| @@ -2146,17 +2146,17 @@ static int packedCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 2145 unmodified lines |
| 2146 | 2146 | if (c->opts.dither == SWS_DITHER_NONE) {\ |
| 2147 | 2147 | for (i = 0; i < height; i++) {\ |
| 2148 | 2148 | for (j = 0; j < length-7; j+=8) {\ |
| 2149 | | − tmp = (bswap(src[j+0]) + bias)>>shift; dst[j+0] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2150 | | − tmp = (bswap(src[j+1]) + bias)>>shift; dst[j+1] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2151 | | − tmp = (bswap(src[j+2]) + bias)>>shift; dst[j+2] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2152 | | − tmp = (bswap(src[j+3]) + bias)>>shift; dst[j+3] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2153 | | − tmp = (bswap(src[j+4]) + bias)>>shift; dst[j+4] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2154 | | − tmp = (bswap(src[j+5]) + bias)>>shift; dst[j+5] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2155 | | − tmp = (bswap(src[j+6]) + bias)>>shift; dst[j+6] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2156 | | − tmp = (bswap(src[j+7]) + bias)>>shift; dst[j+7] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2149 | + tmp = ((bswap(src[j+0]) >> src_shift) + bias)>>shift; dst[j+0] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2150 | + tmp = ((bswap(src[j+1]) >> src_shift) + bias)>>shift; dst[j+1] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2151 | + tmp = ((bswap(src[j+2]) >> src_shift) + bias)>>shift; dst[j+2] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2152 | + tmp = ((bswap(src[j+3]) >> src_shift) + bias)>>shift; dst[j+3] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2153 | + tmp = ((bswap(src[j+4]) >> src_shift) + bias)>>shift; dst[j+4] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2154 | + tmp = ((bswap(src[j+5]) >> src_shift) + bias)>>shift; dst[j+5] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2155 | + tmp = ((bswap(src[j+6]) >> src_shift) + bias)>>shift; dst[j+6] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2156 | + tmp = ((bswap(src[j+7]) >> src_shift) + bias)>>shift; dst[j+7] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2157 | 2157 | }\ |
| 2158 | 2158 | for (; j < length; j++) {\ |
| 2159 | | − tmp = (bswap(src[j]) + bias)>>shift; dst[j] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2159 | + tmp = (bswap(src[j]) + bias)>>shift; dst[j] = dbswap(tmp - (tmp>>dst_depth) << dst_shift);\ |
| 2160 | 2160 | }\ |
| 2161 | 2161 | dst += dstStride;\ |
| 2162 | 2162 | src += srcStride;\ |
| @@ -2165,17 +2165,17 @@ static int packedCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 2 unmodified lines |
| 2165 | 2165 | for (i = 0; i < height; i++) {\ |
| 2166 | 2166 | const uint8_t *dither= dithers[shift-1][i&7];\ |
| 2167 | 2167 | for (j = 0; j < length-7; j+=8) {\ |
| 2168 | | − tmp = (bswap(src[j+0]) + dither[0])>>shift; dst[j+0] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2169 | | − tmp = (bswap(src[j+1]) + dither[1])>>shift; dst[j+1] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2170 | | − tmp = (bswap(src[j+2]) + dither[2])>>shift; dst[j+2] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2171 | | − tmp = (bswap(src[j+3]) + dither[3])>>shift; dst[j+3] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2172 | | − tmp = (bswap(src[j+4]) + dither[4])>>shift; dst[j+4] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2173 | | − tmp = (bswap(src[j+5]) + dither[5])>>shift; dst[j+5] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2174 | | − tmp = (bswap(src[j+6]) + dither[6])>>shift; dst[j+6] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2175 | | − tmp = (bswap(src[j+7]) + dither[7])>>shift; dst[j+7] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2168 | + tmp = ((bswap(src[j+0]) >> src_shift) + dither[0])>>shift; dst[j+0] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2169 | + tmp = ((bswap(src[j+1]) >> src_shift) + dither[1])>>shift; dst[j+1] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2170 | + tmp = ((bswap(src[j+2]) >> src_shift) + dither[2])>>shift; dst[j+2] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2171 | + tmp = ((bswap(src[j+3]) >> src_shift) + dither[3])>>shift; dst[j+3] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2172 | + tmp = ((bswap(src[j+4]) >> src_shift) + dither[4])>>shift; dst[j+4] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2173 | + tmp = ((bswap(src[j+5]) >> src_shift) + dither[5])>>shift; dst[j+5] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2174 | + tmp = ((bswap(src[j+6]) >> src_shift) + dither[6])>>shift; dst[j+6] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2175 | + tmp = ((bswap(src[j+7]) >> src_shift) + dither[7])>>shift; dst[j+7] = dbswap((tmp - (tmp>>dst_depth)) << dst_shift);\ |
| 2176 | 2176 | }\ |
| 2177 | 2177 | for (; j < length; j++) {\ |
| 2178 | | − tmp = (bswap(src[j]) + dither[j&7])>>shift; dst[j] = dbswap(tmp - (tmp>>dst_depth));\ |
| 2178 | + tmp = (bswap(src[j]) + dither[j&7])>>shift; dst[j] = dbswap(tmp - (tmp>>dst_depth) << dst_shift);\ |
| 2179 | 2179 | }\ |
| 2180 | 2180 | dst += dstStride;\ |
| 2181 | 2181 | src += srcStride;\ |
| @@ -2184,14 +2184,14 @@ static int packedCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 2 unmodified lines |
| 2184 | 2184 | for (i = 0; i < height; i++) {\ |
| 2185 | 2185 | const uint8_t *dither= dithers[shift-1][i&7];\ |
| 2186 | 2186 | for (j = 0; j < length-7; j+=8) {\ |
| 2187 | | − tmp = bswap(src[j+0]); dst[j+0] = dbswap((tmp - (tmp>>dst_depth) + dither[0])>>shift);\ |
| 2188 | | − tmp = bswap(src[j+1]); dst[j+1] = dbswap((tmp - (tmp>>dst_depth) + dither[1])>>shift);\ |
| 2189 | | − tmp = bswap(src[j+2]); dst[j+2] = dbswap((tmp - (tmp>>dst_depth) + dither[2])>>shift);\ |
| 2190 | | − tmp = bswap(src[j+3]); dst[j+3] = dbswap((tmp - (tmp>>dst_depth) + dither[3])>>shift);\ |
| 2191 | | − tmp = bswap(src[j+4]); dst[j+4] = dbswap((tmp - (tmp>>dst_depth) + dither[4])>>shift);\ |
| 2192 | | − tmp = bswap(src[j+5]); dst[j+5] = dbswap((tmp - (tmp>>dst_depth) + dither[5])>>shift);\ |
| 2193 | | − tmp = bswap(src[j+6]); dst[j+6] = dbswap((tmp - (tmp>>dst_depth) + dither[6])>>shift);\ |
| 2194 | | − tmp = bswap(src[j+7]); dst[j+7] = dbswap((tmp - (tmp>>dst_depth) + dither[7])>>shift);\ |
| 2187 | + tmp = bswap(src[j+0]) >> src_shift; dst[j+0] = dbswap(((tmp - (tmp>>dst_depth) + dither[0])>>shift) << dst_shift);\ |
| 2188 | + tmp = bswap(src[j+1]) >> src_shift; dst[j+1] = dbswap(((tmp - (tmp>>dst_depth) + dither[1])>>shift) << dst_shift);\ |
| 2189 | + tmp = bswap(src[j+2]) >> src_shift; dst[j+2] = dbswap(((tmp - (tmp>>dst_depth) + dither[2])>>shift) << dst_shift);\ |
| 2190 | + tmp = bswap(src[j+3]) >> src_shift; dst[j+3] = dbswap(((tmp - (tmp>>dst_depth) + dither[3])>>shift) << dst_shift);\ |
| 2191 | + tmp = bswap(src[j+4]) >> src_shift; dst[j+4] = dbswap(((tmp - (tmp>>dst_depth) + dither[4])>>shift) << dst_shift);\ |
| 2192 | + tmp = bswap(src[j+5]) >> src_shift; dst[j+5] = dbswap(((tmp - (tmp>>dst_depth) + dither[5])>>shift) << dst_shift);\ |
| 2193 | + tmp = bswap(src[j+6]) >> src_shift; dst[j+6] = dbswap(((tmp - (tmp>>dst_depth) + dither[6])>>shift) << dst_shift);\ |
| 2194 | + tmp = bswap(src[j+7]) >> src_shift; dst[j+7] = dbswap(((tmp - (tmp>>dst_depth) + dither[7])>>shift) << dst_shift);\ |
| 2195 | 2195 | }\ |
| 2196 | 2196 | for (; j < length; j++) {\ |
| 2197 | 2197 | tmp = bswap(src[j]); dst[j] = dbswap((tmp - (tmp>>dst_depth) + dither[j&7])>>shift);\ |
| @@ -2235,6 +2235,8 @@ static int planarCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 37 unmodified lines |
| 2235 | 2235 | ) { |
| 2236 | 2236 | const int src_depth = desc_src->comp[plane].depth; |
| 2237 | 2237 | const int dst_depth = desc_dst->comp[plane].depth; |
| 2238 | + const int src_shift = desc_src->comp[plane].shift; |
| 2239 | + const int dst_shift = desc_dst->comp[plane].shift; |
| 2238 | 2240 | const uint16_t *srcPtr2 = (const uint16_t *) srcPtr; |
| 2239 | 2241 | uint16_t *dstPtr2 = (uint16_t*)dstPtr; |
| 2240 | 2242 | |
| @@ -2250,11 +2252,11 @@ static int planarCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 9 unmodified lines |
| 2250 | 2252 | #define COPY816(w)\ |
| 2251 | 2253 | if (shiftonly) {\ |
| 2252 | 2254 | for (j = 0; j < length; j++)\ |
| 2253 | | − w(&dstPtr2[j], srcPtr[j]<<(dst_depth-8));\ |
| 2255 | + w(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) << dst_shift);\ |
| 2254 | 2256 | } else {\ |
| 2255 | 2257 | for (j = 0; j < length; j++)\ |
| 2256 | | − w(&dstPtr2[j], (srcPtr[j]<<(dst_depth-8)) |\ |
| 2257 | | − (srcPtr[j]>>(2*8-dst_depth)));\ |
| 2258 | + w(&dstPtr2[j], ((srcPtr[j]<<(dst_depth-8)) |\ |
| 2259 | + (srcPtr[j]>>(2*8-dst_depth))) << dst_shift);\ |
| 2258 | 2260 | } |
| 2259 | 2261 | if(isBE(c->opts.dst_format)){ |
| 2260 | 2262 | COPY816(AV_WB16) |
| @@ -2274,13 +2276,13 @@ static int planarCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 13 unmodified lines |
| 2274 | 2276 | #if HAVE_FAST_64BIT |
| 2275 | 2277 | #define FAST_COPY_UP(shift) \ |
| 2276 | 2278 | for (; j < length - 3; j += 4) { \ |
| 2277 | | − uint64_t v = AV_RN64A(srcPtr2 + j); \ |
| 2279 | + uint64_t v = AV_RN64A(srcPtr2 + j) >> src_shift; \ |
| 2278 | 2280 | AV_WN64A(dstPtr2 + j, v << shift); \ |
| 2279 | 2281 | } |
| 2280 | 2282 | #else |
| 2281 | 2283 | #define FAST_COPY_UP(shift) \ |
| 2282 | 2284 | for (; j < length - 1; j += 2) { \ |
| 2283 | | − uint32_t v = AV_RN32A(srcPtr2 + j); \ |
| 2285 | + uint32_t v = AV_RN32A(srcPtr2 + j) >> src_shift; \ |
| 2284 | 2286 | AV_WN32A(dstPtr2 + j, v << shift); \ |
| 2285 | 2287 | } |
| 2286 | 2288 | #endif |
| @@ -2293,14 +2295,14 @@ static int planarCopyWrapper(SwsInternal *c, const uint8_t *const src[], · 6 unmodified lines |
| 2293 | 2295 | #define COPY_UP(r,w) \ |
| 2294 | 2296 | if(shiftonly){\ |
| 2295 | 2297 | for (; j < length; j++){ \ |
| 2296 | | − unsigned int v= r(&srcPtr2[j]);\ |
| 2297 | | − w(&dstPtr2[j], v<<(dst_depth-src_depth));\ |
| 2298 | + unsigned int v= r(&srcPtr2[j]) >> src_shift;\ |
| 2299 | + w(&dstPtr2[j], (v << (dst_depth-src_depth)) << dst_shift);\ |
| 2298 | 2300 | }\ |
| 2299 | 2301 | }else{\ |
| 2300 | 2302 | for (; j < length; j++){ \ |
| 2301 | | − unsigned int v= r(&srcPtr2[j]);\ |
| 2302 | | − w(&dstPtr2[j], (v<<(dst_depth-src_depth)) | \ |
| 2303 | | − (v>>(2*src_depth-dst_depth)));\ |
| 2303 | + unsigned int v= r(&srcPtr2[j]) >> src_shift;\ |
| 2304 | + w(&dstPtr2[j], ((v<<(dst_depth-src_depth)) | \ |
| 2305 | + (v>>(2*src_depth-dst_depth))) << dst_shift);\ |
| 2304 | 2306 | }\ |
| 2305 | 2307 | } |
| 2306 | 2308 | if(isBE(c->opts.src_format)){ |