vrshoot

diff libs/libjpeg/jidctint.c @ 0:b2f14e535253

initial commit
author John Tsiombikas <nuclear@member.fsf.org>
date Sat, 01 Feb 2014 19:58:19 +0200
parents
children
line diff
     1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/libs/libjpeg/jidctint.c	Sat Feb 01 19:58:19 2014 +0200
     1.3 @@ -0,0 +1,389 @@
     1.4 +/*
     1.5 + * jidctint.c
     1.6 + *
     1.7 + * Copyright (C) 1991-1998, Thomas G. Lane.
     1.8 + * This file is part of the Independent JPEG Group's software.
     1.9 + * For conditions of distribution and use, see the accompanying README file.
    1.10 + *
    1.11 + * This file contains a slow-but-accurate integer implementation of the
    1.12 + * inverse DCT (Discrete Cosine Transform).  In the IJG code, this routine
    1.13 + * must also perform dequantization of the input coefficients.
    1.14 + *
    1.15 + * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT
    1.16 + * on each row (or vice versa, but it's more convenient to emit a row at
    1.17 + * a time).  Direct algorithms are also available, but they are much more
    1.18 + * complex and seem not to be any faster when reduced to code.
    1.19 + *
    1.20 + * This implementation is based on an algorithm described in
    1.21 + *   C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT
    1.22 + *   Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics,
    1.23 + *   Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991.
    1.24 + * The primary algorithm described there uses 11 multiplies and 29 adds.
    1.25 + * We use their alternate method with 12 multiplies and 32 adds.
    1.26 + * The advantage of this method is that no data path contains more than one
    1.27 + * multiplication; this allows a very simple and accurate implementation in
    1.28 + * scaled fixed-point arithmetic, with a minimal number of shifts.
    1.29 + */
    1.30 +
    1.31 +#define JPEG_INTERNALS
    1.32 +#include "jinclude.h"
    1.33 +#include "jpeglib.h"
    1.34 +#include "jdct.h"		/* Private declarations for DCT subsystem */
    1.35 +
    1.36 +#ifdef DCT_ISLOW_SUPPORTED
    1.37 +
    1.38 +
    1.39 +/*
    1.40 + * This module is specialized to the case DCTSIZE = 8.
    1.41 + */
    1.42 +
    1.43 +#if DCTSIZE != 8
    1.44 +  Sorry, this code only copes with 8x8 DCTs. /* deliberate syntax err */
    1.45 +#endif
    1.46 +
    1.47 +
    1.48 +/*
    1.49 + * The poop on this scaling stuff is as follows:
    1.50 + *
    1.51 + * Each 1-D IDCT step produces outputs which are a factor of sqrt(N)
    1.52 + * larger than the true IDCT outputs.  The final outputs are therefore
    1.53 + * a factor of N larger than desired; since N=8 this can be cured by
    1.54 + * a simple right shift at the end of the algorithm.  The advantage of
    1.55 + * this arrangement is that we save two multiplications per 1-D IDCT,
    1.56 + * because the y0 and y4 inputs need not be divided by sqrt(N).
    1.57 + *
    1.58 + * We have to do addition and subtraction of the integer inputs, which
    1.59 + * is no problem, and multiplication by fractional constants, which is
    1.60 + * a problem to do in integer arithmetic.  We multiply all the constants
    1.61 + * by CONST_SCALE and convert them to integer constants (thus retaining
    1.62 + * CONST_BITS bits of precision in the constants).  After doing a
    1.63 + * multiplication we have to divide the product by CONST_SCALE, with proper
    1.64 + * rounding, to produce the correct output.  This division can be done
    1.65 + * cheaply as a right shift of CONST_BITS bits.  We postpone shifting
    1.66 + * as long as possible so that partial sums can be added together with
    1.67 + * full fractional precision.
    1.68 + *
    1.69 + * The outputs of the first pass are scaled up by PASS1_BITS bits so that
    1.70 + * they are represented to better-than-integral precision.  These outputs
    1.71 + * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word
    1.72 + * with the recommended scaling.  (To scale up 12-bit sample data further, an
    1.73 + * intermediate INT32 array would be needed.)
    1.74 + *
    1.75 + * To avoid overflow of the 32-bit intermediate results in pass 2, we must
    1.76 + * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26.  Error analysis
    1.77 + * shows that the values given below are the most effective.
    1.78 + */
    1.79 +
    1.80 +#if BITS_IN_JSAMPLE == 8
    1.81 +#define CONST_BITS  13
    1.82 +#define PASS1_BITS  2
    1.83 +#else
    1.84 +#define CONST_BITS  13
    1.85 +#define PASS1_BITS  1		/* lose a little precision to avoid overflow */
    1.86 +#endif
    1.87 +
    1.88 +/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus
    1.89 + * causing a lot of useless floating-point operations at run time.
    1.90 + * To get around this we use the following pre-calculated constants.
    1.91 + * If you change CONST_BITS you may want to add appropriate values.
    1.92 + * (With a reasonable C compiler, you can just rely on the FIX() macro...)
    1.93 + */
    1.94 +
    1.95 +#if CONST_BITS == 13
    1.96 +#define FIX_0_298631336  ((INT32)  2446)	/* FIX(0.298631336) */
    1.97 +#define FIX_0_390180644  ((INT32)  3196)	/* FIX(0.390180644) */
    1.98 +#define FIX_0_541196100  ((INT32)  4433)	/* FIX(0.541196100) */
    1.99 +#define FIX_0_765366865  ((INT32)  6270)	/* FIX(0.765366865) */
   1.100 +#define FIX_0_899976223  ((INT32)  7373)	/* FIX(0.899976223) */
   1.101 +#define FIX_1_175875602  ((INT32)  9633)	/* FIX(1.175875602) */
   1.102 +#define FIX_1_501321110  ((INT32)  12299)	/* FIX(1.501321110) */
   1.103 +#define FIX_1_847759065  ((INT32)  15137)	/* FIX(1.847759065) */
   1.104 +#define FIX_1_961570560  ((INT32)  16069)	/* FIX(1.961570560) */
   1.105 +#define FIX_2_053119869  ((INT32)  16819)	/* FIX(2.053119869) */
   1.106 +#define FIX_2_562915447  ((INT32)  20995)	/* FIX(2.562915447) */
   1.107 +#define FIX_3_072711026  ((INT32)  25172)	/* FIX(3.072711026) */
   1.108 +#else
   1.109 +#define FIX_0_298631336  FIX(0.298631336)
   1.110 +#define FIX_0_390180644  FIX(0.390180644)
   1.111 +#define FIX_0_541196100  FIX(0.541196100)
   1.112 +#define FIX_0_765366865  FIX(0.765366865)
   1.113 +#define FIX_0_899976223  FIX(0.899976223)
   1.114 +#define FIX_1_175875602  FIX(1.175875602)
   1.115 +#define FIX_1_501321110  FIX(1.501321110)
   1.116 +#define FIX_1_847759065  FIX(1.847759065)
   1.117 +#define FIX_1_961570560  FIX(1.961570560)
   1.118 +#define FIX_2_053119869  FIX(2.053119869)
   1.119 +#define FIX_2_562915447  FIX(2.562915447)
   1.120 +#define FIX_3_072711026  FIX(3.072711026)
   1.121 +#endif
   1.122 +
   1.123 +
   1.124 +/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result.
   1.125 + * For 8-bit samples with the recommended scaling, all the variable
   1.126 + * and constant values involved are no more than 16 bits wide, so a
   1.127 + * 16x16->32 bit multiply can be used instead of a full 32x32 multiply.
   1.128 + * For 12-bit samples, a full 32-bit multiplication will be needed.
   1.129 + */
   1.130 +
   1.131 +#if BITS_IN_JSAMPLE == 8
   1.132 +#define MULTIPLY(var,const)  MULTIPLY16C16(var,const)
   1.133 +#else
   1.134 +#define MULTIPLY(var,const)  ((var) * (const))
   1.135 +#endif
   1.136 +
   1.137 +
   1.138 +/* Dequantize a coefficient by multiplying it by the multiplier-table
   1.139 + * entry; produce an int result.  In this module, both inputs and result
   1.140 + * are 16 bits or less, so either int or short multiply will work.
   1.141 + */
   1.142 +
   1.143 +#define DEQUANTIZE(coef,quantval)  (((ISLOW_MULT_TYPE) (coef)) * (quantval))
   1.144 +
   1.145 +
   1.146 +/*
   1.147 + * Perform dequantization and inverse DCT on one block of coefficients.
   1.148 + */
   1.149 +
   1.150 +GLOBAL(void)
   1.151 +jpeg_idct_islow (j_decompress_ptr cinfo, jpeg_component_info * compptr,
   1.152 +		 JCOEFPTR coef_block,
   1.153 +		 JSAMPARRAY output_buf, JDIMENSION output_col)
   1.154 +{
   1.155 +  INT32 tmp0, tmp1, tmp2, tmp3;
   1.156 +  INT32 tmp10, tmp11, tmp12, tmp13;
   1.157 +  INT32 z1, z2, z3, z4, z5;
   1.158 +  JCOEFPTR inptr;
   1.159 +  ISLOW_MULT_TYPE * quantptr;
   1.160 +  int * wsptr;
   1.161 +  JSAMPROW outptr;
   1.162 +  JSAMPLE *range_limit = IDCT_range_limit(cinfo);
   1.163 +  int ctr;
   1.164 +  int workspace[DCTSIZE2];	/* buffers data between passes */
   1.165 +  SHIFT_TEMPS
   1.166 +
   1.167 +  /* Pass 1: process columns from input, store into work array. */
   1.168 +  /* Note results are scaled up by sqrt(8) compared to a true IDCT; */
   1.169 +  /* furthermore, we scale the results by 2**PASS1_BITS. */
   1.170 +
   1.171 +  inptr = coef_block;
   1.172 +  quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table;
   1.173 +  wsptr = workspace;
   1.174 +  for (ctr = DCTSIZE; ctr > 0; ctr--) {
   1.175 +    /* Due to quantization, we will usually find that many of the input
   1.176 +     * coefficients are zero, especially the AC terms.  We can exploit this
   1.177 +     * by short-circuiting the IDCT calculation for any column in which all
   1.178 +     * the AC terms are zero.  In that case each output is equal to the
   1.179 +     * DC coefficient (with scale factor as needed).
   1.180 +     * With typical images and quantization tables, half or more of the
   1.181 +     * column DCT calculations can be simplified this way.
   1.182 +     */
   1.183 +    
   1.184 +    if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 &&
   1.185 +	inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 &&
   1.186 +	inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 &&
   1.187 +	inptr[DCTSIZE*7] == 0) {
   1.188 +      /* AC terms all zero */
   1.189 +      int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS;
   1.190 +      
   1.191 +      wsptr[DCTSIZE*0] = dcval;
   1.192 +      wsptr[DCTSIZE*1] = dcval;
   1.193 +      wsptr[DCTSIZE*2] = dcval;
   1.194 +      wsptr[DCTSIZE*3] = dcval;
   1.195 +      wsptr[DCTSIZE*4] = dcval;
   1.196 +      wsptr[DCTSIZE*5] = dcval;
   1.197 +      wsptr[DCTSIZE*6] = dcval;
   1.198 +      wsptr[DCTSIZE*7] = dcval;
   1.199 +      
   1.200 +      inptr++;			/* advance pointers to next column */
   1.201 +      quantptr++;
   1.202 +      wsptr++;
   1.203 +      continue;
   1.204 +    }
   1.205 +    
   1.206 +    /* Even part: reverse the even part of the forward DCT. */
   1.207 +    /* The rotator is sqrt(2)*c(-6). */
   1.208 +    
   1.209 +    z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]);
   1.210 +    z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]);
   1.211 +    
   1.212 +    z1 = MULTIPLY(z2 + z3, FIX_0_541196100);
   1.213 +    tmp2 = z1 + MULTIPLY(z3, - FIX_1_847759065);
   1.214 +    tmp3 = z1 + MULTIPLY(z2, FIX_0_765366865);
   1.215 +    
   1.216 +    z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]);
   1.217 +    z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]);
   1.218 +
   1.219 +    tmp0 = (z2 + z3) << CONST_BITS;
   1.220 +    tmp1 = (z2 - z3) << CONST_BITS;
   1.221 +    
   1.222 +    tmp10 = tmp0 + tmp3;
   1.223 +    tmp13 = tmp0 - tmp3;
   1.224 +    tmp11 = tmp1 + tmp2;
   1.225 +    tmp12 = tmp1 - tmp2;
   1.226 +    
   1.227 +    /* Odd part per figure 8; the matrix is unitary and hence its
   1.228 +     * transpose is its inverse.  i0..i3 are y7,y5,y3,y1 respectively.
   1.229 +     */
   1.230 +    
   1.231 +    tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]);
   1.232 +    tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]);
   1.233 +    tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]);
   1.234 +    tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]);
   1.235 +    
   1.236 +    z1 = tmp0 + tmp3;
   1.237 +    z2 = tmp1 + tmp2;
   1.238 +    z3 = tmp0 + tmp2;
   1.239 +    z4 = tmp1 + tmp3;
   1.240 +    z5 = MULTIPLY(z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */
   1.241 +    
   1.242 +    tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */
   1.243 +    tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */
   1.244 +    tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */
   1.245 +    tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */
   1.246 +    z1 = MULTIPLY(z1, - FIX_0_899976223); /* sqrt(2) * (c7-c3) */
   1.247 +    z2 = MULTIPLY(z2, - FIX_2_562915447); /* sqrt(2) * (-c1-c3) */
   1.248 +    z3 = MULTIPLY(z3, - FIX_1_961570560); /* sqrt(2) * (-c3-c5) */
   1.249 +    z4 = MULTIPLY(z4, - FIX_0_390180644); /* sqrt(2) * (c5-c3) */
   1.250 +    
   1.251 +    z3 += z5;
   1.252 +    z4 += z5;
   1.253 +    
   1.254 +    tmp0 += z1 + z3;
   1.255 +    tmp1 += z2 + z4;
   1.256 +    tmp2 += z2 + z3;
   1.257 +    tmp3 += z1 + z4;
   1.258 +    
   1.259 +    /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */
   1.260 +    
   1.261 +    wsptr[DCTSIZE*0] = (int) DESCALE(tmp10 + tmp3, CONST_BITS-PASS1_BITS);
   1.262 +    wsptr[DCTSIZE*7] = (int) DESCALE(tmp10 - tmp3, CONST_BITS-PASS1_BITS);
   1.263 +    wsptr[DCTSIZE*1] = (int) DESCALE(tmp11 + tmp2, CONST_BITS-PASS1_BITS);
   1.264 +    wsptr[DCTSIZE*6] = (int) DESCALE(tmp11 - tmp2, CONST_BITS-PASS1_BITS);
   1.265 +    wsptr[DCTSIZE*2] = (int) DESCALE(tmp12 + tmp1, CONST_BITS-PASS1_BITS);
   1.266 +    wsptr[DCTSIZE*5] = (int) DESCALE(tmp12 - tmp1, CONST_BITS-PASS1_BITS);
   1.267 +    wsptr[DCTSIZE*3] = (int) DESCALE(tmp13 + tmp0, CONST_BITS-PASS1_BITS);
   1.268 +    wsptr[DCTSIZE*4] = (int) DESCALE(tmp13 - tmp0, CONST_BITS-PASS1_BITS);
   1.269 +    
   1.270 +    inptr++;			/* advance pointers to next column */
   1.271 +    quantptr++;
   1.272 +    wsptr++;
   1.273 +  }
   1.274 +  
   1.275 +  /* Pass 2: process rows from work array, store into output array. */
   1.276 +  /* Note that we must descale the results by a factor of 8 == 2**3, */
   1.277 +  /* and also undo the PASS1_BITS scaling. */
   1.278 +
   1.279 +  wsptr = workspace;
   1.280 +  for (ctr = 0; ctr < DCTSIZE; ctr++) {
   1.281 +    outptr = output_buf[ctr] + output_col;
   1.282 +    /* Rows of zeroes can be exploited in the same way as we did with columns.
   1.283 +     * However, the column calculation has created many nonzero AC terms, so
   1.284 +     * the simplification applies less often (typically 5% to 10% of the time).
   1.285 +     * On machines with very fast multiplication, it's possible that the
   1.286 +     * test takes more time than it's worth.  In that case this section
   1.287 +     * may be commented out.
   1.288 +     */
   1.289 +    
   1.290 +#ifndef NO_ZERO_ROW_TEST
   1.291 +    if (wsptr[1] == 0 && wsptr[2] == 0 && wsptr[3] == 0 && wsptr[4] == 0 &&
   1.292 +	wsptr[5] == 0 && wsptr[6] == 0 && wsptr[7] == 0) {
   1.293 +      /* AC terms all zero */
   1.294 +      JSAMPLE dcval = range_limit[(int) DESCALE((INT32) wsptr[0], PASS1_BITS+3)
   1.295 +				  & RANGE_MASK];
   1.296 +      
   1.297 +      outptr[0] = dcval;
   1.298 +      outptr[1] = dcval;
   1.299 +      outptr[2] = dcval;
   1.300 +      outptr[3] = dcval;
   1.301 +      outptr[4] = dcval;
   1.302 +      outptr[5] = dcval;
   1.303 +      outptr[6] = dcval;
   1.304 +      outptr[7] = dcval;
   1.305 +
   1.306 +      wsptr += DCTSIZE;		/* advance pointer to next row */
   1.307 +      continue;
   1.308 +    }
   1.309 +#endif
   1.310 +    
   1.311 +    /* Even part: reverse the even part of the forward DCT. */
   1.312 +    /* The rotator is sqrt(2)*c(-6). */
   1.313 +    
   1.314 +    z2 = (INT32) wsptr[2];
   1.315 +    z3 = (INT32) wsptr[6];
   1.316 +    
   1.317 +    z1 = MULTIPLY(z2 + z3, FIX_0_541196100);
   1.318 +    tmp2 = z1 + MULTIPLY(z3, - FIX_1_847759065);
   1.319 +    tmp3 = z1 + MULTIPLY(z2, FIX_0_765366865);
   1.320 +    
   1.321 +    tmp0 = ((INT32) wsptr[0] + (INT32) wsptr[4]) << CONST_BITS;
   1.322 +    tmp1 = ((INT32) wsptr[0] - (INT32) wsptr[4]) << CONST_BITS;
   1.323 +    
   1.324 +    tmp10 = tmp0 + tmp3;
   1.325 +    tmp13 = tmp0 - tmp3;
   1.326 +    tmp11 = tmp1 + tmp2;
   1.327 +    tmp12 = tmp1 - tmp2;
   1.328 +    
   1.329 +    /* Odd part per figure 8; the matrix is unitary and hence its
   1.330 +     * transpose is its inverse.  i0..i3 are y7,y5,y3,y1 respectively.
   1.331 +     */
   1.332 +    
   1.333 +    tmp0 = (INT32) wsptr[7];
   1.334 +    tmp1 = (INT32) wsptr[5];
   1.335 +    tmp2 = (INT32) wsptr[3];
   1.336 +    tmp3 = (INT32) wsptr[1];
   1.337 +    
   1.338 +    z1 = tmp0 + tmp3;
   1.339 +    z2 = tmp1 + tmp2;
   1.340 +    z3 = tmp0 + tmp2;
   1.341 +    z4 = tmp1 + tmp3;
   1.342 +    z5 = MULTIPLY(z3 + z4, FIX_1_175875602); /* sqrt(2) * c3 */
   1.343 +    
   1.344 +    tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* sqrt(2) * (-c1+c3+c5-c7) */
   1.345 +    tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* sqrt(2) * ( c1+c3-c5+c7) */
   1.346 +    tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* sqrt(2) * ( c1+c3+c5-c7) */
   1.347 +    tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* sqrt(2) * ( c1+c3-c5-c7) */
   1.348 +    z1 = MULTIPLY(z1, - FIX_0_899976223); /* sqrt(2) * (c7-c3) */
   1.349 +    z2 = MULTIPLY(z2, - FIX_2_562915447); /* sqrt(2) * (-c1-c3) */
   1.350 +    z3 = MULTIPLY(z3, - FIX_1_961570560); /* sqrt(2) * (-c3-c5) */
   1.351 +    z4 = MULTIPLY(z4, - FIX_0_390180644); /* sqrt(2) * (c5-c3) */
   1.352 +    
   1.353 +    z3 += z5;
   1.354 +    z4 += z5;
   1.355 +    
   1.356 +    tmp0 += z1 + z3;
   1.357 +    tmp1 += z2 + z4;
   1.358 +    tmp2 += z2 + z3;
   1.359 +    tmp3 += z1 + z4;
   1.360 +    
   1.361 +    /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */
   1.362 +    
   1.363 +    outptr[0] = range_limit[(int) DESCALE(tmp10 + tmp3,
   1.364 +					  CONST_BITS+PASS1_BITS+3)
   1.365 +			    & RANGE_MASK];
   1.366 +    outptr[7] = range_limit[(int) DESCALE(tmp10 - tmp3,
   1.367 +					  CONST_BITS+PASS1_BITS+3)
   1.368 +			    & RANGE_MASK];
   1.369 +    outptr[1] = range_limit[(int) DESCALE(tmp11 + tmp2,
   1.370 +					  CONST_BITS+PASS1_BITS+3)
   1.371 +			    & RANGE_MASK];
   1.372 +    outptr[6] = range_limit[(int) DESCALE(tmp11 - tmp2,
   1.373 +					  CONST_BITS+PASS1_BITS+3)
   1.374 +			    & RANGE_MASK];
   1.375 +    outptr[2] = range_limit[(int) DESCALE(tmp12 + tmp1,
   1.376 +					  CONST_BITS+PASS1_BITS+3)
   1.377 +			    & RANGE_MASK];
   1.378 +    outptr[5] = range_limit[(int) DESCALE(tmp12 - tmp1,
   1.379 +					  CONST_BITS+PASS1_BITS+3)
   1.380 +			    & RANGE_MASK];
   1.381 +    outptr[3] = range_limit[(int) DESCALE(tmp13 + tmp0,
   1.382 +					  CONST_BITS+PASS1_BITS+3)
   1.383 +			    & RANGE_MASK];
   1.384 +    outptr[4] = range_limit[(int) DESCALE(tmp13 - tmp0,
   1.385 +					  CONST_BITS+PASS1_BITS+3)
   1.386 +			    & RANGE_MASK];
   1.387 +    
   1.388 +    wsptr += DCTSIZE;		/* advance pointer to next row */
   1.389 +  }
   1.390 +}
   1.391 +
   1.392 +#endif /* DCT_ISLOW_SUPPORTED */