Source-Channel Separation Theorems for Distortion Perception Coding
Chao Tian 1 , Jun Chen 2 , and Krishna Narayanan 3 · 2025
Read Source-Channel Separation Theorems for Distortion Perception Coding by Chao Tian 1 , Jun Chen 2 , and Krishna Narayanan 3 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-17706.txt.
Description
Chao Tian, Jun Chen, and Krishna Narayanan revisit a foundational question in information theory in this two thousand twenty-five paper. Classic theory says separating lossy source coding from channel coding loses nothing, at least for classical distortion measures. Newer work judges reconstructions also by perceptual quality or realism, which connects to generative models and semantic communication. The authors ask whether separation stays optimal under this distortion-perception standard.
The answer splits by how perception is measured. When quality is judged per block, separation remains optimal if encoder and decoder share common randomness, but fails without it. When quality is the average per-symbol measure, separation stays optimal either way. The results map where a convenient engineering shortcut remains safe in perception-aware coding systems.
Catalog record
| CREATOR | Chao Tian 1 , Jun Chen 2 , and Krishna Narayanan 3 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-17706