoder made. The platform decodes that and compresses it again, and the second encoder cannot tell your compression artefacts apart from real detail, so it spends data preserving them.
The second is that the platform's target bitrate is usually below yours, because it is optimising cost across an enormous library rather than quality for one video.
| Where the quality goes | What you can do |
|---|---|
| Platform re-encodes at its own bitrate | Nothing, this is not adjustable |
| Your compression artefacts get preserved as detail | Upload at a generous bitrate so there are fewer |
| Resolution ladder rebuilt from your source | Upload at the resolution you want served, not above |
| Colour metadata dropped or misread | Tag the export properly |
The one thing that genuinely helps
Give the platform's encoder a clean source. That means a higher bitrate than the platform will use, because every artefact in what you send becomes something the second encoder wastes data on.
This feels wasteful and is not. The upload bandwidth is spent once; the quality of every version served afterwards depends on it. Uploading a 1080p file at 20 Mbps when the platform will serve 8 does not mean the viewer gets 20. It means the 8 is spent on your picture rather than on your compression.
Uploading a smaller file to be considerate achieves the opposite. The platform re-encodes regardless, and a compressed source produces a worse result at exactly the same delivered bitrate.
Resolution: upload what you want served
Platforms allocate bitrate by resolution tier, and the higher tiers usually get a disproportionately generous allowance. Uploading 4K when your source is genuinely 1080p is a common trick to access that allowance.
It sometimes works and it is not free. Upscaling adds no detail, the file is several times larger to upload, and some platforms detect and penalise it. Try it if you like, but know that you are gaming a bitrate ladder rather than improving your video.
What to check afterwards
- Look at the platform's version at full size, not in a small preview. Compression artefacts hide at small sizes.
- Check the gradients: skies, walls, anything smooth. Banding is the first thing to appear when bitrate is short.
- Check fast motion. Smearing there is the encoder running out of budget where it needs it most.
- Check the colour. Washed out or shifted usually means metadata was dropped rather than anything to do with bitrate.
A sensible upload recipe
- Export at the resolution you want served, without upscaling.
- Use a generous bitrate, comfortably above what the platform delivers.
- H.264 unless the platform specifically prefers otherwise, since it is what every ingest pipeline handles best.
- AAC audio at 192 kbps, which is a small cost and one fewer thing to go wrong.
- Tag the colour space, so the platform does not have to guess.
- Keep your master. The upload is a copy, and one day you will need to produce a different version.
Common questions
Why does my video look worse after uploading?
The platform re-encodes it at its own bitrate, and that second encoder also has to spend data preserving the compression artefacts already in your file. Both effects reduce quality and neither is under your control after upload.
Does uploading a bigger file improve the result?
Yes, up to a point, because a cleaner source means the platform's encoder spends its budget on your picture rather than on your artefacts. It does not increase what viewers are served, it improves what that amount of data represents.
Should I upload in 4K when my video is 1080p?
It sometimes gains a more generous bitrate tier, but upscaling adds no real detail, costs a much larger upload, and some platforms detect it. It is gaming the ladder rather than improving the video.
Which codec should I upload in?
H.264 unless the platform states otherwise. Every ingest pipeline handles it well, and any efficiency you gain from a newer codec is discarded the moment the platform re-encodes anyway.