Did the Platform Recompress My Video? Compare and See
The same file plays crisply on your machine and turns mushy after upload — platform’s fault or your parameters? A ten-minute comparison experiment answers it without guessing.
Why platforms recompress at all
Platforms need multiple quality tiers per video and predictable storage/bandwidth costs — whatever you upload gets re-encoded by their encoder. The ceiling of that pass is set by their input bitrate cap: a source far above the cap gets visibly hammered; a source near the cap barely suffers.
So "being recompressed" is not the problem — "being recompressed too hard" is, and the hardness is a function of how your upload parameters match the platform’s tier. Matching starts with knowing the tier (per-platform spec articles cover the numbers).
The comparison experiment
Three steps: before upload, record three parameters of your source with a codec inspection — video codec, average bitrate, resolution; after publishing, download your video from the platform (or use its original-quality download entry if offered); run the same inspection on the downloaded file.
Reading the results: codec downgraded, bitrate halved-plus, resolution dropped — confirmed hard recompression; go back and check your upload parameters against the cap. Bitrate roughly intact, resolution unchanged — recompression was gentle; the softness comes from the source itself. Downloaded bitrate higher than your source — you under-bitrated the upload; raise the source bitrate and pocket free quality.
The fix: pre-compress to the platform’s tier
After confirming hard recompression, the right move is not switching platforms but pre-compressing the source to the platform’s tier yourself: CRF-encode to the target quality so "your controlled, inspectable compression" replaces "their blunt pass" — smaller uploads, milder recompression, better final quality.
The inspect-compare-precompress loop also serves multi-platform distribution: one file per platform tier beats "one 4K high-bitrate file for everywhere" on final quality. The inspection runs locally — parameter comparisons cost no upload bandwidth.