*
* Authors: Laurent Aimar <fenrir@via.ecp.fr>
* Loren Merritt <lorenm@u.washington.edu>
+ * Fiona Glaser <fiona@x264.com>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
h->param.rc.i_rc_method = X264_RC_CQP;
h->param.rc.f_ip_factor = 1;
h->param.rc.f_pb_factor = 1;
- h->param.analyse.b_transform_8x8 = 0;
h->param.analyse.b_psnr = 0;
h->param.analyse.b_ssim = 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;
+ h->param.analyse.f_psy_rd = 0;
+ h->param.i_bframe = 0;
+ /* 8x8dct is not useful at all in CAVLC lossless */
+ if( !h->param.b_cabac )
+ h->param.analyse.b_transform_8x8 = 0;
}
if( h->param.rc.i_rc_method == X264_RC_CQP )
{
if( h->param.i_keyint_max <= 0 )
h->param.i_keyint_max = 1;
h->param.i_keyint_min = x264_clip3( h->param.i_keyint_min, 1, h->param.i_keyint_max/2+1 );
-
+ if( !h->param.analyse.i_subpel_refine && h->param.analyse.i_direct_mv_pred > X264_DIRECT_PRED_SPATIAL )
+ {
+ x264_log( h, X264_LOG_WARNING, "subme=0 + direct=temporal is not supported\n" );
+ h->param.analyse.i_direct_mv_pred = X264_DIRECT_PRED_SPATIAL;
+ }
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;
+ if( !h->param.i_bframe )
+ h->param.i_bframe_adaptive = X264_B_ADAPT_NONE;
h->param.analyse.b_weighted_bipred = h->param.analyse.b_weighted_bipred && h->param.i_bframe > 0;
h->mb.b_direct_auto_write = h->param.analyse.i_direct_mv_pred == X264_DIRECT_PRED_AUTO
&& h->param.i_bframe
if( h->param.analyse.i_me_method == X264_ME_TESA &&
(h->mb.b_lossless || h->param.analyse.i_subpel_refine <= 1) )
h->param.analyse.i_me_method = X264_ME_ESA;
- h->param.analyse.i_subpel_refine = x264_clip3( h->param.analyse.i_subpel_refine, 1, 7 );
- h->param.analyse.b_bframe_rdo = h->param.analyse.b_bframe_rdo && h->param.analyse.i_subpel_refine >= 6;
+ h->param.analyse.i_subpel_refine = x264_clip3( h->param.analyse.i_subpel_refine, 0, 9 );
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;
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.rc.i_aq_mode = x264_clip3( h->param.rc.i_aq_mode, 0, 2 );
- if( h->param.rc.f_aq_strength <= 0 )
+ if( !h->param.analyse.i_trellis )
+ h->param.analyse.f_psy_trellis = 0;
+ h->param.analyse.f_psy_rd = x264_clip3f( h->param.analyse.f_psy_rd, 0, 10 );
+ h->param.analyse.f_psy_trellis = x264_clip3f( h->param.analyse.f_psy_trellis, 0, 10 );
+ if( h->param.analyse.i_subpel_refine < 6 )
+ h->param.analyse.f_psy_rd = 0;
+ h->mb.i_psy_rd = FIX8( h->param.analyse.f_psy_rd );
+ /* Psy RDO increases overall quantizers to improve the quality of luma--this indirectly hurts chroma quality */
+ /* so we lower the chroma QP offset to compensate */
+ /* This can be triggered repeatedly on multiple calls to parameter_validate, but since encoding
+ * uses the pps chroma qp offset not the param chroma qp offset, this is not a problem. */
+ if( h->mb.i_psy_rd )
+ h->param.analyse.i_chroma_qp_offset -= h->param.analyse.f_psy_rd < 0.25 ? 1 : 2;
+ h->mb.i_psy_trellis = FIX8( h->param.analyse.f_psy_trellis / 4 );
+ /* Psy trellis has a similar effect. */
+ if( h->mb.i_psy_trellis )
+ h->param.analyse.i_chroma_qp_offset -= h->param.analyse.f_psy_trellis < 0.25 ? 1 : 2;
+ else
+ h->mb.i_psy_trellis = 0;
+ h->param.analyse.i_chroma_qp_offset = x264_clip3(h->param.analyse.i_chroma_qp_offset, -12, 12);
+ h->param.rc.i_aq_mode = x264_clip3( h->param.rc.i_aq_mode, 0, 1 );
+ h->param.rc.f_aq_strength = x264_clip3f( h->param.rc.f_aq_strength, 0, 3 );
+ if( h->param.rc.f_aq_strength == 0 )
h->param.rc.i_aq_mode = 0;
- /* VAQ effectively replaces qcomp, so qcomp is raised towards 1 to compensate. */
- if( h->param.rc.i_aq_mode == X264_AQ_GLOBAL )
- h->param.rc.f_qcompress = x264_clip3f(h->param.rc.f_qcompress + h->param.rc.f_aq_strength / 0.7, 0, 1);
h->param.analyse.i_noise_reduction = x264_clip3( h->param.analyse.i_noise_reduction, 0, 1<<16 );
{
while( l[1].level_idc && x264_validate_levels( h, 0 ) && l++ );
if( h->param.rc.i_vbv_buffer_size <= 0 )
h->param.rc.i_vbv_max_bitrate = 0;
- x264_log( h, X264_LOG_DEBUG, "level_idc: %d\n", h->param.i_level_idc );
}
else
{
h->param.analyse.i_mv_range = l->mv_range >> h->param.b_interlaced;
else
h->param.analyse.i_mv_range = x264_clip3(h->param.analyse.i_mv_range, 32, 512 >> h->param.b_interlaced);
- if( h->param.analyse.i_direct_8x8_inference < 0 )
- h->param.analyse.i_direct_8x8_inference = l->direct8x8;
}
if( h->param.i_threads > 1 )
BOOLIFY( b_deblocking_filter );
BOOLIFY( b_interlaced );
BOOLIFY( analyse.b_transform_8x8 );
- BOOLIFY( analyse.i_direct_8x8_inference );
- BOOLIFY( analyse.b_bidir_me );
BOOLIFY( analyse.b_chroma_me );
BOOLIFY( analyse.b_fast_pskip );
BOOLIFY( rc.b_stat_write );
static void mbcmp_init( x264_t *h )
{
int satd = !h->mb.b_lossless && h->param.analyse.i_subpel_refine > 1;
- memcpy( h->pixf.mbcmp, satd ? h->pixf.satd : h->pixf.sad, sizeof(h->pixf.mbcmp) );
+ memcpy( h->pixf.mbcmp, satd ? h->pixf.satd : h->pixf.sad_aligned, sizeof(h->pixf.mbcmp) );
+ memcpy( h->pixf.mbcmp_unaligned, satd ? h->pixf.satd : h->pixf.sad, sizeof(h->pixf.mbcmp_unaligned) );
+ h->pixf.intra_mbcmp_x3_16x16 = satd ? h->pixf.intra_satd_x3_16x16 : h->pixf.intra_sad_x3_16x16;
satd &= h->param.analyse.i_me_method == X264_ME_TESA;
memcpy( h->pixf.fpelcmp, satd ? h->pixf.satd : h->pixf.sad, sizeof(h->pixf.fpelcmp) );
memcpy( h->pixf.fpelcmp_x3, satd ? h->pixf.satd_x3 : h->pixf.sad_x3, sizeof(h->pixf.fpelcmp_x3) );
h->mb.i_mb_count = h->sps->i_mb_width * h->sps->i_mb_height;
/* Init frames. */
- h->frames.i_delay = h->param.i_bframe + h->param.i_threads - 1;
+ if( h->param.i_bframe_adaptive == X264_B_ADAPT_TRELLIS )
+ h->frames.i_delay = X264_MAX(h->param.i_bframe,3)*4 + h->param.i_threads - 1;
+ else
+ h->frames.i_delay = h->param.i_bframe + h->param.i_threads - 1;
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;
h->frames.b_have_lowres = !h->param.rc.b_stat_read
&& ( h->param.rc.i_rc_method == X264_RC_ABR
|| h->param.rc.i_rc_method == X264_RC_CRF
- || h->param.b_bframe_adaptive
+ || h->param.i_bframe_adaptive
|| h->param.b_pre_scenecut );
h->frames.b_have_lowres |= (h->param.rc.b_stat_read && h->param.rc.i_vbv_buffer_size > 0);
+ h->frames.b_have_sub8x8_esa = !!(h->param.analyse.inter & X264_ANALYSE_PSUB8x8);
h->frames.i_last_idr = - h->param.i_keyint_max;
h->frames.i_input = 0;
/* init CPU functions */
x264_predict_16x16_init( h->param.cpu, h->predict_16x16 );
x264_predict_8x8c_init( h->param.cpu, h->predict_8x8c );
- x264_predict_8x8_init( h->param.cpu, h->predict_8x8 );
+ x264_predict_8x8_init( h->param.cpu, h->predict_8x8, &h->predict_8x8_filter );
x264_predict_4x4_init( h->param.cpu, h->predict_4x4 );
-
+ if( !h->param.b_cabac );
+ x264_init_vlc_tables();
x264_pixel_init( h->param.cpu, &h->pixf );
x264_dct_init( h->param.cpu, &h->dctf );
x264_zigzag_init( h->param.cpu, &h->zigzagf, h->param.b_interlaced );
if( !strcmp(x264_cpu_names[i].name, "SSE3")
&& (param->cpu & X264_CPU_SSSE3 || !(param->cpu & X264_CPU_CACHELINE_64)) )
continue;
+ if( !strcmp(x264_cpu_names[i].name, "SSE4.1")
+ && (param->cpu & X264_CPU_SSE42) )
+ continue;
if( (param->cpu & x264_cpu_names[i].flags) == x264_cpu_names[i].flags
&& (!i || x264_cpu_names[i].flags != x264_cpu_names[i-1].flags) )
p += sprintf( p, " %s", x264_cpu_names[i].name );
}
}
+ x264_log( h, X264_LOG_INFO, "profile %s, level %d.%d\n",
+ h->sps->i_profile_idc == PROFILE_BASELINE ? "Baseline" :
+ h->sps->i_profile_idc == PROFILE_MAIN ? "Main" :
+ h->sps->i_profile_idc == PROFILE_HIGH ? "High" :
+ "High 4:4:4 Predictive", h->sps->i_level_idc/10, h->sps->i_level_idc%10 );
+
return h;
}
COPY( analyse.intra );
COPY( analyse.inter );
COPY( analyse.i_direct_mv_pred );
- COPY( analyse.i_me_method );
- COPY( analyse.i_me_range );
+ /* Scratch buffer prevents me_range from being increased for esa/tesa */
+ if( h->param.analyse.i_me_method < X264_ME_ESA || param->analyse.i_me_range < h->param.analyse.i_me_range )
+ COPY( analyse.i_me_range );
COPY( analyse.i_noise_reduction );
- COPY( analyse.i_subpel_refine );
+ /* We can't switch out of subme=0 during encoding. */
+ if( h->param.analyse.i_subpel_refine )
+ COPY( analyse.i_subpel_refine );
COPY( analyse.i_trellis );
- COPY( analyse.b_bidir_me );
- COPY( analyse.b_bframe_rdo );
COPY( analyse.b_chroma_me );
COPY( analyse.b_dct_decimate );
COPY( analyse.b_fast_pskip );
COPY( analyse.b_mixed_references );
+ COPY( analyse.f_psy_rd );
+ COPY( analyse.f_psy_trellis );
// can only twiddle these if they were enabled to begin with:
+ if( h->param.analyse.i_me_method >= X264_ME_ESA || param->analyse.i_me_method < X264_ME_ESA )
+ COPY( analyse.i_me_method );
+ if( h->param.analyse.i_me_method >= X264_ME_ESA && !h->frames.b_have_sub8x8_esa )
+ h->param.analyse.inter &= ~X264_ANALYSE_PSUB8x8;
if( h->pps->b_transform_8x8_mode )
COPY( analyse.b_transform_8x8 );
if( h->frames.i_max_ref1 > 1 )
if( b_hpel )
{
x264_frame_expand_border( h, h->fdec, min_y, b_end );
- x264_frame_filter( h, h->fdec, min_y, b_end );
- x264_frame_expand_border_filtered( h, h->fdec, min_y, b_end );
+ if( h->param.analyse.i_subpel_refine )
+ {
+ x264_frame_filter( h, h->fdec, min_y, b_end );
+ x264_frame_expand_border_filtered( h, h->fdec, min_y, b_end );
+ }
}
if( h->param.i_threads > 1 && h->fdec->b_kept_as_ref )
x264_pixel_ssim_wxh( &h->pixf,
h->fdec->plane[0] + 2+min_y*h->fdec->i_stride[0], h->fdec->i_stride[0],
h->fenc->plane[0] + 2+min_y*h->fenc->i_stride[0], h->fenc->i_stride[0],
- h->param.i_width-2, max_y-min_y );
+ h->param.i_width-2, max_y-min_y, h->scratch_buffer );
}
}
/* Compute misc bits */
h->stat.frame.i_misc_bits = bs_pos( &h->out.bs )
+ NALU_OVERHEAD * 8
- - h->stat.frame.i_itex_bits
- - h->stat.frame.i_ptex_bits
- - h->stat.frame.i_hdr_bits;
+ - h->stat.frame.i_tex_bits
+ - h->stat.frame.i_mv_bits;
}
static void x264_thread_sync_context( x264_t *dst, x264_t *src )
{
int i_frame_size;
+#ifdef HAVE_MMX
+ /* Misalign mask has to be set separately for each thread. */
+ if( h->param.cpu&X264_CPU_SSE_MISALIGN )
+ x264_cpu_mask_misalign_sse();
+#endif
+
#if VISUALIZE
if( h->param.b_visualize )
x264_visualize_init( h );
if( h->frames.b_have_lowres )
x264_frame_init_lowres( h, fenc );
+ if( h->param.rc.i_aq_mode )
+ x264_adaptive_quant_frame( h, fenc );
+
if( h->frames.i_input <= h->frames.i_delay + 1 - h->param.i_threads )
{
/* Nothing yet to encode */
return 0;
}
- x264_slicetype_decide( h );
+ x264_stack_align( x264_slicetype_decide, h );
/* 3: move some B-frames and 1 non-B to encode queue */
while( IS_X264_TYPE_B( h->frames.next[bframes]->i_type ) )
* we can't assign an I-frame. Instead, change the previous
* B-frame to P, and rearrange coding order. */
- if( h->param.b_bframe_adaptive || b > 1 )
+ if( h->param.i_bframe_adaptive || b > 1 )
h->fenc->i_type = X264_TYPE_AUTO;
x264_frame_sort_pts( h->frames.current );
x264_frame_unshift( h->frames.next, h->fenc );
for( i = 0; i < X264_PARTTYPE_MAX; i++ )
for( j = 0; j < 2; j++ )
{
- int l0 = x264_mb_type_list0_table[i][j];
- int l1 = x264_mb_type_list1_table[i][j];
+ int l0 = x264_mb_type_list_table[i][0][j];
+ int l1 = x264_mb_type_list_table[i][1][j];
if( l0 || l1 )
list_count[l1+l0*l1] += h->stat.i_mb_count[SLICE_TYPE_B][i] * 2;
}