Transcoding

Turning one uploaded video into the several quality versions an adaptive stream needs — the step that makes smooth playback possible for every viewer.

Updated: May 2026 5 min read Encoding
In short

Transcoding turns one uploaded file into a ladder of resolutions and bitrates, so a single video can play smoothly on a phone on 3G and a TV on fibre alike.

What is transcoding?

Transcoding is the process of re-encoding an already-encoded video into different versions — changing the codecCodecThe algorithm that compresses and decompresses video — H.264, H.265 (HEVC) and AV1 are the most common in modern streaming.Open definition →, resolutionResolutionThe pixel dimensions of a video frame — 480p, 720p, 1080p, 2160p (4K). Determines how sharp the picture looks on a given screen.Open definition →, or bitrateBitrateThe amount of data used per second of video. Higher bitrate means more detail and a larger file; the core lever behind streaming quality.Open definition →. In streaming, it means taking one source upload and producing several renditions of it, so a player can switch between quality levels as network conditions change.

It's easy to confuse with encoding. Encoding compresses raw video into a format for the first time; transcoding takes that compressed file and re-encodes it into new variants. When you upload a finished MP4 to a video platform, what happens next is transcoding.

Without transcoding, you could only deliver a single fixed quality — which buffers on slow connections and looks soft on fast ones. It is the step that makes adaptive bitrateAdaptive Bitrate (ABR)Automatically adjusts video quality in real time to match each viewer’s connection speed — smooth playback from 480p to 4K with no buffering.Open definition → streaming possible.

How transcoding works

A streaming transcode turns one file into a ready-to-stream package: decode the source, build a ladder of renditions, segment and package it, and publish to the edge.

The transcoding pipeline

Transcoding sits between the upload and the player: it produces every rendition the adaptive stream will switch between.

Stage by stage:

Decode the source

The uploaded file is decoded back to raw frames so it can be re-encoded cleanly. Working from the original source (rather than an already-compressed rendition) keeps quality high.

Build the rendition ladder

Each rung is re-encoded at a target resolution and bitrate — for example 1080p, 720p, 480p and 360p. A 4K4K (2160p)Ultra HD at 3840×2160 — four times the pixels of Full HD, and the point where codec choice starts paying for itself.Open definition → source adds a 2160p rung; the ladder is tuned to the source so nothing is upscaled.

Segment and package

Each rendition is cut into short segments and packaged for delivery protocols — HLSHLS (HTTP Live Streaming)Apple’s adaptive streaming protocol. Uses .m3u8 manifests and segmented delivery; supported on every browser, device and smart TV.Open definition → and DASHDASHDynamic Adaptive Streaming over HTTP — an open ISO standard using .mpd manifests, common for DRM-protected premium content.Open definition → — with a manifestManifestA small index file (.m3u8 or .mpd) that lists every available rendition and segment so the player knows what it can request.Open definition → that lists every rendition and segment.

Publish to the CDN

The packaged renditions are pushed to a CDNCDN (Content Delivery Network)A global network of edge servers that caches video close to viewers, cutting latency and absorbing traffic spikes.Open definition → so viewers stream from an edge server near them. Parallel transcoding means playback can start on the first ready rendition while higher ones finish.

Encoding vs. transcoding

AspectEncodingTranscoding
InputRaw / uncompressed videoAlready-encoded file
PurposeCompress for the first timeRe-encode into new variants
OutputOne compressed fileMultiple renditions / formats
Where it happens in streamingCamera, editor, exportOn the platform, after upload

Each rung of the ladder is a full re-encode at its own bitrate. Get a feel for how storage and bandwidth scale across the ladder:

Per-rendition size & bandwidth

Resolution
Codec

Estimated size

Connection needed

A single-rendition estimate at the recommended bitrate. Real adaptive delivery stores the whole ladder — Kinescope builds and serves it for you.

Further reading

Denis Konnov

Written by

Denis Konnov Verified

CMO at Kinescope

Denis leads marketing and content at Kinescope, working closely with the platform's engineers. This glossary is reviewed and kept up to date by the team building Kinescope's streaming, encoding and CDN delivery infrastructure.

FAQ

Can’t find the answer you need? Check out our Help Center.

Encoding turns raw or uncompressed video into a compressed format for the first time. Transcoding takes an already-encoded file and re-encodes it — into a different codec, resolution, or bitrate. In streaming, "transcoding" usually means producing the multiple renditions an adaptive stream needs from a single source upload.

Adaptive streaming needs several versions of the same video at different resolutions and bitrates so the player can switch quality to match each viewer’s connection. Transcoding produces that rendition ladder. Without it you could only serve one fixed quality, which buffers on slow connections and looks soft on fast ones.

A transcoding (or "bitrate") ladder is the set of renditions generated from one source — for example 1080p, 720p, 480p and 360p, each at a target bitrate. The player picks a rung of the ladder based on real-time bandwidth. Platforms tune the ladder to the source resolution to avoid wasting storage.

Each re-encode is lossy, so a poorly configured transcode can lose quality. Done well — with modern codecs and sensible target bitrates — the loss is imperceptible, and the gain (smooth adaptive playback for every viewer) far outweighs it. Kinescope transcodes from your original source to keep quality high.

It depends on length, resolution and how many renditions are produced, but modern platforms transcode in parallel so playback can start before the whole job finishes. On Kinescope most files are embed-ready within minutes, with higher resolutions finishing in the background.

No. Kinescope transcodes on its side — upload MP4, MOV, MKV and most common formats up to 4K, and the platform builds the full adaptive ladder, writes the manifests and delivers via CDN. No pre-processing on your machine.