code wiki / _hdl_build / nx_vtransform_dct_gate.nx

nx_vtransform_dct_gate.nx

buildroot/runtime/_hdl_build/nx_vtransform_dct_gate.nx

5915 B109 linesdepth 3pulls 5 transitivereach 0 importersview sourcekind gate/prooftopic vtransform
docsdependenciesstructsconstsfunctions

about

nx_vtransform_dct_gate.nx -- verifies the sovereign H.264 integer DCT (nx_vtransform_dct) IN ISOLATION before it touches the codec: (1) the full pipeline fwd->quant->dequant->inv->inv_scale reconstructs real 4x4 blocks faithfully (a wrong table/shift would blow this up), and (2) the DCT compacts natural-image energy into the low-frequency coefficients BETTER than the current Walsh-Hadamard transform (the reason to swap). Blocks come from the real 768x768 H.264-decoded luma frame (ref_frame0.yuv), centered to signed residual-like values.

dependencies 4 imports · 0 importers

nx_syscalls.nx nx_gate_emit_lib.nx nx_vtransform_dct.nx nx_vtransform.nx nx_vtransform_dct_gate.nx

imports: nx_syscalls.nxnx_gate_emit_lib.nxnx_vtransform_dct.nxnx_vtransform.nx

imported by: nobody (leaf or entry point)

call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown

main g_puts sys_write sys_mmap sys_read_file sys_openat_rd sys_lseek sys_mmap ↻ sys_read sys_close sys_exit g_check g_puts ↻ load_block vtd_fwd vtd_fwd1 vtd_quant vtd_class vtd_abs vtd_mf vtd_dequant vtd_class ↻ vtd_vi vtd_inv vtd_inv1 vtd_inv_scale g_abs g_pn sys_mmap ↻ sys_write ↻ vt_fwd vt_wht4 top4_frac1000

structs

none

consts

none

functions

11func g_abs(v: i64) -> i64 { if v < 0 { return 0 - v } return v }
called by 1: main
14func load_block(src: *u8, W: i64, px: i64, py: i64, blk: *i64) -> i64
called by 1: main
23func top4_frac1000(c: *i64) -> i64
called by 1: main
37func main() -> i64