Transcoding
Transcoding is the direct digital-to-digital conversion of encoded media content from one format, bitrate, or resolution to another, decoding the original stream and re-encoding it in a target format for device or channel compatibility.
What Is Transcoding?
Transcoding is the direct digital-to-digital conversion of encoded media content from one format, bitrate, or resolution to another, typically to achieve compatibility with a target device or delivery channel. The process first decodes the original compressed stream back to an intermediate uncompressed representation, then re-encodes it in the target format. Transcoding applies to video, audio, image, and character data, though its most computationally intensive and widely studied form involves video. Because the intermediate representation discards format-specific compression artifacts, transcoding introduces generation loss unless lossless codecs are involved at both stages.
Transcoding sits at the intersection of data compression, multimedia communication, and signal processing. The field draws from rate-distortion theory, which characterizes the relationship between compressed file size and the fidelity of the reconstructed signal, and from codec standards developed by bodies such as ITU-T, ISO/IEC, and the Alliance for Open Media.
Video Coding and Codec Conversion
Video transcoding frequently involves conversion between codec standards. A media file encoded in H.264/AVC may need to be transcoded to H.265/HEVC or AV1 to reduce bandwidth consumption on streaming platforms, or to a legacy codec for compatibility with older playback devices. Each major codec generation achieves roughly a 40 to 50 percent bitrate reduction at equivalent quality compared to its predecessor, making transcoding to a newer codec an effective way to lower storage and delivery costs.
Compressed-domain transcoding techniques reduce computational load by operating partly in the compressed domain rather than fully decoding each frame. Motion vectors estimated during decoding can be reused or adapted for re-encoding, and discrete cosine transform (DCT) coefficients can sometimes be manipulated directly. Research published in IEEE Transactions on Multimedia has examined rate control algorithms designed specifically for transcoding, where the encoder must match a target bitrate without access to the original uncompressed source. IEEE Xplore papers on universal rate control for video transcoding describe methods that adapt quantization parameters frame-by-frame to hit a delivery bitrate target.
Image Coding and Format Interoperability
Image transcoding converts between still-image formats such as JPEG, JPEG 2000, WebP, and AVIF. Each format reflects different tradeoffs between compression efficiency, decoding complexity, and feature support. JPEG remains dominant on the web due to broad decoder availability, while newer formats such as AVIF offer substantially better quality-to-size ratios. Transcoding a large archive from JPEG to a modern format reduces storage and bandwidth at the cost of one-time processing.
Spatial resolution changes are common in image transcoding for responsive web design, where a single high-resolution source image must be adapted to multiple display sizes and pixel densities. Transcoding pipelines in content delivery infrastructure routinely generate dozens of resolution and format variants from a single uploaded master, a process sometimes called image transformation or image resizing at the application layer.
Adaptive Bitrate Streaming and Multimedia Delivery
In adaptive bitrate (ABR) streaming, a source video is transcoded into multiple representations at different bitrates and resolutions, with each representation segmented into short chunks. The player client selects the representation whose bitrate best matches the available network throughput, switching among representations at segment boundaries to maintain smooth playback. HLS and DASH are the dominant ABR delivery protocols, and cloud-based transcoding services now process content at scale using GPU-accelerated encoder pipelines.
Transcoding is computationally intensive relative to decoding alone, and hardware accelerators including GPU encoder units and application-specific integrated circuits (ASICs) are used to meet throughput requirements in production. The ScienceDirect overview of video compression standards provides context on the codec landscape that transcoding pipelines must navigate.
Applications
Transcoding has a central role across many multimedia and communications contexts, including:
- Adaptive bitrate streaming for over-the-top video services
- Broadcast contribution and distribution workflows converting between production and delivery formats
- Video conferencing systems adapting resolution and bitrate to participant bandwidth
- Content archival and format migration in media libraries
- Real-time image format conversion in web content delivery networks