* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
*****************************************************************************/
-#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <math.h>
******************************* x264 libs **********************************
*
****************************************************************************/
-static int64_t x264_sqe( x264_t *h, uint8_t *pix1, int i_pix_stride, uint8_t *pix2, int i_pix2_stride, int i_width, int i_height )
-{
- int64_t i_sqe = 0;
- int x, y;
-
-#define SSD(size) i_sqe += h->pixf.ssd[size]( pix1+y*i_pix_stride+x, i_pix_stride, \
- pix2+y*i_pix2_stride+x, i_pix2_stride );
- for( y = 0; y < i_height-15; y += 16 )
- {
- for( x = 0; x < i_width-15; x += 16 )
- SSD(PIXEL_16x16);
- if( x < i_width-7 )
- SSD(PIXEL_8x16);
- }
- if( y < i_height-7 )
- for( x = 0; x < i_width-7; x += 8 )
- SSD(PIXEL_8x8);
-#undef SSD
- x264_cpu_restore( h->param.cpu );
-
- return i_sqe;
-}
-
static float x264_psnr( int64_t i_sqe, int64_t i_size )
{
double f_mse = (double)i_sqe / ((double)65025.0 * (double)i_size);
bs_write_ue( s, sh->i_cabac_init_idc );
}
bs_write_se( s, sh->i_qp_delta ); /* slice qp delta */
-#if 0
- if( sh->i_type == SLICE_TYPE_SP || sh->i_type == SLICE_TYPE_SI )
- {
- if( sh->i_type == SLICE_TYPE_SP )
- {
- bs_write1( s, sh->b_sp_for_swidth );
- }
- bs_write_se( s, sh->i_qs_delta );
- }
-#endif
if( sh->pps->b_deblocking_filter_control )
{
return -1;
}
- if( h->param.i_width % 16 != 0 || h->param.i_height % 16 != 0 )
+ if( h->param.i_width % 2 || h->param.i_height % 2 )
{
- x264_log( h, X264_LOG_ERROR, "width %% 16 != 0 or height %% 16 != 0 (%dx%d)\n",
+ x264_log( h, X264_LOG_ERROR, "width or height not divisible by 2 (%dx%d)\n",
h->param.i_width, h->param.i_height );
return -1;
}
}
h->param.i_threads = x264_clip3( h->param.i_threads, 1, X264_SLICE_MAX );
- h->param.i_threads = X264_MIN( h->param.i_threads, h->param.i_height / 16 );
+ h->param.i_threads = X264_MIN( h->param.i_threads, (h->param.i_height + 15) / 16 );
#if !(HAVE_PTHREAD)
if( h->param.i_threads > 1 )
{
}
#endif
+ if( h->param.rc.b_cbr )
+ h->param.rc.i_rf_constant = 0;
+ if( h->param.rc.i_rf_constant > 0 )
+ h->param.rc.i_qp_constant = h->param.rc.i_rf_constant;
+ h->param.rc.i_rf_constant = x264_clip3( h->param.rc.i_rf_constant, 0, 51 );
+ h->param.rc.i_qp_constant = x264_clip3( h->param.rc.i_qp_constant, 0, 51 );
if( !h->param.rc.b_cbr && h->param.rc.i_qp_constant == 0 )
{
h->mb.b_lossless = 1;
h->param.rc.f_ip_factor = 1;
h->param.rc.f_pb_factor = 1;
h->param.analyse.b_psnr = 0;
+ h->param.analyse.i_chroma_qp_offset = 0;
+ h->param.analyse.i_trellis = 0;
+ h->param.analyse.b_fast_pskip = 0;
+ h->param.analyse.i_noise_reduction = 0;
+ }
+
+ if( ( h->param.i_width % 16 || h->param.i_height % 16 ) && !h->mb.b_lossless )
+ {
+ x264_log( h, X264_LOG_WARNING,
+ "width or height not divisible by 16 (%dx%d), compression will suffer.\n",
+ h->param.i_width, h->param.i_height );
}
h->param.i_frame_reference = x264_clip3( h->param.i_frame_reference, 1, 16 );
h->param.i_bframe = x264_clip3( h->param.i_bframe, 0, X264_BFRAME_MAX );
h->param.i_bframe_bias = x264_clip3( h->param.i_bframe_bias, -90, 100 );
h->param.b_bframe_pyramid = h->param.b_bframe_pyramid && h->param.i_bframe > 1;
+ h->param.b_bframe_adaptive = h->param.b_bframe_adaptive && h->param.i_bframe > 0;
h->param.i_deblocking_filter_alphac0 = x264_clip3( h->param.i_deblocking_filter_alphac0, -6, 6 );
h->param.i_deblocking_filter_beta = x264_clip3( h->param.i_deblocking_filter_beta, -6, 6 );
- h->param.i_cabac_init_idc = x264_clip3( h->param.i_cabac_init_idc, -1, 2 );
- /* don't yet support merging of cabac stats */
- if( h->param.i_threads > 1 && h->param.i_cabac_init_idc == -1 )
- h->param.i_cabac_init_idc = 0;
+ h->param.i_cabac_init_idc = x264_clip3( h->param.i_cabac_init_idc, 0, 2 );
if( h->param.i_cqm_preset < X264_CQM_FLAT || h->param.i_cqm_preset > X264_CQM_CUSTOM )
h->param.i_cqm_preset = X264_CQM_FLAT;
if( h->param.analyse.i_me_range > 16 && h->param.analyse.i_me_method <= X264_ME_HEX )
h->param.analyse.i_me_range = 16;
h->param.analyse.i_subpel_refine = x264_clip3( h->param.analyse.i_subpel_refine, 1, 6 );
+ h->param.analyse.b_bframe_rdo = h->param.analyse.b_bframe_rdo && h->param.analyse.i_subpel_refine >= 6;
+ h->param.analyse.b_mixed_references = h->param.analyse.b_mixed_references && h->param.i_frame_reference > 1;
+ h->param.analyse.inter &= X264_ANALYSE_PSUB16x16|X264_ANALYSE_PSUB8x8|X264_ANALYSE_BSUB16x16|
+ X264_ANALYSE_I4x4|X264_ANALYSE_I8x8;
+ h->param.analyse.intra &= X264_ANALYSE_I4x4|X264_ANALYSE_I8x8;
if( !(h->param.analyse.inter & X264_ANALYSE_PSUB16x16) )
h->param.analyse.inter &= ~X264_ANALYSE_PSUB8x8;
if( !h->param.analyse.b_transform_8x8 )
h->param.analyse.intra &= ~X264_ANALYSE_I8x8;
}
h->param.analyse.i_chroma_qp_offset = x264_clip3(h->param.analyse.i_chroma_qp_offset, -12, 12);
- h->param.analyse.i_mv_range = x264_clip3(h->param.analyse.i_mv_range, 32, 2048);
+ if( !h->param.b_cabac )
+ h->param.analyse.i_trellis = 0;
+ h->param.analyse.i_trellis = x264_clip3( h->param.analyse.i_trellis, 0, 2 );
+ h->param.analyse.i_noise_reduction = x264_clip3( h->param.analyse.i_noise_reduction, 0, 1<<16 );
+
+ {
+ const x264_level_t *l = x264_levels;
+ while( l->level_idc != 0 && l->level_idc != h->param.i_level_idc )
+ l++;
+ if( l->level_idc == 0 )
+ {
+ x264_log( h, X264_LOG_ERROR, "invalid level_idc: %d\n", h->param.i_level_idc );
+ return -1;
+ }
+ if( h->param.analyse.i_mv_range <= 0 )
+ h->param.analyse.i_mv_range = l->mv_range;
+ else
+ h->param.analyse.i_mv_range = x264_clip3(h->param.analyse.i_mv_range, 32, 2048);
+ }
if( h->param.rc.f_qblur < 0 )
h->param.rc.f_qblur = 0;
if( h->param.rc.f_complexity_blur < 0 )
h->param.rc.f_complexity_blur = 0;
- h->param.rc.i_qp_constant = x264_clip3(h->param.rc.i_qp_constant, 0, 51);
+
+ /* ensure the booleans are 0 or 1 so they can be used in math */
+#define BOOLIFY(x) h->param.x = !!h->param.x
+ BOOLIFY( b_cabac );
+ BOOLIFY( b_deblocking_filter );
+ BOOLIFY( analyse.b_transform_8x8 );
+ BOOLIFY( analyse.b_weighted_bipred );
+ BOOLIFY( analyse.b_bidir_me );
+ BOOLIFY( analyse.b_chroma_me );
+ BOOLIFY( analyse.b_fast_pskip );
+ BOOLIFY( rc.b_cbr );
+ BOOLIFY( rc.b_stat_write );
+ BOOLIFY( rc.b_stat_read );
+#undef BOOLIFY
return 0;
}
h->pps = &h->pps_array[0];
x264_pps_init( h->pps, 0, &h->param, h->sps);
+ x264_validate_levels( h );
+
x264_cqm_init( h );
h->mb.i_mb_count = h->sps->i_mb_width * h->sps->i_mb_height;
h->frames.i_max_ref0 = h->param.i_frame_reference;
h->frames.i_max_ref1 = h->sps->vui.i_num_reorder_frames;
h->frames.i_max_dpb = h->sps->vui.i_max_dec_frame_buffering + 1;
+ h->frames.b_have_lowres = !h->param.rc.b_stat_read
+ && ( h->param.rc.b_cbr || h->param.rc.i_rf_constant || h->param.b_bframe_adaptive );
for( i = 0; i < X264_BFRAME_MAX + 3; i++ )
{
h->fdec = h->frames.reference[0];
- /* init mb cache */
x264_macroblock_cache_init( h );
-
- /* init cabac adaptive model */
- x264_cabac_model_init( &h->cabac );
+ x264_rdo_init( );
/* init CPU functions */
x264_predict_16x16_init( h->param.cpu, h->predict_16x16 );
x264_mc_init( h->param.cpu, &h->mc );
x264_csp_init( h->param.cpu, h->param.i_csp, &h->csp );
x264_quant_init( h, h->param.cpu, &h->quantf );
+ x264_deblock_init( h->param.cpu, &h->loopf );
memcpy( h->pixf.mbcmp,
( h->mb.b_lossless || h->param.analyse.i_subpel_refine <= 1 ) ? h->pixf.sad : h->pixf.satd,
h->param.analyse.i_me_method = param->analyse.i_me_method;
h->param.analyse.i_me_range = param->analyse.i_me_range;
h->param.analyse.i_subpel_refine = param->analyse.i_subpel_refine;
+ h->param.analyse.i_trellis = param->analyse.i_trellis;
h->param.analyse.intra = param->analyse.intra;
h->param.analyse.inter = param->analyse.inter;
- if( h->sps->b_direct8x8_inference && h->param.i_bframe
- && h->param.analyse.i_direct_mv_pred == X264_DIRECT_PRED_TEMPORAL )
- h->param.analyse.inter &= ~X264_ANALYSE_PSUB8x8;
memcpy( h->pixf.mbcmp,
( h->mb.b_lossless || h->param.analyse.i_subpel_refine <= 1 ) ? h->pixf.sad : h->pixf.satd,
{
/* identify ourself */
x264_nal_start( h, NAL_SEI, NAL_PRIORITY_DISPOSABLE );
- x264_sei_version_write( &h->out.bs );
+ x264_sei_version_write( h, &h->out.bs );
x264_nal_end( h );
/* generate sequence parameters */
: dtime > 0;
if( swap )
{
- x264_frame_t *tmp = list[i+1];
- list[i+1] = list[i];
- list[i] = tmp;
+ XCHG( x264_frame_t*, list[i], list[i+1] );
b_ok = 0;
}
}
{
if( h->fref0[i]->i_poc < h->fref0[i+1]->i_poc )
{
- x264_frame_t *tmp = h->fref0[i+1];
-
- h->fref0[i+1] = h->fref0[i];
- h->fref0[i] = tmp;
+ XCHG( x264_frame_t*, h->fref0[i], h->fref0[i+1] );
b_ok = 0;
break;
}
{
if( h->fref1[i]->i_poc > h->fref1[i+1]->i_poc )
{
- x264_frame_t *tmp = h->fref1[i+1];
-
- h->fref1[i+1] = h->fref1[i];
- h->fref1[i] = tmp;
+ XCHG( x264_frame_t*, h->fref1[i], h->fref1[i+1] );
b_ok = 0;
break;
}
/* move lowres copy of the image to the ref frame */
for( i = 0; i < 4; i++)
- {
- uint8_t *tmp = h->fdec->lowres[i];
- h->fdec->lowres[i] = h->fenc->lowres[i];
- h->fenc->lowres[i] = tmp;
- }
+ XCHG( uint8_t*, h->fdec->lowres[i], h->fenc->lowres[i] );
/* adaptive B decision needs a pointer, since it can't use the ref lists */
if( h->sh.i_type != SLICE_TYPE_B )
/* Nothing to do ? */
}
- /* get adapative cabac model if needed */
- if( h->param.b_cabac )
- {
- if( h->param.i_cabac_init_idc == -1 )
- {
- h->sh.i_cabac_init_idc = x264_cabac_model_get( &h->cabac, i_slice_type );
- }
- }
-
x264_macroblock_slice_init( h );
}
if( h->param.b_cabac )
{
- int i_cabac_word;
x264_cabac_encode_flush( &h->cabac );
- /* TODO cabac stuffing things (p209) */
- i_cabac_word = (((3 * h->cabac.i_sym_cnt - 3 * 96 * h->sps->i_mb_width * h->sps->i_mb_height)/32) - bs_pos( &h->out.bs)/8)/3;
- while( i_cabac_word > 0 )
- {
- bs_write( &h->out.bs, 16, 0x0000 );
- i_cabac_word--;
- }
}
else
{
x264_frame_copy_picture( h, fenc, pic_in );
+ if( h->param.i_width % 16 || h->param.i_height % 16 )
+ x264_frame_expand_border_mod16( h, fenc );
+
fenc->i_frame = h->frames.i_input++;
x264_frame_put( h->frames.next, fenc );
- x264_frame_init_lowres( h->param.cpu, fenc );
+ if( h->frames.b_have_lowres )
+ x264_frame_init_lowres( h->param.cpu, fenc );
if( h->frames.i_input <= h->frames.i_delay )
{
h->i_nal_ref_idc = i_nal_ref_idc;
/* Write SPS and PPS */
- if( i_nal_type == NAL_SLICE_IDR )
+ if( i_nal_type == NAL_SLICE_IDR && h->param.b_repeat_headers )
{
if( h->fenc->i_frame == 0 )
{
/* identify ourself */
x264_nal_start( h, NAL_SEI, NAL_PRIORITY_DISPOSABLE );
- x264_sei_version_write( &h->out.bs );
+ x264_sei_version_write( h, &h->out.bs );
x264_nal_end( h );
}
}
/* ---------------------- Update encoder state ------------------------- */
- /* update cabac */
- if( h->param.b_cabac && h->param.i_cabac_init_idc == -1 )
- {
- x264_cabac_model_update( &h->cabac, i_slice_type, h->sh.i_qp );
- }
/* handle references */
if( i_nal_ref_idc != NAL_PRIORITY_DISPOSABLE )
x264_frame_put( h->frames.unused, h->fenc );
+ x264_noise_reduction_update( h );
+
TIMER_STOP( i_mtime_encode_frame );
/* ---------------------- Compute/Print statistics --------------------- */
int64_t i_sqe_y, i_sqe_u, i_sqe_v;
/* PSNR */
- i_sqe_y = x264_sqe( h, frame_psnr->plane[0], frame_psnr->i_stride[0], h->fenc->plane[0], h->fenc->i_stride[0], h->param.i_width, h->param.i_height );
- i_sqe_u = x264_sqe( h, frame_psnr->plane[1], frame_psnr->i_stride[1], h->fenc->plane[1], h->fenc->i_stride[1], h->param.i_width/2, h->param.i_height/2);
- i_sqe_v = x264_sqe( h, frame_psnr->plane[2], frame_psnr->i_stride[2], h->fenc->plane[2], h->fenc->i_stride[2], h->param.i_width/2, h->param.i_height/2);
+ i_sqe_y = x264_pixel_ssd_wxh( &h->pixf, frame_psnr->plane[0], frame_psnr->i_stride[0], h->fenc->plane[0], h->fenc->i_stride[0], h->param.i_width, h->param.i_height );
+ i_sqe_u = x264_pixel_ssd_wxh( &h->pixf, frame_psnr->plane[1], frame_psnr->i_stride[1], h->fenc->plane[1], h->fenc->i_stride[1], h->param.i_width/2, h->param.i_height/2);
+ i_sqe_v = x264_pixel_ssd_wxh( &h->pixf, frame_psnr->plane[2], frame_psnr->i_stride[2], h->fenc->plane[2], h->fenc->i_stride[2], h->param.i_width/2, h->param.i_height/2);
+ x264_cpu_restore( h->param.cpu );
h->stat.i_sqe_global[i_slice_type] += i_sqe_y + i_sqe_u + i_sqe_v;
h->stat.f_psnr_average[i_slice_type] += x264_psnr( i_sqe_y + i_sqe_u + i_sqe_v, 3 * h->param.i_width * h->param.i_height / 2 );
#ifdef DEBUG_DUMP_FRAME
/* Dump reconstructed frame */
x264_frame_dump( h, frame_psnr, "fdec.yuv" );
-#endif
-#if 0
- if( h->i_ref0 > 0 )
- {
- x264_frame_dump( h, h->fref0[0], "ref0.yuv" );
- }
- if( h->i_ref1 > 0 )
- {
- x264_frame_dump( h, h->fref1[0], "ref1.yuv" );
- }
#endif
return 0;
}
i_mb_count[B_SKIP] / i_count );
}
+ x264_ratecontrol_summary( h );
+
if( h->stat.i_slice_count[SLICE_TYPE_I] + h->stat.i_slice_count[SLICE_TYPE_P] + h->stat.i_slice_count[SLICE_TYPE_B] > 0 )
{
const int i_count = h->stat.i_slice_count[SLICE_TYPE_I] +
projects / x264 / blobdiff