doc update

This commit is contained in:
minjaesong
2025-11-04 00:43:14 +09:00
parent 9d98cc1a21
commit 61b0bdaed7

View File

@@ -677,7 +677,7 @@ TSVM Enhanced Video (TEV) Format
Created by CuriousTorvald and Claude on 2025-08-17
TEV is a modern video codec optimized for TSVM's 4096-color hardware, featuring
DCT-based compression, motion compensation, and efficient temporal coding.
DCT-based compression, optional motion compensation, and efficient temporal coding.
## Version History
- Version 2.0: YCoCg-R 4:2:0 with 16x16/8x8 DCT blocks
@@ -1091,7 +1091,7 @@ This packet contains multiple frames encoded as a single spacetime block for opt
temporal compression.
uint8 Packet Type (0x12/0x13)
uint8 GOP Size (number of frames in this GOP, typically 16)
uint8 GOP Size (number of frames in this GOP)
<if packet type is 0x13>
uint32 Compressed Size
* Zstd-compressed Motion Data
@@ -1122,16 +1122,15 @@ This layout enables Zstd to find patterns across both spatial and temporal dimen
resulting in superior compression compared to per-frame encoding.
### Motion Vectors
- Stored in 1/16-pixel units (divide by 16.0 for pixel displacement)
- Used for global motion compensation (camera movement, scene translation)
- Computed using FFT-based phase correlation for accurate frame alignment
- Stored in 1/4-pixel units (divide by 4.0 for pixel displacement)
- Computed using dense optical flow
- Cumulative relative to frame 0 (not frame-to-frame deltas)
- First frame (frame 0) always has motion vector (0, 0)
### Temporal 3D DWT Process
1. Detect inter-frame motion using phase correlation
2. Align frames and expand canvas to preserve all original pixels
3. Apply 1D DWT across temporal axis (GOP frames) on expanded canvas
1. Detect where the scene change is happening on the first pass
2. Determine GOP slicing from the scene detection
3. Apply 1D DWT across temporal axis (GOP frames)
4. Apply 2D DWT on each spatial slice of temporal subbands
5. Perceptual quantization with temporal-spatial awareness
6. Unified significance map preprocessing across all frames/channels