mirror of
https://github.com/curioustorvald/tsvm.git
synced 2026-06-14 16:34:04 +09:00
working YCoCg-R variant
This commit is contained in:
@@ -698,17 +698,17 @@ DCT-based compression, motion compensation, and efficient temporal coding.
|
||||
byte[5] Reserved
|
||||
|
||||
## Packet Types
|
||||
0x10, 0x00: I-frame (intra-coded frame)
|
||||
0x11, 0x00: P-frame (predicted frame)
|
||||
0x20, 0x00: MP2 audio packet
|
||||
0xFF, 0xFF: sync packet
|
||||
0x10: I-frame (intra-coded frame)
|
||||
0x11: P-frame (predicted frame)
|
||||
0x20: MP2 audio packet
|
||||
0xFF: sync packet
|
||||
|
||||
## Video Packet Structure
|
||||
uint16 Packet Type
|
||||
uint8 Packet Type
|
||||
uint32 Compressed Size
|
||||
* Zstd-compressed Block Data
|
||||
* Gzip-compressed Block Data
|
||||
|
||||
## Block Data (per 8x8 block)
|
||||
## Block Data (per 8x8/16x16 block)
|
||||
uint8 Mode: encoding mode
|
||||
0x00 = SKIP (copy from previous frame)
|
||||
0x01 = INTRA (DCT-coded, no prediction)
|
||||
@@ -731,15 +731,13 @@ of 8, while AC coefficients are quantized according to quality tables.
|
||||
- Uses Sum of Absolute Differences (SAD) for motion estimation
|
||||
- Bilinear interpolation for sub-pixel motion vectors
|
||||
|
||||
## Color Space
|
||||
TEV operates in native 4096-color mode (4:4:4 RGB, 4 bits per channel).
|
||||
No color space conversion required - direct compatibility with graphics mode 2.
|
||||
## Colour Space
|
||||
TEV operates in 8-Bit colour mode, colour space conversion required
|
||||
|
||||
## Compression Features
|
||||
- 8x8 DCT blocks (vs 4x4 in iPF)
|
||||
- 16x16 DCT blocks (vs 4x4 in iPF)
|
||||
- Temporal prediction with motion compensation
|
||||
- Rate-distortion optimized mode selection
|
||||
- Zstd compression with video-optimized settings
|
||||
- Hardware-accelerated encoding/decoding functions
|
||||
|
||||
## Performance Comparison
|
||||
|
||||
Reference in New Issue
Block a user