From f3fb4fe22f28b73dd9642f5973af008003133f16 Mon Sep 17 00:00:00 2001 From: Oliver Schinagl Date: Mon, 16 May 2005 20:02:56 +0000 Subject: Some small stuff changed --- src/fdct.c | 187 +++++++++---------------------------------------------------- src/idct.c | 2 +- 2 files changed, 27 insertions(+), 162 deletions(-) diff --git a/src/fdct.c b/src/fdct.c index 10a222b..2dab7c5 100644 --- a/src/fdct.c +++ b/src/fdct.c @@ -1,114 +1,14 @@ - /* - * jfdctint.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a slow-but-accurate integer implementation of the - * forward DCT (Discrete Cosine Transform). - * - * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT - * on each column. Direct algorithms are also available, but they are - * much more complex and seem not to be any faster when reduced to code. - * - * This implementation is based on an algorithm described in - * C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT - * Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics, - * Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991. - * The primary algorithm described there uses 11 multiplies and 29 adds. - * We use their alternate method with 12 multiplies and 32 adds. - * The advantage of this method is that no data path contains more than one - * multiplication; this allows a very simple and accurate implementation in - * scaled fixed-point arithmetic, with a minimal number of shifts. - */ -/* -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - -*/ #include "dct.h" /* Private declarations for DCT subsystem */ #define DCTSIZE 8 #define INT32 int -#ifdef RIGHT_SHIFT_IS_UNSIGNED -#define SHIFT_TEMPS INT32 shift_temp; -#define RIGHT_SHIFT(x,shft) \ - ((shift_temp = (x)) < 0 ? \ - (shift_temp >> (shft)) | ((~((INT32) 0)) << (32-(shft))) : \ - (shift_temp >> (shft))) -#else -#define SHIFT_TEMPSy -#define RIGHT_SHIFT(x,shft) ((x) >> (shft)) -#endif - -#ifndef USE_ACCURATE_ROUNDING -#undef DESCALE -#define DESCALE(x,n) RIGHT_SHIFT(x, n) -#endif - - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCTs. /* deliberate syntax err */ -#endif - -/* - * The poop on this scaling stuff is as follows: - * - * Each 1-D DCT step produces outputs which are a factor of sqrt(N) - * larger than the true DCT outputs. The final outputs are therefore - * a factor of N larger than desired; since N=8 this can be cured by - * a simple right shift at the end of the algorithm. The advantage of - * this arrangement is that we save two multiplications per 1-D DCT, - * because the y0 and y4 outputs need not be divided by sqrt(N). - * In the IJG code, this factor of 8 is removed by the quantization step - * (in jcdctmgr.c), NOT in this module. - * - * We have to do addition and subtraction of the integer inputs, which - * is no problem, and multiplication by fractional constants, which is - * a problem to do in integer arithmetic. We multiply all the constants - * by CONST_SCALE and convert them to integer constants (thus retaining - * CONST_BITS bits of precision in the constants). After doing a - * multiplication we have to divide the product by CONST_SCALE, with proper - * rounding, to produce the correct output. This division can be done - * cheaply as a right shift of CONST_BITS bits. We postpone shifting - * as long as possible so that partial sums can be added together with - * full fractional precision. - * - * The outputs of the first pass are scaled up by PASS1_BITS bits so that - * they are represented to better-than-integral precision. These outputs - * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word - * with the recommended scaling. (For 12-bit sample data, the intermediate - * array is INT32 anyway.) - * - * To avoid overflow of the 32-bit intermediate results in pass 2, we must - * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26. Error analysis - * shows that the values given below are the most effective. - */ +#define DESCALE(x,shft) ((x) >> (shft)) +#define MULTIPLY(var,const) ((var) * (const)) -#if BITS_IN_JSAMPLE == 8 -#define CONST_BITS 13 -#define PASS1_BITS 2 -#else #define CONST_BITS 13 #define PASS1_BITS 1 /* lose a little precision to avoid overflow */ -#endif - -/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus - * causing a lot of useless floating-point operations at run time. - * To get around this we use the following pre-calculated constants. - * If you change CONST_BITS you may want to add appropriate values. - * (With a reasonable C compiler, you can just rely on the FIX() macro...) - */ -#if CONST_BITS == 13 #define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ #define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ #define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ @@ -121,43 +21,12 @@ #define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ #define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ #define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ -#else -#define FIX_0_298631336 FIX(0.298631336) -#define FIX_0_390180644 FIX(0.390180644) -#define FIX_0_541196100 FIX(0.541196100) -#define FIX_0_765366865 FIX(0.765366865) -#define FIX_0_899976223 FIX(0.899976223) -#define FIX_1_175875602 FIX(1.175875602) -#define FIX_1_501321110 FIX(1.501321110) -#define FIX_1_847759065 FIX(1.847759065) -#define FIX_1_961570560 FIX(1.961570560) -#define FIX_2_053119869 FIX(2.053119869) -#define FIX_2_562915447 FIX(2.562915447) -#define FIX_3_072711026 FIX(3.072711026) -#endif - - -/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. - * For 8-bit samples with the recommended scaling, all the variable - * and constant values involved are no more than 16 bits wide, so a - * 16x16->32 bit multiply can be used instead of a full 32x32 multiply. - * For 12-bit samples, a full 32-bit multiplication will be needed. - */ - -#if BITS_IN_JSAMPLE == 8 -#define MULTIPLY(var,const) MULTIPLY16C16(var,const) -#else -#define MULTIPLY(var,const) ((var) * (const)) -#endif /* * Perform the forward DCT on one block of samples. */ -//GLOBAL(void) -//jpeg_fdct_islow (DCTELEM * data) -// void jpeg_fdct_ifast (DCTELEM * data) { INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; @@ -181,36 +50,36 @@ void jpeg_fdct_ifast (DCTELEM * data) tmp5 = dataptr[2] - dataptr[5]; tmp3 = dataptr[3] + dataptr[4]; tmp4 = dataptr[3] - dataptr[4]; - + /* Even part per LL&M figure 1 --- note that published figure is faulty; * rotator "sqrt(2)*c1" should be "sqrt(2)*c6". */ - + tmp10 = tmp0 + tmp3; tmp13 = tmp0 - tmp3; tmp11 = tmp1 + tmp2; tmp12 = tmp1 - tmp2; - + dataptr[0] = (DCTELEM) ((tmp10 + tmp11) << PASS1_BITS); dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << PASS1_BITS); - + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); dataptr[2] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp13, FIX_0_765366865), CONST_BITS-PASS1_BITS); dataptr[6] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp12, - FIX_1_847759065), CONST_BITS-PASS1_BITS); - + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). * cK represents cos(K*pi/16). * i0..i3 in the paper are tmp4..tmp7 here. */ - + z1 = tmp4 + tmp7; z2 = tmp5 + tmp6; z3 = tmp4 + tmp6; z4 = tmp5 + tmp7; z5 = MULTIPLY(z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */ - + tmp4 = MULTIPLY(tmp4, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */ tmp5 = MULTIPLY(tmp5, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */ tmp6 = MULTIPLY(tmp6, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */ @@ -219,15 +88,15 @@ void jpeg_fdct_ifast (DCTELEM * data) z2 = MULTIPLY(z2, - FIX_2_562915447); /* sqrt(2) * (-c1-c3) */ z3 = MULTIPLY(z3, - FIX_1_961570560); /* sqrt(2) * (-c3-c5) */ z4 = MULTIPLY(z4, - FIX_0_390180644); /* sqrt(2) * (c5-c3) */ - + z3 += z5; z4 += z5; - + dataptr[7] = (DCTELEM) DESCALE(tmp4 + z1 + z3, CONST_BITS-PASS1_BITS); dataptr[5] = (DCTELEM) DESCALE(tmp5 + z2 + z4, CONST_BITS-PASS1_BITS); dataptr[3] = (DCTELEM) DESCALE(tmp6 + z2 + z3, CONST_BITS-PASS1_BITS); dataptr[1] = (DCTELEM) DESCALE(tmp7 + z1 + z4, CONST_BITS-PASS1_BITS); - + dataptr += DCTSIZE; /* advance pointer to next row */ } @@ -246,36 +115,36 @@ void jpeg_fdct_ifast (DCTELEM * data) tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; tmp4 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; - + /* Even part per LL&M figure 1 --- note that published figure is faulty; * rotator "sqrt(2)*c1" should be "sqrt(2)*c6". */ - + tmp10 = tmp0 + tmp3; tmp13 = tmp0 - tmp3; tmp11 = tmp1 + tmp2; tmp12 = tmp1 - tmp2; - + dataptr[DCTSIZE*0] = (DCTELEM) DESCALE(tmp10 + tmp11, PASS1_BITS); dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp10 - tmp11, PASS1_BITS); - + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp13, FIX_0_765366865), CONST_BITS+PASS1_BITS); dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 + MULTIPLY(tmp12, - FIX_1_847759065), CONST_BITS+PASS1_BITS); - + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). * cK represents cos(K*pi/16). * i0..i3 in the paper are tmp4..tmp7 here. */ - + z1 = tmp4 + tmp7; z2 = tmp5 + tmp6; z3 = tmp4 + tmp6; z4 = tmp5 + tmp7; z5 = MULTIPLY(z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */ - + tmp4 = MULTIPLY(tmp4, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */ tmp5 = MULTIPLY(tmp5, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */ tmp6 = MULTIPLY(tmp6, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */ @@ -284,19 +153,15 @@ void jpeg_fdct_ifast (DCTELEM * data) z2 = MULTIPLY(z2, - FIX_2_562915447); /* sqrt(2) * (-c1-c3) */ z3 = MULTIPLY(z3, - FIX_1_961570560); /* sqrt(2) * (-c3-c5) */ z4 = MULTIPLY(z4, - FIX_0_390180644); /* sqrt(2) * (c5-c3) */ - + z3 += z5; z4 += z5; - - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp4 + z1 + z3, - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp5 + z2 + z4, - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp6 + z2 + z3, - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp7 + z1 + z4, - CONST_BITS+PASS1_BITS); - + + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp4 + z1 + z3, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp5 + z2 + z4, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp6 + z2 + z3, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp7 + z1 + z4, CONST_BITS+PASS1_BITS); + dataptr++; /* advance pointer to next column */ } } diff --git a/src/idct.c b/src/idct.c index db83be8..fce305a 100644 --- a/src/idct.c +++ b/src/idct.c @@ -3,7 +3,7 @@ /* Author : Pierre Guerrier, march 1998 */ /* IDCT code by Geert Janssen */ /*---------------------------------------------*/ -#include "idct.h" +#include "dct.h" /*-------------------------------------------------------*/ /* JPEG data types here */ -- cgit v0.12