Blackfin optimized YUV420 to RGB CSC Color Space Converters.

YUV2 -> RGB BGR for 565, 555 and 888 a.k.a. 24bit color.
Speed-up compared to C version compiled with -O3  187.28%
Patch by Marc Hoffman %mmh A pleasantst P com%
Original thread:
Date: May 9, 2007 2:46 AM
Subject: [FFmpeg-devel] PATCH BlackFin yuv2rgb color space conversion


git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@23307 b3059339-0415-0410-9bf9-f77b7e298cf2

1 files changed, 205 insertions, 0 deletions

diff --git a/libswscale/yuv2rgb_bfin.c b/libswscale/yuv2rgb_bfin.c
new file mode 100644
index 0000000000..98c86c06f9
--- /dev/null
+++ b/libswscale/yuv2rgb_bfin.c
@@ -0,0 +1,205 @@
+/*
+ * Copyright (C) 2007 Marc Hoffman <marc.hoffman@analog.com>
+ *                    April 20, 2007
+ *
+ * Blackfin Video Color Space Converters Operations
+ *  convert I420 YV12 to RGB in various formats,
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <inttypes.h>
+#include <assert.h>
+#include "config.h"
+#ifdef HAVE_MALLOC_H
+#include <malloc.h>
+#endif
+#include <unistd.h>
+#include <bits/bfin_sram.h>
+#include "rgb2rgb.h"
+#include "swscale.h"
+#include "swscale_internal.h"
+
+
+#define L1CODE __attribute__ ((l1_text))
+
+extern void ff_bfin_yuv2rgb555_line (uint8_t *Y, uint8_t *U, uint8_t *V, uint8_t *out,
+                                     int w, uint32_t *coeffs) L1CODE;
+
+extern void ff_bfin_yuv2rgb565_line (uint8_t *Y, uint8_t *U, uint8_t *V, uint8_t *out,
+                                     int w, uint32_t *coeffs) L1CODE;
+
+extern void ff_bfin_yuv2rgb24_line (uint8_t *Y, uint8_t *U, uint8_t *V, uint8_t *out,
+                                    int w, uint32_t *coeffs) L1CODE;
+
+typedef void (* ltransform_t)(uint8_t *Y, uint8_t *U, uint8_t *V, uint8_t *out,
+                              int w, uint32_t *coeffs);
+
+
+static void bfin_prepare_coefficients (SwsContext *c, int rgb, int masks)
+{
+    int oy;
+    oy      = c->yOffset&0xffff;
+    oy      = oy >> 3; // keep everything U8.0 for offset calculation
+
+    c->oc   = 128*0x01010101U;
+    c->oy   =  oy*0x01010101U;
+
+    /* copy 64bit vector coeffs down to 32bit vector coeffs */
+    c->cy  = c->yCoeff;
+    c->zero = 0;
+
+    if (rgb) {
+        c->crv = c->vrCoeff;
+        c->cbu = c->ubCoeff;
+        c->cgu = c->ugCoeff;
+        c->cgv = c->vgCoeff;
+    } else {
+        c->crv = c->ubCoeff;
+        c->cbu = c->vrCoeff;
+        c->cgu = c->vgCoeff;
+        c->cgv = c->ugCoeff;
+    }
+
+
+    if (masks == 555) {
+        c->rmask = 0x001f * 0x00010001U;
+        c->gmask = 0x03e0 * 0x00010001U;
+        c->bmask = 0x7c00 * 0x00010001U;
+    } else if (masks == 565) {
+        c->rmask = 0x001f * 0x00010001U;
+        c->gmask = 0x07e0 * 0x00010001U;
+        c->bmask = 0xf800 * 0x00010001U;
+    }
+}
+
+static int core_yuv420_rgb (SwsContext *c,
+                            uint8_t **in, int *instrides,
+                            int srcSliceY, int srcSliceH,
+                            uint8_t **oplanes, int *outstrides,
+                            ltransform_t lcscf, int rgb, int masks)
+{
+    uint8_t *py,*pu,*pv,*op;
+    int w  = instrides[0];
+    int h2 = srcSliceH>>1;
+    int i;
+
+    bfin_prepare_coefficients (c, rgb, masks);
+
+    py = in[0];
+    pu = in[1+(1^rgb)];
+    pv = in[1+(0^rgb)];
+
+    op = oplanes[0] + srcSliceY*outstrides[0];
+
+    for (i=0;i<h2;i++) {
+
+        lcscf (py,pu,pv,op,w,&c->oy);
+
+        py += instrides[0];
+        op += outstrides[0];
+
+        lcscf (py,pu,pv,op,w,&c->oy);
+
+        py += instrides[0];
+        pu += instrides[1];
+        pv += instrides[2];
+        op += outstrides[0];
+    }
+
+    return srcSliceH;
+}
+
+
+static int bfin_yuv420_rgb555 (SwsContext *c,
+                               uint8_t **in, int *instrides,
+                               int srcSliceY, int srcSliceH,
+                               uint8_t **oplanes, int *outstrides)
+{
+    return core_yuv420_rgb (c,in,instrides,srcSliceY,srcSliceH,oplanes,outstrides,
+                            ff_bfin_yuv2rgb555_line, 1, 555);
+}
+
+static int bfin_yuv420_bgr555 (SwsContext *c,
+                               uint8_t **in, int *instrides,
+                               int srcSliceY, int srcSliceH,
+                               uint8_t **oplanes, int *outstrides)
+{
+    return core_yuv420_rgb (c,in,instrides,srcSliceY,srcSliceH,oplanes,outstrides,
+                            ff_bfin_yuv2rgb555_line, 0, 555);
+}
+
+static int bfin_yuv420_rgb24 (SwsContext *c,
+                              uint8_t **in, int *instrides,
+                              int srcSliceY, int srcSliceH,
+                              uint8_t **oplanes, int *outstrides)
+{
+    return core_yuv420_rgb (c,in,instrides,srcSliceY,srcSliceH,oplanes,outstrides,
+                            ff_bfin_yuv2rgb24_line, 1, 888);
+}
+
+static int bfin_yuv420_bgr24 (SwsContext *c,
+                              uint8_t **in, int *instrides,
+                              int srcSliceY, int srcSliceH,
+                              uint8_t **oplanes, int *outstrides)
+{
+    return core_yuv420_rgb (c,in,instrides,srcSliceY,srcSliceH,oplanes,outstrides,
+                            ff_bfin_yuv2rgb24_line, 0, 888);
+}
+
+static int bfin_yuv420_rgb565 (SwsContext *c,
+                               uint8_t **in, int *instrides,
+                               int srcSliceY, int srcSliceH,
+                               uint8_t **oplanes, int *outstrides)
+{
+    return core_yuv420_rgb (c,in,instrides,srcSliceY,srcSliceH,oplanes,outstrides,
+                            ff_bfin_yuv2rgb565_line, 1, 565);
+}
+
+static int bfin_yuv420_bgr565 (SwsContext *c,
+                               uint8_t **in, int *instrides,
+                               int srcSliceY, int srcSliceH,
+                               uint8_t **oplanes, int *outstrides)
+{
+    return core_yuv420_rgb (c,in,instrides,srcSliceY,srcSliceH,oplanes,outstrides,
+                            ff_bfin_yuv2rgb565_line, 0, 565);
+}
+
+
+SwsFunc ff_bfin_yuv2rgb_get_func_ptr (SwsContext *c)
+{
+    SwsFunc f;
+
+    switch(c->dstFormat) {
+    case PIX_FMT_RGB555: f = bfin_yuv420_rgb555; break;
+    case PIX_FMT_BGR555: f = bfin_yuv420_bgr555; break;
+    case PIX_FMT_RGB565: f = bfin_yuv420_rgb565; break;
+    case PIX_FMT_BGR565: f = bfin_yuv420_bgr565; break;
+    case PIX_FMT_RGB24:  f = bfin_yuv420_rgb24;  break;
+    case PIX_FMT_BGR24:  f = bfin_yuv420_bgr24;  break;
+    default:
+        return 0;
+    }
+
+    av_log(c, AV_LOG_INFO, "BlackFin Accelerated Color Space Converter %s\n",
+           sws_format_name (c->dstFormat));
+
+    return f;
+}