gambatte/common/videolink/rgb32conv.cpp
sinamas d44f804f03 common/videolink: cleanup/minor tweaks
use ptrdiff prevline offset in hqx filters to facilitate use of lea.

use endptr rathern than loop ctr in rgb2yuv to avoid gcc/x86 rmw
spill.

avoid reloading rhs input pixel after storing lhs output in rgb2yuv
(no restrict keyword in c++98).

fix hypothetical component overflow in hq2x blend9.
2013-04-21 16:16:47 +02:00

199 lines
6.8 KiB
C++

/***************************************************************************
* Copyright (C) 2009 by Sindre Aamås *
* sinamas@users.sourceforge.net *
* *
* This program is free software; you can redistribute it and/or modify *
* it under the terms of the GNU General Public License version 2 as *
* published by the Free Software Foundation. *
* *
* This program 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 General Public License version 2 for more details. *
* *
* You should have received a copy of the GNU General Public License *
* version 2 along with this program; if not, write to the *
* Free Software Foundation, Inc., *
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. *
***************************************************************************/
#include "rgb32conv.h"
#include "array.h"
#include "gbint.h"
#include "videolink.h"
#include <algorithm>
namespace {
static bool isBigEndian() {
union {
gambatte::uint_least32_t ul32;
unsigned char uc[sizeof(gambatte::uint_least32_t)];
} u;
u.ul32 = -0x10000;
return u.uc[0];
}
class Rgb32ToUyvy {
public:
Rgb32ToUyvy();
void operator()(gambatte::uint_least32_t *d, std::ptrdiff_t dstPitch,
gambatte::uint_least32_t const *s, std::ptrdiff_t srcPitch,
unsigned w, unsigned h);
private:
struct CacheUnit {
gambatte::uint_least32_t rgb32;
gambatte::uint_least32_t uyvy;
};
enum { cache_size = 0x100 };
enum { cache_mask = cache_size - 1 };
CacheUnit cache_[cache_size];
};
Rgb32ToUyvy::Rgb32ToUyvy() {
if (isBigEndian()) {
CacheUnit c = { 0, 128ul << 24 | 16ul << 16 | 128u << 8 | 16 };
std::fill(cache_, cache_ + cache_size, c);
} else {
CacheUnit c = { 0, 16ul << 24 | 128ul << 16 | 16 << 8 | 128 };
std::fill(cache_, cache_ + cache_size, c);
}
}
void Rgb32ToUyvy::operator()(gambatte::uint_least32_t *dst,
std::ptrdiff_t const dstPitch,
gambatte::uint_least32_t const *src,
std::ptrdiff_t const srcPitch,
unsigned const w,
unsigned h)
{
while (h--) {
gambatte::uint_least32_t *d = dst;
gambatte::uint_least32_t const *s = src;
gambatte::uint_least32_t const *const sEnd = s + w - 1;
while (s < sEnd) {
if ((cache_[s[0] & cache_mask].rgb32 - s[0]) | (cache_[s[1] & cache_mask].rgb32 - s[1])) {
cache_[s[0] & cache_mask].rgb32 = s[0];
cache_[s[1] & cache_mask].rgb32 = s[1];
unsigned long const r = (s[0] >> 16 & 0x000000FF) | (s[1] & 0x00FF0000);
unsigned long const g = (s[0] >> 8 & 0x000000FF) | (s[1] << 8 & 0x00FF0000);
unsigned long const b = (s[0] & 0x000000FF) | (s[1] << 16 & 0x00FF0000);
unsigned long const y = r * 66 + g * 129 + b * 25 + ( 16 * 256 + 128) * 0x00010001ul;
unsigned long const u = b * 112 - r * 38 - g * 74 + (128 * 256 + 128) * 0x00010001ul;
unsigned long const v = r * 112 - g * 94 - b * 18 + (128 * 256 + 128) * 0x00010001ul;
if (isBigEndian()) {
cache_[s[0] & cache_mask].uyvy = (u << 16 & 0xFF000000)
| (y << 8 & 0x00FF0000)
| (v & 0x0000FF00)
| (y >> 8 & 0x000000FF);
cache_[s[1] & cache_mask].uyvy = (u & 0xFF000000)
| (y >> 8 & 0x00FF0000)
| (v >> 16 & 0x0000FF00)
| y >> 24 ;
} else {
cache_[s[0] & cache_mask].uyvy = (y << 16 & 0xFF000000)
| (v << 8 & 0x00FF0000)
| (y & 0x0000FF00)
| (u >> 8 & 0x000000FF);
cache_[s[1] & cache_mask].uyvy = (y & 0xFF000000)
| (v >> 8 & 0x00FF0000)
| (y >> 16 & 0x0000FF00)
| u >> 24 ;
}
}
gambatte::uint_least32_t const s0 = s[0], s1 = s[1];
d[0] = cache_[s0 & cache_mask].uyvy;
d[1] = cache_[s1 & cache_mask].uyvy;
s += 2;
d += 2;
}
src += srcPitch;
dst += dstPitch;
}
}
static void rgb32ToRgb16(gambatte::uint_least16_t *d,
std::ptrdiff_t const dstPitch,
gambatte::uint_least32_t const *s,
std::ptrdiff_t const srcPitch,
unsigned const w,
unsigned h)
{
do {
std::ptrdiff_t i = -static_cast<std::ptrdiff_t>(w);
s += w;
d += w;
do {
d[i] = (s[i] >> 8 & 0xF800) | (s[i] & 0xFC00) >> 5 | (s[i] & 0xFF) >> 3;
} while (++i);
s += srcPitch - static_cast<std::ptrdiff_t>(w);
d += dstPitch - static_cast<std::ptrdiff_t>(w);
} while (--h);
}
class Rgb32ToUyvyLink : public VideoLink {
public:
Rgb32ToUyvyLink(unsigned width, unsigned height)
: inbuf_(static_cast<std::size_t>(width) * height)
, width_(width)
, height_(height)
{
}
virtual void * inBuf() const { return inbuf_; }
virtual std::ptrdiff_t inPitch() const { return width_; }
virtual void draw(void *dst, std::ptrdiff_t dstPitch) {
rgb32ToUyvy_(static_cast<gambatte::uint_least32_t *>(dst), dstPitch,
inbuf_, width_, width_, height_);
}
private:
SimpleArray<gambatte::uint_least32_t> const inbuf_;
Rgb32ToUyvy rgb32ToUyvy_;
unsigned const width_;
unsigned const height_;
};
class Rgb32ToRgb16Link : public VideoLink {
public:
Rgb32ToRgb16Link(unsigned width, unsigned height)
: inbuf_(static_cast<std::size_t>(width) * height)
, width_(width)
, height_(height)
{
}
virtual void * inBuf() const { return inbuf_; }
virtual std::ptrdiff_t inPitch() const { return width_; }
virtual void draw(void *dst, std::ptrdiff_t dstPitch) {
if (!inbuf_)
return;
rgb32ToRgb16(static_cast<gambatte::uint_least16_t *>(dst), dstPitch,
inbuf_, width_, width_, height_);
}
private:
SimpleArray<gambatte::uint_least32_t> const inbuf_;
unsigned const width_;
unsigned const height_;
};
} // anon namespace
VideoLink * Rgb32Conv::create(PixelFormat pf, unsigned width, unsigned height) {
switch (pf) {
case RGB16: return new Rgb32ToRgb16Link(width, height);
case UYVY: return new Rgb32ToUyvyLink(width, height);
default: return 0;
}
}