yading@10: /*
yading@10:  * S3 Texture Compression (S3TC) decoding functions
yading@10:  * Copyright (c) 2007 by Ivo van Poorten
yading@10:  *
yading@10:  * see also: http://wiki.multimedia.cx/index.php?title=S3TC
yading@10:  *
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: #include "libavcodec/bytestream.h"
yading@10: #include "avcodec.h"
yading@10: #include "s3tc.h"
yading@10: 
yading@10: static inline void dxt1_decode_pixels(GetByteContext *gb, uint32_t *d,
yading@10:                                       unsigned int qstride, unsigned int flag,
yading@10:                                       uint64_t alpha) {
yading@10:     unsigned int x, y, c0, c1, a = (!flag * 255u) << 24;
yading@10:     unsigned int rb0, rb1, rb2, rb3, g0, g1, g2, g3;
yading@10:     uint32_t colors[4], pixels;
yading@10: 
yading@10:     c0 = bytestream2_get_le16(gb);
yading@10:     c1 = bytestream2_get_le16(gb);
yading@10: 
yading@10:     rb0  = (c0<<3 | c0<<8) & 0xf800f8;
yading@10:     rb1  = (c1<<3 | c1<<8) & 0xf800f8;
yading@10:     rb0 +=        (rb0>>5) & 0x070007;
yading@10:     rb1 +=        (rb1>>5) & 0x070007;
yading@10:     g0   =        (c0 <<5) & 0x00fc00;
yading@10:     g1   =        (c1 <<5) & 0x00fc00;
yading@10:     g0  +=        (g0 >>6) & 0x000300;
yading@10:     g1  +=        (g1 >>6) & 0x000300;
yading@10: 
yading@10:     colors[0] = rb0 + g0 + a;
yading@10:     colors[1] = rb1 + g1 + a;
yading@10: 
yading@10:     if (c0 > c1 || flag) {
yading@10:         rb2 = (((2*rb0+rb1) * 21) >> 6) & 0xff00ff;
yading@10:         rb3 = (((2*rb1+rb0) * 21) >> 6) & 0xff00ff;
yading@10:         g2  = (((2*g0 +g1 ) * 21) >> 6) & 0x00ff00;
yading@10:         g3  = (((2*g1 +g0 ) * 21) >> 6) & 0x00ff00;
yading@10:         colors[3] = rb3 + g3 + a;
yading@10:     } else {
yading@10:         rb2 = ((rb0+rb1) >> 1) & 0xff00ff;
yading@10:         g2  = ((g0 +g1 ) >> 1) & 0x00ff00;
yading@10:         colors[3] = 0;
yading@10:     }
yading@10: 
yading@10:     colors[2] = rb2 + g2 + a;
yading@10: 
yading@10:     pixels = bytestream2_get_le32(gb);
yading@10:     for (y=0; y<4; y++) {
yading@10:         for (x=0; x<4; x++) {
yading@10:             a        = (alpha & 0x0f) << 28;
yading@10:             a       += a >> 4;
yading@10:             d[x]     = a + colors[pixels&3];
yading@10:             pixels >>= 2;
yading@10:             alpha  >>= 4;
yading@10:         }
yading@10:         d += qstride;
yading@10:     }
yading@10: }
yading@10: 
yading@10: void ff_decode_dxt1(GetByteContext *gb, uint8_t *dst,
yading@10:                     const unsigned int w, const unsigned int h,
yading@10:                     const unsigned int stride) {
yading@10:     unsigned int bx, by, qstride = stride/4;
yading@10:     uint32_t *d = (uint32_t *) dst;
yading@10: 
yading@10:     for (by=0; by < h/4; by++, d += stride-w)
yading@10:         for (bx = 0; bx < w / 4; bx++, d += 4)
yading@10:             dxt1_decode_pixels(gb, d, qstride, 0, 0LL);
yading@10: }
yading@10: 
yading@10: void ff_decode_dxt3(GetByteContext *gb, uint8_t *dst,
yading@10:                     const unsigned int w, const unsigned int h,
yading@10:                     const unsigned int stride) {
yading@10:     unsigned int bx, by, qstride = stride/4;
yading@10:     uint32_t *d = (uint32_t *) dst;
yading@10: 
yading@10:     for (by=0; by < h/4; by++, d += stride-w)
yading@10:         for (bx = 0; bx < w / 4; bx++, d += 4)
yading@10:             dxt1_decode_pixels(gb, d, qstride, 1, bytestream2_get_le64(gb));
yading@10: }