123 |
int16_t qcoeff[6 * 64], |
int16_t qcoeff[6 * 64], |
124 |
int16_t data[6*64]) |
int16_t data[6*64]) |
125 |
{ |
{ |
126 |
int i; |
int mpeg; |
127 |
|
int scaler_lum, scaler_chr; |
128 |
|
|
129 |
for (i = 0; i < 6; i++) { |
quanth263_intraFuncPtr const quant[2] = |
130 |
uint32_t iDcScaler = get_dc_scaler(pMB->quant, i < 4); |
{ |
131 |
|
(quanth263_intraFuncPtr)quant_intra, |
132 |
|
(quanth263_intraFuncPtr)quant4_intra |
133 |
|
}; |
134 |
|
|
135 |
|
mpeg = !!(pParam->vol_flags & XVID_VOL_MPEGQUANT); |
136 |
|
scaler_lum = get_dc_scaler(pMB->quant, 1); |
137 |
|
scaler_chr = get_dc_scaler(pMB->quant, 0); |
138 |
|
|
139 |
/* Quantize the block */ |
/* Quantize the block */ |
140 |
start_timer(); |
start_timer(); |
141 |
if (!(pParam->vol_flags & XVID_VOL_MPEGQUANT)) { |
quant[mpeg](&data[0 * 64], &qcoeff[0 * 64], pMB->quant, scaler_lum); |
142 |
quant_intra(&data[i * 64], &qcoeff[i * 64], pMB->quant, iDcScaler); |
quant[mpeg](&data[1 * 64], &qcoeff[1 * 64], pMB->quant, scaler_lum); |
143 |
} else { |
quant[mpeg](&data[2 * 64], &qcoeff[2 * 64], pMB->quant, scaler_lum); |
144 |
quant4_intra(&data[i * 64], &qcoeff[i * 64], pMB->quant, iDcScaler); |
quant[mpeg](&data[3 * 64], &qcoeff[3 * 64], pMB->quant, scaler_lum); |
145 |
} |
quant[mpeg](&data[4 * 64], &qcoeff[4 * 64], pMB->quant, scaler_chr); |
146 |
|
quant[mpeg](&data[5 * 64], &qcoeff[5 * 64], pMB->quant, scaler_chr); |
147 |
stop_quant_timer(); |
stop_quant_timer(); |
148 |
} |
} |
|
} |
|
149 |
|
|
150 |
/* DeQuantize all blocks -- Intra mode */ |
/* DeQuantize all blocks -- Intra mode */ |
151 |
static __inline void |
static __inline void |
154 |
int16_t qcoeff[6 * 64], |
int16_t qcoeff[6 * 64], |
155 |
int16_t data[6*64]) |
int16_t data[6*64]) |
156 |
{ |
{ |
157 |
int i; |
int mpeg; |
158 |
|
int scaler_lum, scaler_chr; |
159 |
|
|
160 |
for (i = 0; i < 6; i++) { |
quanth263_intraFuncPtr const dequant[2] = |
161 |
uint32_t iDcScaler = get_dc_scaler(iQuant, i < 4); |
{ |
162 |
|
(quanth263_intraFuncPtr)dequant_intra, |
163 |
|
(quanth263_intraFuncPtr)dequant4_intra |
164 |
|
}; |
165 |
|
|
166 |
|
mpeg = !!(pParam->vol_flags & XVID_VOL_MPEGQUANT); |
167 |
|
scaler_lum = get_dc_scaler(iQuant, 1); |
168 |
|
scaler_chr = get_dc_scaler(iQuant, 0); |
169 |
|
|
170 |
start_timer(); |
start_timer(); |
171 |
if (!(pParam->vol_flags & XVID_VOL_MPEGQUANT)) |
dequant[mpeg](&qcoeff[0 * 64], &data[0 * 64], iQuant, scaler_lum); |
172 |
dequant_intra(&qcoeff[i * 64], &data[i * 64], iQuant, iDcScaler); |
dequant[mpeg](&qcoeff[1 * 64], &data[1 * 64], iQuant, scaler_lum); |
173 |
else |
dequant[mpeg](&qcoeff[2 * 64], &data[2 * 64], iQuant, scaler_lum); |
174 |
dequant4_intra(&qcoeff[i * 64], &data[i * 64], iQuant, iDcScaler); |
dequant[mpeg](&qcoeff[3 * 64], &data[3 * 64], iQuant, scaler_lum); |
175 |
|
dequant[mpeg](&qcoeff[4 * 64], &data[4 * 64], iQuant, scaler_chr); |
176 |
|
dequant[mpeg](&qcoeff[5 * 64], &data[5 * 64], iQuant, scaler_chr); |
177 |
stop_iquant_timer(); |
stop_iquant_timer(); |
178 |
} |
} |
|
} |
|
179 |
|
|
180 |
|
|
181 |
static int |
typedef int (*trellis_func_ptr_t)(int16_t *const Out, |
182 |
dct_quantize_trellis_h263_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero); |
const int16_t *const In, |
183 |
|
int Q, |
184 |
|
const uint16_t * const Zigzag, |
185 |
|
int Non_Zero); |
186 |
|
|
187 |
static int |
static int |
188 |
dct_quantize_trellis_mpeg_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero); |
dct_quantize_trellis_h263_c(int16_t *const Out, |
189 |
|
const int16_t *const In, |
190 |
|
int Q, |
191 |
|
const uint16_t * const Zigzag, |
192 |
|
int Non_Zero); |
193 |
|
|
194 |
|
static int |
195 |
|
dct_quantize_trellis_mpeg_c(int16_t *const Out, |
196 |
|
const int16_t *const In, |
197 |
|
int Q, |
198 |
|
const uint16_t * const Zigzag, |
199 |
|
int Non_Zero); |
200 |
|
|
201 |
/* Quantize all blocks -- Inter mode */ |
/* Quantize all blocks -- Inter mode */ |
202 |
static __inline uint8_t |
static __inline uint8_t |
212 |
int i; |
int i; |
213 |
uint8_t cbp = 0; |
uint8_t cbp = 0; |
214 |
int sum; |
int sum; |
215 |
int code_block; |
int code_block, mpeg; |
216 |
|
|
217 |
|
quanth263_interFuncPtr const quant[2] = |
218 |
|
{ |
219 |
|
(quanth263_interFuncPtr)quant_inter, |
220 |
|
(quanth263_interFuncPtr)quant4_inter |
221 |
|
}; |
222 |
|
|
223 |
|
trellis_func_ptr_t const trellis[2] = |
224 |
|
{ |
225 |
|
(trellis_func_ptr_t)dct_quantize_trellis_h263_c, |
226 |
|
(trellis_func_ptr_t)dct_quantize_trellis_mpeg_c |
227 |
|
}; |
228 |
|
|
229 |
|
mpeg = !!(pParam->vol_flags & XVID_VOL_MPEGQUANT); |
230 |
|
|
231 |
for (i = 0; i < 6; i++) { |
for (i = 0; i < 6; i++) { |
232 |
|
|
233 |
/* Quantize the block */ |
/* Quantize the block */ |
234 |
start_timer(); |
start_timer(); |
235 |
if (!(pParam->vol_flags & XVID_VOL_MPEGQUANT)) { |
|
236 |
sum = quant_inter(&qcoeff[i*64], &data[i*64], pMB->quant); |
sum = quant[mpeg](&qcoeff[i*64], &data[i*64], pMB->quant); |
237 |
if ( (sum) && (frame->vop_flags & XVID_VOP_TRELLISQUANT) ) { |
|
238 |
sum = dct_quantize_trellis_h263_c(&qcoeff[i*64], &data[i*64], pMB->quant, &scan_tables[0][0], 63)+1; |
if(sum && (frame->vop_flags & XVID_VOP_TRELLISQUANT)) { |
239 |
limit = 1; |
sum = trellis[mpeg](&qcoeff[i*64], &data[i*64], pMB->quant, &scan_tables[0][0], 63); |
|
} |
|
|
} else { |
|
|
sum = quant4_inter(&qcoeff[i * 64], &data[i * 64], pMB->quant); |
|
|
// if ( (sum) && (frame->vop_flags & XVID_VOP_TRELLISQUANT) ) |
|
|
// sum = dct_quantize_trellis_mpeg_c (&qcoeff[i*64], &data[i*64], pMB->quant)+1; |
|
240 |
} |
} |
241 |
stop_quant_timer(); |
stop_quant_timer(); |
242 |
|
|
275 |
int16_t qcoeff[6 * 64], |
int16_t qcoeff[6 * 64], |
276 |
const uint8_t cbp) |
const uint8_t cbp) |
277 |
{ |
{ |
278 |
int i; |
int mpeg; |
279 |
|
|
280 |
|
quanth263_interFuncPtr const dequant[2] = |
281 |
|
{ |
282 |
|
(quanth263_interFuncPtr)dequant_inter, |
283 |
|
(quanth263_interFuncPtr)dequant4_inter |
284 |
|
}; |
285 |
|
|
286 |
|
mpeg = !!(pParam->vol_flags & XVID_VOL_MPEGQUANT); |
287 |
|
|
|
for (i = 0; i < 6; i++) { |
|
|
if (cbp & (1 << (5 - i))) { |
|
288 |
start_timer(); |
start_timer(); |
289 |
if (!(pParam->vol_flags & XVID_VOL_MPEGQUANT)) |
if(cbp & (1 << (5 - 0))) dequant[mpeg](&data[0 * 64], &qcoeff[0 * 64], iQuant); |
290 |
dequant_inter(&data[i * 64], &qcoeff[i * 64], iQuant); |
if(cbp & (1 << (5 - 1))) dequant[mpeg](&data[1 * 64], &qcoeff[1 * 64], iQuant); |
291 |
else |
if(cbp & (1 << (5 - 2))) dequant[mpeg](&data[2 * 64], &qcoeff[2 * 64], iQuant); |
292 |
dequant4_inter(&data[i * 64], &qcoeff[i * 64], iQuant); |
if(cbp & (1 << (5 - 3))) dequant[mpeg](&data[3 * 64], &qcoeff[3 * 64], iQuant); |
293 |
|
if(cbp & (1 << (5 - 4))) dequant[mpeg](&data[4 * 64], &qcoeff[4 * 64], iQuant); |
294 |
|
if(cbp & (1 << (5 - 5))) dequant[mpeg](&data[5 * 64], &qcoeff[5 * 64], iQuant); |
295 |
stop_iquant_timer(); |
stop_iquant_timer(); |
296 |
} |
} |
|
} |
|
|
} |
|
297 |
|
|
298 |
typedef void (transfer_operation_8to16_t) (int16_t *Dst, const uint8_t *Src, int BpS); |
typedef void (transfer_operation_8to16_t) (int16_t *Dst, const uint8_t *Src, int BpS); |
299 |
typedef void (transfer_operation_16to8_t) (uint8_t *Dst, const int16_t *Src, int BpS); |
typedef void (transfer_operation_16to8_t) (uint8_t *Dst, const int16_t *Src, int BpS); |
311 |
uint32_t stride2 = stride / 2; |
uint32_t stride2 = stride / 2; |
312 |
uint32_t next_block = stride * 8; |
uint32_t next_block = stride * 8; |
313 |
int32_t cst; |
int32_t cst; |
314 |
|
int vop_reduced; |
315 |
uint8_t *pY_Cur, *pU_Cur, *pV_Cur; |
uint8_t *pY_Cur, *pU_Cur, *pV_Cur; |
316 |
const IMAGE * const pCurrent = &frame->image; |
const IMAGE * const pCurrent = &frame->image; |
317 |
|
transfer_operation_8to16_t * const functions[2] = |
318 |
|
{ |
319 |
|
(transfer_operation_8to16_t *)transfer_8to16copy, |
320 |
|
(transfer_operation_8to16_t *)filter_18x18_to_8x8 |
321 |
|
}; |
322 |
transfer_operation_8to16_t *transfer_op = NULL; |
transfer_operation_8to16_t *transfer_op = NULL; |
323 |
|
|
324 |
if ((frame->vop_flags & XVID_VOP_REDUCED)) { |
vop_reduced = !!(frame->vop_flags & XVID_VOP_REDUCED); |
325 |
|
|
326 |
/* Image pointers */ |
/* Image pointers */ |
327 |
pY_Cur = pCurrent->y + (y_pos << 5) * stride + (x_pos << 5); |
pY_Cur = pCurrent->y + (y_pos << (4+vop_reduced)) * stride + (x_pos << (4+vop_reduced)); |
328 |
pU_Cur = pCurrent->u + (y_pos << 4) * stride2 + (x_pos << 4); |
pU_Cur = pCurrent->u + (y_pos << (3+vop_reduced)) * stride2 + (x_pos << (3+vop_reduced)); |
329 |
pV_Cur = pCurrent->v + (y_pos << 4) * stride2 + (x_pos << 4); |
pV_Cur = pCurrent->v + (y_pos << (3+vop_reduced)) * stride2 + (x_pos << (3+vop_reduced)); |
330 |
|
|
331 |
/* Block size */ |
/* Block size */ |
332 |
cst = 16; |
cst = 8<<vop_reduced; |
333 |
|
|
334 |
/* Operation function */ |
/* Operation function */ |
335 |
transfer_op = (transfer_operation_8to16_t*)filter_18x18_to_8x8; |
transfer_op = functions[vop_reduced]; |
|
} else { |
|
|
|
|
|
/* Image pointers */ |
|
|
pY_Cur = pCurrent->y + (y_pos << 4) * stride + (x_pos << 4); |
|
|
pU_Cur = pCurrent->u + (y_pos << 3) * stride2 + (x_pos << 3); |
|
|
pV_Cur = pCurrent->v + (y_pos << 3) * stride2 + (x_pos << 3); |
|
|
|
|
|
/* Block size */ |
|
|
cst = 8; |
|
|
|
|
|
/* Operation function */ |
|
|
transfer_op = (transfer_operation_8to16_t*)transfer_8to16copy; |
|
|
} |
|
336 |
|
|
337 |
/* Do the transfer */ |
/* Do the transfer */ |
338 |
start_timer(); |
start_timer(); |
352 |
const uint32_t x_pos, |
const uint32_t x_pos, |
353 |
const uint32_t y_pos, |
const uint32_t y_pos, |
354 |
int16_t data[6 * 64], |
int16_t data[6 * 64], |
355 |
const uint32_t add, |
const uint32_t add, /* Must be 1 or 0 */ |
356 |
const uint8_t cbp) |
const uint8_t cbp) |
357 |
{ |
{ |
358 |
uint8_t *pY_Cur, *pU_Cur, *pV_Cur; |
uint8_t *pY_Cur, *pU_Cur, *pV_Cur; |
360 |
uint32_t stride2 = stride / 2; |
uint32_t stride2 = stride / 2; |
361 |
uint32_t next_block = stride * 8; |
uint32_t next_block = stride * 8; |
362 |
uint32_t cst; |
uint32_t cst; |
363 |
|
int vop_reduced; |
364 |
const IMAGE * const pCurrent = &frame->image; |
const IMAGE * const pCurrent = &frame->image; |
365 |
|
/* Array of function pointers, indexed by [vop_reduced<<1+add] */ |
366 |
|
transfer_operation_16to8_t * const functions[4] = |
367 |
|
{ |
368 |
|
(transfer_operation_16to8_t*)transfer_16to8copy, |
369 |
|
(transfer_operation_16to8_t*)transfer_16to8add, |
370 |
|
(transfer_operation_16to8_t*)copy_upsampled_8x8_16to8, |
371 |
|
(transfer_operation_16to8_t*)add_upsampled_8x8_16to8 |
372 |
|
}; |
373 |
|
|
374 |
transfer_operation_16to8_t *transfer_op = NULL; |
transfer_operation_16to8_t *transfer_op = NULL; |
375 |
|
|
376 |
if (pMB->field_dct) { |
if (pMB->field_dct) { |
378 |
stride *= 2; |
stride *= 2; |
379 |
} |
} |
380 |
|
|
381 |
if ((frame->vop_flags & XVID_VOP_REDUCED)) { |
/* Makes this vars booleans */ |
382 |
|
vop_reduced = !!(frame->vop_flags & XVID_VOP_REDUCED); |
383 |
|
|
384 |
/* Image pointers */ |
/* Image pointers */ |
385 |
pY_Cur = pCurrent->y + (y_pos << 5) * stride + (x_pos << 5); |
pY_Cur = pCurrent->y + (y_pos << (4+vop_reduced)) * stride + (x_pos << (4+vop_reduced)); |
386 |
pU_Cur = pCurrent->u + (y_pos << 4) * stride2 + (x_pos << 4); |
pU_Cur = pCurrent->u + (y_pos << (3+vop_reduced)) * stride2 + (x_pos << (3+vop_reduced)); |
387 |
pV_Cur = pCurrent->v + (y_pos << 4) * stride2 + (x_pos << 4); |
pV_Cur = pCurrent->v + (y_pos << (3+vop_reduced)) * stride2 + (x_pos << (3+vop_reduced)); |
388 |
|
|
389 |
/* Block size */ |
/* Block size */ |
390 |
cst = 16; |
cst = 8<<vop_reduced; |
391 |
|
|
392 |
/* Operation function */ |
/* Operation function */ |
393 |
if(add) |
transfer_op = functions[(vop_reduced<<1) + add]; |
|
transfer_op = (transfer_operation_16to8_t*)add_upsampled_8x8_16to8; |
|
|
else |
|
|
transfer_op = (transfer_operation_16to8_t*)copy_upsampled_8x8_16to8; |
|
|
} else { |
|
|
|
|
|
/* Image pointers */ |
|
|
pY_Cur = pCurrent->y + (y_pos << 4) * stride + (x_pos << 4); |
|
|
pU_Cur = pCurrent->u + (y_pos << 3) * stride2 + (x_pos << 3); |
|
|
pV_Cur = pCurrent->v + (y_pos << 3) * stride2 + (x_pos << 3); |
|
|
|
|
|
/* Block size */ |
|
|
cst = 8; |
|
|
|
|
|
/* Operation function */ |
|
|
if(add) |
|
|
transfer_op = (transfer_operation_16to8_t*)transfer_16to8add; |
|
|
else |
|
|
transfer_op = (transfer_operation_16to8_t*)transfer_16to8copy; |
|
|
} |
|
394 |
|
|
395 |
/* Do the operation */ |
/* Do the operation */ |
396 |
start_timer(); |
start_timer(); |
460 |
/* Set the limit threshold */ |
/* Set the limit threshold */ |
461 |
limit = PVOP_TOOSMALL_LIMIT + ((pMB->quant == 1)? 1 : 0); |
limit = PVOP_TOOSMALL_LIMIT + ((pMB->quant == 1)? 1 : 0); |
462 |
|
|
463 |
|
if (frame->vop_flags & XVID_VOP_CARTOON) |
464 |
|
limit *= 3; |
465 |
|
|
466 |
/* Quantize the block */ |
/* Quantize the block */ |
467 |
cbp = MBQuantInter(pParam, frame, pMB, data, qcoeff, 0, limit); |
cbp = MBQuantInter(pParam, frame, pMB, data, qcoeff, 0, limit); |
468 |
|
|
501 |
/* Set the limit threshold */ |
/* Set the limit threshold */ |
502 |
limit = BVOP_TOOSMALL_LIMIT; |
limit = BVOP_TOOSMALL_LIMIT; |
503 |
|
|
504 |
|
if (frame->vop_flags & XVID_VOP_CARTOON) |
505 |
|
limit *= 2; |
506 |
|
|
507 |
/* Quantize the block */ |
/* Quantize the block */ |
508 |
cbp = MBQuantInter(pParam, frame, pMB, data, qcoeff, 1, limit); |
cbp = MBQuantInter(pParam, frame, pMB, data, qcoeff, 1, limit); |
509 |
|
|
582 |
|
|
583 |
/* left blocks */ |
/* left blocks */ |
584 |
|
|
585 |
// 1=2, 2=4, 4=8, 8=1 |
/* 1=2, 2=4, 4=8, 8=1 */ |
586 |
MOVLINE(tmp, LINE(0, 1)); |
MOVLINE(tmp, LINE(0, 1)); |
587 |
MOVLINE(LINE(0, 1), LINE(0, 2)); |
MOVLINE(LINE(0, 1), LINE(0, 2)); |
588 |
MOVLINE(LINE(0, 2), LINE(0, 4)); |
MOVLINE(LINE(0, 2), LINE(0, 4)); |
589 |
MOVLINE(LINE(0, 4), LINE(2, 0)); |
MOVLINE(LINE(0, 4), LINE(2, 0)); |
590 |
MOVLINE(LINE(2, 0), tmp); |
MOVLINE(LINE(2, 0), tmp); |
591 |
|
|
592 |
// 3=6, 6=12, 12=9, 9=3 |
/* 3=6, 6=12, 12=9, 9=3 */ |
593 |
MOVLINE(tmp, LINE(0, 3)); |
MOVLINE(tmp, LINE(0, 3)); |
594 |
MOVLINE(LINE(0, 3), LINE(0, 6)); |
MOVLINE(LINE(0, 3), LINE(0, 6)); |
595 |
MOVLINE(LINE(0, 6), LINE(2, 4)); |
MOVLINE(LINE(0, 6), LINE(2, 4)); |
596 |
MOVLINE(LINE(2, 4), LINE(2, 1)); |
MOVLINE(LINE(2, 4), LINE(2, 1)); |
597 |
MOVLINE(LINE(2, 1), tmp); |
MOVLINE(LINE(2, 1), tmp); |
598 |
|
|
599 |
// 5=10, 10=5 |
/* 5=10, 10=5 */ |
600 |
MOVLINE(tmp, LINE(0, 5)); |
MOVLINE(tmp, LINE(0, 5)); |
601 |
MOVLINE(LINE(0, 5), LINE(2, 2)); |
MOVLINE(LINE(0, 5), LINE(2, 2)); |
602 |
MOVLINE(LINE(2, 2), tmp); |
MOVLINE(LINE(2, 2), tmp); |
603 |
|
|
604 |
// 7=14, 14=13, 13=11, 11=7 |
/* 7=14, 14=13, 13=11, 11=7 */ |
605 |
MOVLINE(tmp, LINE(0, 7)); |
MOVLINE(tmp, LINE(0, 7)); |
606 |
MOVLINE(LINE(0, 7), LINE(2, 6)); |
MOVLINE(LINE(0, 7), LINE(2, 6)); |
607 |
MOVLINE(LINE(2, 6), LINE(2, 5)); |
MOVLINE(LINE(2, 6), LINE(2, 5)); |
610 |
|
|
611 |
/* right blocks */ |
/* right blocks */ |
612 |
|
|
613 |
// 1=2, 2=4, 4=8, 8=1 |
/* 1=2, 2=4, 4=8, 8=1 */ |
614 |
MOVLINE(tmp, LINE(1, 1)); |
MOVLINE(tmp, LINE(1, 1)); |
615 |
MOVLINE(LINE(1, 1), LINE(1, 2)); |
MOVLINE(LINE(1, 1), LINE(1, 2)); |
616 |
MOVLINE(LINE(1, 2), LINE(1, 4)); |
MOVLINE(LINE(1, 2), LINE(1, 4)); |
617 |
MOVLINE(LINE(1, 4), LINE(3, 0)); |
MOVLINE(LINE(1, 4), LINE(3, 0)); |
618 |
MOVLINE(LINE(3, 0), tmp); |
MOVLINE(LINE(3, 0), tmp); |
619 |
|
|
620 |
// 3=6, 6=12, 12=9, 9=3 |
/* 3=6, 6=12, 12=9, 9=3 */ |
621 |
MOVLINE(tmp, LINE(1, 3)); |
MOVLINE(tmp, LINE(1, 3)); |
622 |
MOVLINE(LINE(1, 3), LINE(1, 6)); |
MOVLINE(LINE(1, 3), LINE(1, 6)); |
623 |
MOVLINE(LINE(1, 6), LINE(3, 4)); |
MOVLINE(LINE(1, 6), LINE(3, 4)); |
624 |
MOVLINE(LINE(3, 4), LINE(3, 1)); |
MOVLINE(LINE(3, 4), LINE(3, 1)); |
625 |
MOVLINE(LINE(3, 1), tmp); |
MOVLINE(LINE(3, 1), tmp); |
626 |
|
|
627 |
// 5=10, 10=5 |
/* 5=10, 10=5 */ |
628 |
MOVLINE(tmp, LINE(1, 5)); |
MOVLINE(tmp, LINE(1, 5)); |
629 |
MOVLINE(LINE(1, 5), LINE(3, 2)); |
MOVLINE(LINE(1, 5), LINE(3, 2)); |
630 |
MOVLINE(LINE(3, 2), tmp); |
MOVLINE(LINE(3, 2), tmp); |
631 |
|
|
632 |
// 7=14, 14=13, 13=11, 11=7 |
/* 7=14, 14=13, 13=11, 11=7 */ |
633 |
MOVLINE(tmp, LINE(1, 7)); |
MOVLINE(tmp, LINE(1, 7)); |
634 |
MOVLINE(LINE(1, 7), LINE(3, 6)); |
MOVLINE(LINE(1, 7), LINE(3, 6)); |
635 |
MOVLINE(LINE(3, 6), LINE(3, 5)); |
MOVLINE(LINE(3, 6), LINE(3, 5)); |
641 |
|
|
642 |
|
|
643 |
|
|
644 |
/************************************************************************ |
/***************************************************************************** |
645 |
* Trellis based R-D optimal quantization * |
* Trellis based R-D optimal quantization |
646 |
* * |
* |
647 |
* Trellis Quant code (C) 2003 Pascal Massimino skal(at)planet-d.net * |
* Trellis Quant code (C) 2003 Pascal Massimino skal(at)planet-d.net |
648 |
* * |
* |
649 |
************************************************************************/ |
****************************************************************************/ |
650 |
|
|
651 |
|
|
652 |
|
#if 0 |
653 |
static int |
static int |
654 |
dct_quantize_trellis_mpeg_c(int16_t *const Out, const int16_t *const In, int Q, |
dct_quantize_trellis_mpeg_c(int16_t *const Out, |
655 |
const uint16_t * const Zigzag, int Non_Zero) |
const int16_t *const In, |
656 |
{ return 63; } |
int Q, |
657 |
|
const uint16_t * const Zigzag, |
658 |
|
int Non_Zero) |
659 |
////////////////////////////////////////////////////////// |
{ |
660 |
// |
return 63; |
661 |
// Trellis-Based quantization |
} |
662 |
// |
#endif |
663 |
// So far I understand this paper: |
|
664 |
// |
/*---------------------------------------------------------------------------- |
665 |
// "Trellis-Based R-D Optimal Quantization in H.263+" |
* |
666 |
// J.Wen, M.Luttrell, J.Villasenor |
* Trellis-Based quantization |
667 |
// IEEE Transactions on Image Processing, Vol.9, No.8, Aug. 2000. |
* |
668 |
// |
* So far I understand this paper: |
669 |
// we are at stake with a simplified Bellmand-Ford / Dijkstra Single |
* |
670 |
// Source Shorted Path algo. But due to the underlying graph structure |
* "Trellis-Based R-D Optimal Quantization in H.263+" |
671 |
// ("Trellis"), it can be turned into a dynamic programming algo, |
* J.Wen, M.Luttrell, J.Villasenor |
672 |
// partially saving the explicit graph's nodes representation. And |
* IEEE Transactions on Image Processing, Vol.9, No.8, Aug. 2000. |
673 |
// without using a heap, since the open frontier of the DAG is always |
* |
674 |
// known, and of fixed sized. |
* we are at stake with a simplified Bellmand-Ford / Dijkstra Single |
675 |
// |
* Source Shorted Path algo. But due to the underlying graph structure |
676 |
////////////////////////////////////////////////////////// |
* ("Trellis"), it can be turned into a dynamic programming algo, |
677 |
|
* partially saving the explicit graph's nodes representation. And |
678 |
|
* without using a heap, since the open frontier of the DAG is always |
679 |
|
* known, and of fixed sized. |
680 |
|
*--------------------------------------------------------------------------*/ |
681 |
|
|
682 |
|
|
|
////////////////////////////////////////////////////////// |
|
|
// Codes lengths for relevant levels. |
|
683 |
|
|
684 |
// let's factorize: |
/* Codes lengths for relevant levels. */ |
685 |
|
|
686 |
|
/* let's factorize: */ |
687 |
static const uint8_t Code_Len0[64] = { |
static const uint8_t Code_Len0[64] = { |
688 |
30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30, |
30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30, |
689 |
30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30 }; |
30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30 }; |
748 |
3, 4, 5, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 9, 9,10,10,10,10,10,10,10,10,12,12,13,13,12,13,14,15,15, |
3, 4, 5, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 9, 9,10,10,10,10,10,10,10,10,12,12,13,13,12,13,14,15,15, |
749 |
15,16,16,16,16,17,17,17,18,18,19,19,19,19,19,19,19,19,21,21,22,22,30,30,30,30,30,30,30,30,30,30 }; |
15,16,16,16,16,17,17,17,18,18,19,19,19,19,19,19,19,19,21,21,22,22,30,30,30,30,30,30,30,30,30,30 }; |
750 |
|
|
751 |
// a few more table for LAST table: |
/* a few more table for LAST table: */ |
752 |
static const uint8_t Code_Len21[64] = { |
static const uint8_t Code_Len21[64] = { |
753 |
13,20,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30, |
13,20,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30, |
754 |
30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30}; |
30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30,30}; |
763 |
12,13,13,13,13,13,13,13,13,14,16,16,16,16,17,17,17,17,18,18,18,18,18,18,18,18,19,19,19,19,19,19}; |
12,13,13,13,13,13,13,13,13,14,16,16,16,16,17,17,17,17,18,18,18,18,18,18,18,18,19,19,19,19,19,19}; |
764 |
|
|
765 |
|
|
766 |
static const uint8_t * const B16_17_Code_Len[24] = { // levels [1..24] |
static const uint8_t * const B16_17_Code_Len[24] = { /* levels [1..24] */ |
767 |
Code_Len20,Code_Len19,Code_Len18,Code_Len17, |
Code_Len20,Code_Len19,Code_Len18,Code_Len17, |
768 |
Code_Len16,Code_Len15,Code_Len14,Code_Len13, |
Code_Len16,Code_Len15,Code_Len14,Code_Len13, |
769 |
Code_Len12,Code_Len11,Code_Len10,Code_Len9, |
Code_Len12,Code_Len11,Code_Len10,Code_Len9, |
772 |
Code_Len2, Code_Len1, Code_Len1, Code_Len1, |
Code_Len2, Code_Len1, Code_Len1, Code_Len1, |
773 |
}; |
}; |
774 |
|
|
775 |
static const uint8_t * const B16_17_Code_Len_Last[6] = { // levels [1..6] |
static const uint8_t * const B16_17_Code_Len_Last[6] = { /* levels [1..6] */ |
776 |
Code_Len24,Code_Len23,Code_Len22,Code_Len21, Code_Len3, Code_Len1, |
Code_Len24,Code_Len23,Code_Len22,Code_Len21, Code_Len3, Code_Len1, |
777 |
}; |
}; |
778 |
|
|
786 |
}; |
}; |
787 |
#undef TL |
#undef TL |
788 |
|
|
789 |
static __inline int Find_Last(const int16_t *C, const uint16_t *Zigzag, int i) |
static int __inline |
790 |
|
Find_Last(const int16_t *C, const uint16_t *Zigzag, int i) |
791 |
{ |
{ |
792 |
while(i>=0) |
while(i>=0) |
793 |
if (C[Zigzag[i]]) |
if (C[Zigzag[i]]) |
796 |
return -1; |
return -1; |
797 |
} |
} |
798 |
|
|
799 |
////////////////////////////////////////////////////////// |
static int __inline |
800 |
// this routine has been strippen of all debug code |
Compute_Sum(const int16_t *C, int last) |
801 |
////////////////////////////////////////////////////////// |
{ |
802 |
|
int sum = 0; |
803 |
|
|
804 |
|
while(last--) |
805 |
|
sum += abs(C[last]); |
806 |
|
|
807 |
|
return(sum); |
808 |
|
} |
809 |
|
/* this routine has been strippen of all debug code */ |
810 |
|
|
811 |
static int |
static int |
812 |
dct_quantize_trellis_h263_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero) |
dct_quantize_trellis_h263_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero) |
813 |
{ |
{ |
814 |
|
|
815 |
// Note: We should search last non-zero coeffs on *real* DCT input coeffs (In[]), |
/* |
816 |
// not quantized one (Out[]). However, it only improves the result *very* |
* Note: We should search last non-zero coeffs on *real* DCT input coeffs (In[]), |
817 |
// slightly (~0.01dB), whereas speed drops to crawling level :) |
* not quantized one (Out[]). However, it only improves the result *very* |
818 |
// Well, actually, taking 1 more coeff past Non_Zero into account sometimes helps, |
* slightly (~0.01dB), whereas speed drops to crawling level :) |
819 |
|
* Well, actually, taking 1 more coeff past Non_Zero into account sometimes helps. |
820 |
|
*/ |
821 |
typedef struct { int16_t Run, Level; } NODE; |
typedef struct { int16_t Run, Level; } NODE; |
822 |
|
|
823 |
NODE Nodes[65], Last; |
NODE Nodes[65], Last; |
826 |
const int Mult = 2*Q; |
const int Mult = 2*Q; |
827 |
const int Bias = (Q-1) | 1; |
const int Bias = (Q-1) | 1; |
828 |
const int Lev0 = Mult + Bias; |
const int Lev0 = Mult + Bias; |
829 |
const int Lambda = Trellis_Lambda_Tabs[Q-1]; // it's 1/lambda, actually |
const int Lambda = Trellis_Lambda_Tabs[Q-1]; /* it's 1/lambda, actually */ |
830 |
|
|
831 |
int Run_Start = -1; |
int Run_Start = -1; |
832 |
uint32_t Min_Cost = 2<<16; |
uint32_t Min_Cost = 2<<16; |
834 |
int Last_Node = -1; |
int Last_Node = -1; |
835 |
uint32_t Last_Cost = 0; |
uint32_t Last_Cost = 0; |
836 |
|
|
837 |
int i, j; |
int i, j, sum; |
838 |
Run_Costs[-1] = 2<<16; // source (w/ CBP penalty) |
Run_Costs[-1] = 2<<16; /* source (w/ CBP penalty) */ |
839 |
|
|
840 |
Non_Zero = Find_Last(Out, Zigzag, Non_Zero); |
Non_Zero = Find_Last(Out, Zigzag, Non_Zero); |
841 |
if (Non_Zero<0) |
if (Non_Zero<0) |
842 |
return -1; |
return 0; /* Sum is zero if there are only zero coeffs */ |
843 |
|
|
844 |
for(i=0; i<=Non_Zero; i++) |
for(i=0; i<=Non_Zero; i++) { |
|
{ |
|
845 |
const int AC = In[Zigzag[i]]; |
const int AC = In[Zigzag[i]]; |
846 |
const int Level1 = Out[Zigzag[i]]; |
const int Level1 = Out[Zigzag[i]]; |
847 |
const int Dist0 = Lambda* AC*AC; |
const int Dist0 = Lambda* AC*AC; |
848 |
uint32_t Best_Cost = 0xf0000000; |
uint32_t Best_Cost = 0xf0000000; |
849 |
Last_Cost += Dist0; |
Last_Cost += Dist0; |
850 |
|
|
851 |
if ((uint32_t)(Level1+1)<3) // very specialized loop for -1,0,+1 |
/* very specialized loop for -1,0,+1 */ |
852 |
{ |
if ((uint32_t)(Level1+1)<3) { |
853 |
int dQ; |
int dQ; |
854 |
int Run; |
int Run; |
855 |
uint32_t Cost0; |
uint32_t Cost0; |
865 |
|
|
866 |
Nodes[i].Run = 1; |
Nodes[i].Run = 1; |
867 |
Best_Cost = (Code_Len20[0]<<16) + Run_Costs[i-1]+Cost0; |
Best_Cost = (Code_Len20[0]<<16) + Run_Costs[i-1]+Cost0; |
868 |
for(Run=i-Run_Start; Run>0; --Run) |
for(Run=i-Run_Start; Run>0; --Run) { |
|
{ |
|
869 |
const uint32_t Cost_Base = Cost0 + Run_Costs[i-Run]; |
const uint32_t Cost_Base = Cost0 + Run_Costs[i-Run]; |
870 |
const uint32_t Cost = Cost_Base + (Code_Len20[Run-1]<<16); |
const uint32_t Cost = Cost_Base + (Code_Len20[Run-1]<<16); |
871 |
const uint32_t lCost = Cost_Base + (Code_Len24[Run-1]<<16); |
const uint32_t lCost = Cost_Base + (Code_Len24[Run-1]<<16); |
872 |
|
|
873 |
// TODO: what about tie-breaks? Should we favor short runs or |
/* |
874 |
// long runs? Although the error is the same, it would not be |
* TODO: what about tie-breaks? Should we favor short runs or |
875 |
// spread the same way along high and low frequencies... |
* long runs? Although the error is the same, it would not be |
876 |
|
* spread the same way along high and low frequencies... |
877 |
|
*/ |
878 |
|
|
879 |
// (I'd say: favour short runs => hifreq errors (HVS) -- gruel ) |
/* (I'd say: favour short runs => hifreq errors (HVS) -- gruel ) */ |
880 |
|
|
881 |
if (Cost<Best_Cost) { |
if (Cost<Best_Cost) { |
882 |
Best_Cost = Cost; |
Best_Cost = Cost; |
891 |
} |
} |
892 |
if (Last_Node==i) |
if (Last_Node==i) |
893 |
Last.Level = Nodes[i].Level; |
Last.Level = Nodes[i].Level; |
894 |
} |
} else { /* "big" levels */ |
|
else // "big" levels |
|
|
{ |
|
895 |
const uint8_t *Tbl_L1, *Tbl_L2, *Tbl_L1_Last, *Tbl_L2_Last; |
const uint8_t *Tbl_L1, *Tbl_L2, *Tbl_L1_Last, *Tbl_L2_Last; |
896 |
int Level2; |
int Level2; |
897 |
int dQ1, dQ2; |
int dQ1, dQ2; |
907 |
Tbl_L2 = (Level2<=24) ? B16_17_Code_Len[Level2-1] : Code_Len0; |
Tbl_L2 = (Level2<=24) ? B16_17_Code_Len[Level2-1] : Code_Len0; |
908 |
Tbl_L1_Last = (Level1<=6) ? B16_17_Code_Len_Last[Level1-1] : Code_Len0; |
Tbl_L1_Last = (Level1<=6) ? B16_17_Code_Len_Last[Level1-1] : Code_Len0; |
909 |
Tbl_L2_Last = (Level2<=6) ? B16_17_Code_Len_Last[Level2-1] : Code_Len0; |
Tbl_L2_Last = (Level2<=6) ? B16_17_Code_Len_Last[Level2-1] : Code_Len0; |
910 |
} else { // Level1<-1 |
} else { /* Level1<-1 */ |
911 |
dQ1 = Level1*Mult-AC - Bias; |
dQ1 = Level1*Mult-AC - Bias; |
912 |
dQ2 = dQ1 + Mult; |
dQ2 = dQ1 + Mult; |
913 |
Level2 = Level1 + 1; |
Level2 = Level1 + 1; |
916 |
Tbl_L1_Last = (Level1>=- 6) ? B16_17_Code_Len_Last[Level1^-1] : Code_Len0; |
Tbl_L1_Last = (Level1>=- 6) ? B16_17_Code_Len_Last[Level1^-1] : Code_Len0; |
917 |
Tbl_L2_Last = (Level2>=- 6) ? B16_17_Code_Len_Last[Level2^-1] : Code_Len0; |
Tbl_L2_Last = (Level2>=- 6) ? B16_17_Code_Len_Last[Level2^-1] : Code_Len0; |
918 |
} |
} |
919 |
|
|
920 |
Dist1 = Lambda*dQ1*dQ1; |
Dist1 = Lambda*dQ1*dQ1; |
921 |
Dist2 = Lambda*dQ2*dQ2; |
Dist2 = Lambda*dQ2*dQ2; |
922 |
dDist21 = Dist2-Dist1; |
dDist21 = Dist2-Dist1; |
927 |
uint32_t Cost1, Cost2; |
uint32_t Cost1, Cost2; |
928 |
int bLevel; |
int bLevel; |
929 |
|
|
930 |
// for sub-optimal (but slightly worth it, speed-wise) search, uncomment the following: |
/* |
931 |
// if (Cost_Base>=Best_Cost) continue; |
* for sub-optimal (but slightly worth it, speed-wise) search, uncomment the following: |
932 |
// (? doesn't seem to have any effect -- gruel ) |
* if (Cost_Base>=Best_Cost) continue; |
933 |
|
* (? doesn't seem to have any effect -- gruel ) |
934 |
|
*/ |
935 |
|
|
936 |
Cost1 = Cost_Base + (Tbl_L1[Run-1]<<16); |
Cost1 = Cost_Base + (Tbl_L1[Run-1]<<16); |
937 |
Cost2 = Cost_Base + (Tbl_L2[Run-1]<<16) + dDist21; |
Cost2 = Cost_Base + (Tbl_L2[Run-1]<<16) + dDist21; |
939 |
if (Cost2<Cost1) { |
if (Cost2<Cost1) { |
940 |
Cost1 = Cost2; |
Cost1 = Cost2; |
941 |
bLevel = Level2; |
bLevel = Level2; |
942 |
} else |
} else { |
943 |
bLevel = Level1; |
bLevel = Level1; |
944 |
|
} |
945 |
|
|
946 |
if (Cost1<Best_Cost) { |
if (Cost1<Best_Cost) { |
947 |
Best_Cost = Cost1; |
Best_Cost = Cost1; |
955 |
if (Cost2<Cost1) { |
if (Cost2<Cost1) { |
956 |
Cost1 = Cost2; |
Cost1 = Cost2; |
957 |
bLevel = Level2; |
bLevel = Level2; |
958 |
} else |
} else { |
959 |
bLevel = Level1; |
bLevel = Level1; |
960 |
|
} |
961 |
|
|
962 |
if (Cost1<Last_Cost) { |
if (Cost1<Last_Cost) { |
963 |
Last_Cost = Cost1; |
Last_Cost = Cost1; |
965 |
Last.Level = bLevel; |
Last.Level = bLevel; |
966 |
Last_Node = i; |
Last_Node = i; |
967 |
} |
} |
968 |
} //end of "for Run" |
} /* end of "for Run" */ |
969 |
|
|
970 |
} |
} |
971 |
|
|
974 |
if (Best_Cost < Min_Cost + Dist0) { |
if (Best_Cost < Min_Cost + Dist0) { |
975 |
Min_Cost = Best_Cost; |
Min_Cost = Best_Cost; |
976 |
Run_Start = i; |
Run_Start = i; |
977 |
} |
} else { |
978 |
else |
/* |
979 |
{ |
* as noticed by Michael Niedermayer (michaelni at gmx.at), there's |
980 |
// as noticed by Michael Niedermayer (michaelni at gmx.at), there's |
* a code shorter by 1 bit for a larger run (!), same level. We give |
981 |
// a code shorter by 1 bit for a larger run (!), same level. We give |
* it a chance by not moving the left barrier too much. |
982 |
// it a chance by not moving the left barrier too much. |
*/ |
983 |
|
|
984 |
while( Run_Costs[Run_Start]>Min_Cost+(1<<16) ) |
while( Run_Costs[Run_Start]>Min_Cost+(1<<16) ) |
985 |
Run_Start++; |
Run_Start++; |
986 |
|
|
987 |
// spread on preceding coeffs the cost incurred by skipping this one |
/* spread on preceding coeffs the cost incurred by skipping this one */ |
988 |
for(j=Run_Start; j<i; ++j) Run_Costs[j] += Dist0; |
for(j=Run_Start; j<i; ++j) Run_Costs[j] += Dist0; |
989 |
Min_Cost += Dist0; |
Min_Cost += Dist0; |
990 |
} |
} |
991 |
} |
} |
992 |
|
|
993 |
|
/* It seems trellis doesn't give good results... just compute the Out sum and |
994 |
|
* quit (even if we did not modify it, upperlayer relies on this data) */ |
995 |
if (Last_Node<0) |
if (Last_Node<0) |
996 |
return -1; |
return Compute_Sum(Out, Non_Zero); |
997 |
|
|
998 |
// reconstruct optimal sequence backward with surviving paths |
/* reconstruct optimal sequence backward with surviving paths */ |
999 |
memset(Out, 0x00, 64*sizeof(*Out)); |
memset(Out, 0x00, 64*sizeof(*Out)); |
1000 |
Out[Zigzag[Last_Node]] = Last.Level; |
Out[Zigzag[Last_Node]] = Last.Level; |
1001 |
i = Last_Node - Last.Run; |
i = Last_Node - Last.Run; |
1002 |
|
sum = 0; |
1003 |
while(i>=0) { |
while(i>=0) { |
1004 |
Out[Zigzag[i]] = Nodes[i].Level; |
Out[Zigzag[i]] = Nodes[i].Level; |
1005 |
|
sum += abs(Nodes[i].Level); |
1006 |
i -= Nodes[i].Run; |
i -= Nodes[i].Run; |
1007 |
} |
} |
|
return Last_Node; |
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1008 |
|
|
1009 |
|
return sum; |
1010 |
|
} |
1011 |
|
|
1012 |
////////////////////////////////////////////////////////// |
static int |
1013 |
// original version including heavy debugging info |
dct_quantize_trellis_mpeg_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero) |
1014 |
////////////////////////////////////////////////////////// |
{ |
1015 |
|
/* ToDo: Ok ok it's just a place holder for Gruel -- damn write this one :-) */ |
1016 |
|
return Compute_Sum(Out, 63); |
1017 |
|
} |
1018 |
|
|
1019 |
|
/* original version including heavy debugging info */ |
1020 |
|
|
1021 |
#ifdef DBGTRELL |
#ifdef DBGTRELL |
1022 |
|
|
1040 |
int j=0, j0=0; |
int j=0, j0=0; |
1041 |
int Run, Level; |
int Run, Level; |
1042 |
|
|
1043 |
Bits = 2; // CBP |
Bits = 2; /* CBP */ |
1044 |
while(j<Last) { |
while(j<Last) { |
1045 |
while(!C[Zigzag[j]]) |
while(!C[Zigzag[j]]) |
1046 |
j++; |
j++; |
1087 |
dct_quantize_trellis_h263_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero) |
dct_quantize_trellis_h263_c(int16_t *const Out, const int16_t *const In, int Q, const uint16_t * const Zigzag, int Non_Zero) |
1088 |
{ |
{ |
1089 |
|
|
1090 |
// Note: We should search last non-zero coeffs on *real* DCT input coeffs (In[]), |
/* |
1091 |
// not quantized one (Out[]). However, it only improves the result *very* |
* Note: We should search last non-zero coeffs on *real* DCT input coeffs (In[]), |
1092 |
// slightly (~0.01dB), whereas speed drops to crawling level :) |
* not quantized one (Out[]). However, it only improves the result *very* |
1093 |
// Well, actually, taking 1 more coeff past Non_Zero into account sometimes helps, |
* slightly (~0.01dB), whereas speed drops to crawling level :) |
1094 |
|
* Well, actually, taking 1 more coeff past Non_Zero into account sometimes helps. |
1095 |
|
*/ |
1096 |
typedef struct { int16_t Run, Level; } NODE; |
typedef struct { int16_t Run, Level; } NODE; |
1097 |
|
|
1098 |
NODE Nodes[65], Last; |
NODE Nodes[65], Last; |
1101 |
const int Mult = 2*Q; |
const int Mult = 2*Q; |
1102 |
const int Bias = (Q-1) | 1; |
const int Bias = (Q-1) | 1; |
1103 |
const int Lev0 = Mult + Bias; |
const int Lev0 = Mult + Bias; |
1104 |
const int Lambda = Trellis_Lambda_Tabs[Q-1]; // it's 1/lambda, actually |
const int Lambda = Trellis_Lambda_Tabs[Q-1]; /* it's 1/lambda, actually */ |
1105 |
|
|
1106 |
int Run_Start = -1; |
int Run_Start = -1; |
1107 |
Run_Costs[-1] = 2<<16; // source (w/ CBP penalty) |
Run_Costs[-1] = 2<<16; /* source (w/ CBP penalty) */ |
1108 |
uint32_t Min_Cost = 2<<16; |
uint32_t Min_Cost = 2<<16; |
1109 |
|
|
1110 |
int Last_Node = -1; |
int Last_Node = -1; |
1113 |
int i, j; |
int i, j; |
1114 |
|
|
1115 |
#if (DBG>0) |
#if (DBG>0) |
1116 |
Last.Level = 0; Last.Run = -1; // just initialize to smthg |
Last.Level = 0; Last.Run = -1; /* just initialize to smthg */ |
1117 |
#endif |
#endif |
1118 |
|
|
1119 |
Non_Zero = Find_Last(Out, Zigzag, Non_Zero); |
Non_Zero = Find_Last(Out, Zigzag, Non_Zero); |
1128 |
uint32_t Best_Cost = 0xf0000000; |
uint32_t Best_Cost = 0xf0000000; |
1129 |
Last_Cost += Dist0; |
Last_Cost += Dist0; |
1130 |
|
|
1131 |
if ((uint32_t)(Level1+1)<3) // very specialized loop for -1,0,+1 |
if ((uint32_t)(Level1+1)<3) /* very specialized loop for -1,0,+1 */ |
1132 |
{ |
{ |
1133 |
int dQ; |
int dQ; |
1134 |
int Run; |
int Run; |
1151 |
const uint32_t Cost = Cost_Base + (Code_Len20[Run-1]<<16); |
const uint32_t Cost = Cost_Base + (Code_Len20[Run-1]<<16); |
1152 |
const uint32_t lCost = Cost_Base + (Code_Len24[Run-1]<<16); |
const uint32_t lCost = Cost_Base + (Code_Len24[Run-1]<<16); |
1153 |
|
|
1154 |
// TODO: what about tie-breaks? Should we favor short runs or |
/* |
1155 |
// long runs? Although the error is the same, it would not be |
* TODO: what about tie-breaks? Should we favor short runs or |
1156 |
// spread the same way along high and low frequencies... |
* long runs? Although the error is the same, it would not be |
1157 |
|
* spread the same way along high and low frequencies... |
1158 |
|
*/ |
1159 |
if (Cost<Best_Cost) { |
if (Cost<Best_Cost) { |
1160 |
Best_Cost = Cost; |
Best_Cost = Cost; |
1161 |
Nodes[i].Run = Run; |
Nodes[i].Run = Run; |
1185 |
printf( "\n" ); |
printf( "\n" ); |
1186 |
} |
} |
1187 |
} |
} |
1188 |
else // "big" levels |
else /* "big" levels */ |
1189 |
{ |
{ |
1190 |
const uint8_t *Tbl_L1, *Tbl_L2, *Tbl_L1_Last, *Tbl_L2_Last; |
const uint8_t *Tbl_L1, *Tbl_L2, *Tbl_L1_Last, *Tbl_L2_Last; |
1191 |
int Level2; |
int Level2; |
1202 |
Tbl_L2 = (Level2<=24) ? B16_17_Code_Len[Level2-1] : Code_Len0; |
Tbl_L2 = (Level2<=24) ? B16_17_Code_Len[Level2-1] : Code_Len0; |
1203 |
Tbl_L1_Last = (Level1<=6) ? B16_17_Code_Len_Last[Level1-1] : Code_Len0; |
Tbl_L1_Last = (Level1<=6) ? B16_17_Code_Len_Last[Level1-1] : Code_Len0; |
1204 |
Tbl_L2_Last = (Level2<=6) ? B16_17_Code_Len_Last[Level2-1] : Code_Len0; |
Tbl_L2_Last = (Level2<=6) ? B16_17_Code_Len_Last[Level2-1] : Code_Len0; |
1205 |
} else { // Level1<-1 |
} else { /* Level1<-1 */ |
1206 |
dQ1 = Level1*Mult-AC - Bias; |
dQ1 = Level1*Mult-AC - Bias; |
1207 |
dQ2 = dQ1 + Mult; |
dQ2 = dQ1 + Mult; |
1208 |
Level2 = Level1 + 1; |
Level2 = Level1 + 1; |
1221 |
uint32_t Cost1, Cost2; |
uint32_t Cost1, Cost2; |
1222 |
int bLevel; |
int bLevel; |
1223 |
|
|
1224 |
// for sub-optimal (but slightly worth it, speed-wise) search, uncomment the following: |
/* |
1225 |
// if (Cost_Base>=Best_Cost) continue; |
* for sub-optimal (but slightly worth it, speed-wise) search, uncomment the following: |
1226 |
|
* if (Cost_Base>=Best_Cost) continue; |
1227 |
|
*/ |
1228 |
Cost1 = Cost_Base + (Tbl_L1[Run-1]<<16); |
Cost1 = Cost_Base + (Tbl_L1[Run-1]<<16); |
1229 |
Cost2 = Cost_Base + (Tbl_L2[Run-1]<<16) + dDist21; |
Cost2 = Cost_Base + (Tbl_L2[Run-1]<<16) + dDist21; |
1230 |
|
|
1255 |
Last.Level = bLevel; |
Last.Level = bLevel; |
1256 |
Last_Node = i; |
Last_Node = i; |
1257 |
} |
} |
1258 |
} //end of "for Run" |
} /* end of "for Run" */ |
1259 |
|
|
1260 |
if (DBG==1) { |
if (DBG==1) { |
1261 |
Run_Costs[i] = Best_Cost; |
Run_Costs[i] = Best_Cost; |
1281 |
} |
} |
1282 |
else |
else |
1283 |
{ |
{ |
1284 |
// as noticed by Michael Niedermayer (michaelni at gmx.at), there's |
/* |
1285 |
// a code shorter by 1 bit for a larger run (!), same level. We give |
* as noticed by Michael Niedermayer (michaelni at gmx.at), there's |
1286 |
// it a chance by not moving the left barrier too much. |
* a code shorter by 1 bit for a larger run (!), same level. We give |
1287 |
|
* it a chance by not moving the left barrier too much. |
1288 |
|
*/ |
1289 |
|
|
1290 |
while( Run_Costs[Run_Start]>Min_Cost+(1<<16) ) |
while( Run_Costs[Run_Start]>Min_Cost+(1<<16) ) |
1291 |
Run_Start++; |
Run_Start++; |
1292 |
|
|
1293 |
// spread on preceding coeffs the cost incurred by skipping this one |
/* spread on preceding coeffs the cost incurred by skipping this one */ |
1294 |
for(j=Run_Start; j<i; ++j) Run_Costs[j] += Dist0; |
for(j=Run_Start; j<i; ++j) Run_Costs[j] += Dist0; |
1295 |
Min_Cost += Dist0; |
Min_Cost += Dist0; |
1296 |
} |
} |
1308 |
if (Last_Node<0) |
if (Last_Node<0) |
1309 |
return -1; |
return -1; |
1310 |
|
|
1311 |
// reconstruct optimal sequence backward with surviving paths |
/* reconstruct optimal sequence backward with surviving paths */ |
1312 |
memset(Out, 0x00, 64*sizeof(*Out)); |
memset(Out, 0x00, 64*sizeof(*Out)); |
1313 |
Out[Zigzag[Last_Node]] = Last.Level; |
Out[Zigzag[Last_Node]] = Last.Level; |
1314 |
i = Last_Node - Last.Run; |
i = Last_Node - Last.Run; |