/* * Copyright ( c ) 2018 , Alliance for Open Media . All rights reserved . * * This source code is subject to the terms of the BSD 2 Clause License and * the Alliance for Open Media Patent License 1 . 0 . If the BSD 2 Clause License * was not distributed with this source code in the LICENSE file , you can * obtain it at www . aomedia . org / license / software . If the Alliance for Open * Media Patent License 1 . 0 was not distributed with this source code in the * PATENTS file , you can obtain it at www . aomedia . org / license / patent . #include "warp_plane_neon.h" static AOM_FORCE_INLINE int16x8_t horizontal_filter_4x1_f4(const uint8x16_t in,int sx, int alpha) {const int32x4_t add_const = vdupq_n_s32(1 << (8 + FILTER_BITS - 1 ));// Loading the 8 filter taps 4 ];0 ], in16_lo);1 ], vextq_s16(in16_lo, in16_hi, 1 ));2 ], vextq_s16(in16_lo, in16_hi, 2 ));3 ], vextq_s16(in16_lo, in16_hi, 3 ));0 ));return vreinterpretq_s16_u16(res);static AOM_FORCE_INLINE int16x8_t horizontal_filter_8x1_f8(const uint8x16_t in,int sx, int alpha) {const int32x4_t add_const = vdupq_n_s32(1 << (8 + FILTER_BITS - 1 ));// Loading the 8 filter taps 8 ];0 ], in16_lo);1 ], vextq_s16(in16_lo, in16_hi, 1 ));2 ], vextq_s16(in16_lo, in16_hi, 2 ));3 ], vextq_s16(in16_lo, in16_hi, 3 ));4 ], vextq_s16(in16_lo, in16_hi, 4 ));5 ], vextq_s16(in16_lo, in16_hi, 5 ));6 ], vextq_s16(in16_lo, in16_hi, 6 ));7 ], vextq_s16(in16_lo, in16_hi, 7 ));return vreinterpretq_s16_u16(res);static AOM_FORCE_INLINE int16x8_tconst uint8x16_t in, int16x8_t f_s16) {const int32x4_t add_const = vdupq_n_s32(1 << (8 + FILTER_BITS - 1 ));1 ));2 ));3 ));0 ));return vreinterpretq_s16_u16(res);static AOM_FORCE_INLINE int16x8_t horizontal_filter_4x1_f1(const uint8x16_t in,int sx) {return horizontal_filter_4x1_f1_beta0(in, f_s16);static AOM_FORCE_INLINE int16x8_tconst uint8x16_t in, int16x8_t f_s16) {const int32x4_t add_const = vdupq_n_s32(1 << (8 + FILTER_BITS - 1 ));1 ));2 ));3 ));4 ));5 ));6 ));7 ));return vreinterpretq_s16_u16(res);static AOM_FORCE_INLINE int16x8_t horizontal_filter_8x1_f1(const uint8x16_t in,int sx) {return horizontal_filter_8x1_f1_beta0(in, f_s16);static AOM_FORCE_INLINE void vertical_filter_4x1_f4(const int16x8_t *src,int sy,int gamma) {0 ]), vget_low_s16(src[1 ]), vget_low_s16(src[2 ]),3 ]), vget_low_s16(src[4 ]), vget_low_s16(src[5 ]),6 ]), vget_low_s16(src[7 ]), &s0, &s1, &s2, &s3);4 ];0 ]));0 ]));1 ]));1 ]));2 ]));2 ]));3 ]));3 ]));static AOM_FORCE_INLINE void vertical_filter_8x1_f8(const int16x8_t *src,int sy,int gamma) {0 ];1 ];2 ];3 ];4 ];5 ];6 ];7 ];8 ];0 ]));0 ]));1 ]));1 ]));2 ]));2 ]));3 ]));3 ]));4 ]));4 ]));5 ]));5 ]));6 ]));6 ]));7 ]));7 ]));static AOM_FORCE_INLINE void warp_affine_horizontal_neon(const uint8_t *ref, int width, int height, int stride, int p_width,int p_height, int16_t alpha, int16_t beta, const int64_t x4,const int64_t y4, const int i, int16x8_t tmp[]) {const int height_limit = AOMMIN(8 , p_height - i) + 7 ;1 << WARPEDMODEL_PREC_BITS) - 1 );4 ) + beta * (-4 ) + (1 << (WARPEDDIFF_PREC_BITS - 1 )) +1 << WARP_PARAM_REDUCE_BITS) - 1 );if (warp_affine_special_case(ref, ix4, iy4, width, height, stride,return ;static const uint8_t kIotaArr[] = { 0 , 1 , 2 , 3 , 4 , 5 , 6 , 7 ,8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 };const uint8x16_t indx = vld1q_u8(kIotaArr);const int out_of_boundary_left = -(ix4 - 6 );const int out_of_boundary_right = (ix4 + 8 ) - width;if (p_width == 4 ) {if (beta == 0 ) {if (alpha == 0 ) {else {else {if (alpha == 0 ) {3 )));else {3 )),else {if (beta == 0 ) {if (alpha == 0 ) {else {else {if (alpha == 0 ) {3 )));else {3 )),void av1_warp_affine_neon(const int32_t *mat, const uint8_t *ref, int width,int height, int stride, uint8_t *pred, int p_col,int p_row, int p_width, int p_height, int p_stride,int subsampling_x, int subsampling_y,const int w0 = conv_params->fwd_offset;const int w1 = conv_params->bck_offset;const int is_compound = conv_params->is_compound;const dst = conv_params->dst;const int dst_stride = conv_params->dst_stride;const int do_average = conv_params->do_average;const int use_dist_wtd_comp_avg = conv_params->use_dist_wtd_comp_avg;for (int i = 0 ; i < p_height; i += 8 ) {for (int j = 0 ; j < p_width; j += 8 ) {const int32_t src_x = (p_col + j + 4 ) << subsampling_x;const int32_t src_y = (p_row + i + 4 ) << subsampling_y;const int64_t dst_x =2 ] * src_x + (int64_t)mat[3 ] * src_y + (int64_t)mat[0 ];const int64_t dst_y =4 ] * src_x + (int64_t)mat[5 ] * src_y + (int64_t)mat[1 ];const int64_t x4 = dst_x >> subsampling_x;const int64_t y4 = dst_y >> subsampling_y;15 ];
Messung V0.5 in Prozent C=100 H=97 G=98
¤ Dauer der Verarbeitung: 0.10 Sekunden
(vorverarbeitet am 2026-09-30)
¤ *© Formatika GbR, Deutschland
