//{ mb_analyse_int //{ mb_analyse_int //{ mb_analyse_int //{ mb_analyse_int
static void dull_mb_analyse_init_P( x264_t *h, x264_mb_analysis_t *a )
{
x264_mb_analyse_init_qp( h, a, h->mb.i_qp );
h->mb.b_transform_8x8 = 0;
h->mb.b_noise_reduction = 0;
/* I: Intra part */
a->i_satd_i16x16 =
a->i_satd_i8x8 =
a->i_satd_i4x4 =
a->i_satd_i8x8chroma = COST_MAX;
a->b_fast_intra = 0;
a->b_avoid_topright = 0;
h->mb.i_skip_intra =
h->mb.b_lossless ? 0 :
a->i_mbrd ? 2 :
!h->param.analyse.i_trellis && !h->param.analyse.i_noise_reduction;
/* II: Inter part P/B frame */
if( h->sh.i_type != SLICE_TYPE_I )
{
int i;
int i_fmv_range = 4 * h->param.analyse.i_mv_range;
// limit motion search to a slightly smaller range than the theoretical limit,
// since the search may go a few iterations past its given range
int i_fpel_border = 6; // umh: 1 for diamond, 2 for octagon, 2 for hpel
/* Calculate max allowed MV range */
#define CLIP_FMV(mv) x264_clip3( mv, -i_fmv_range, i_fmv_range-1 )
h->mb.mv_min[0] = 4*( -16*h->mb.i_mb_x - 24 );
h->mb.mv_max[0] = 4*( 16*( h->sps->i_mb_width - h->mb.i_mb_x - 1 ) + 24 );
h->mb.mv_min_spel[0] = CLIP_FMV( h->mb.mv_min[0] );
h->mb.mv_max_spel[0] = CLIP_FMV( h->mb.mv_max[0] );
if( h->param.b_intra_refresh && h->sh.i_type == SLICE_TYPE_P )
{
int max_x = (h->fref0[0]->i_pir_end_col * 16 - 3)*4; /* 3 pixels of hpel border */
int max_mv = max_x - 4*16*h->mb.i_mb_x;
/* If we're left of the refresh bar, don't reference right of it. */
if( max_mv > 0 && h->mb.i_mb_x < h->fdec->i_pir_start_col )
h->mb.mv_max_spel[0] = X264_MIN( h->mb.mv_max_spel[0], max_mv );
}
h->mb.mv_min_fpel[0] = (h->mb.mv_min_spel[0]>>2) + i_fpel_border;
h->mb.mv_max_fpel[0] = (h->mb.mv_max_spel[0]>>2) - i_fpel_border;
if( h->mb.i_mb_x == 0 )
{
int mb_y = h->mb.i_mb_y;
int mb_height = h->sps->i_mb_height;
int thread_mvy_range = i_fmv_range;
h->mb.mv_min[1] = 4*( -16*mb_y - 24 );
h->mb.mv_max[1] = 4*( 16*( mb_height - mb_y - 1 ) + 24 );
h->mb.mv_min_spel[1] = x264_clip3( h->mb.mv_min[1], -i_fmv_range, i_fmv_range );
h->mb.mv_max_spel[1] = CLIP_FMV( h->mb.mv_max[1] );
h->mb.mv_max_spel[1] = X264_MIN( h->mb.mv_max_spel[1], thread_mvy_range*4 );
h->mb.mv_min_fpel[1] = (h->mb.mv_min_spel[1]>>2) + i_fpel_border;
h->mb.mv_max_fpel[1] = (h->mb.mv_max_spel[1]>>2) - i_fpel_border;
}
float x264_pixel_ssim_wxh( x264_pixel_function_t *pf,
uint8_t *pix1, int stride1,
uint8_t *pix2, int stride2,
int width, int height )
{
int x, y, z;
float ssim = 0.0;
int (*sum0)[4] = x264_malloc(4 * (width/4+3) * sizeof(int));
int (*sum1)[4] = x264_malloc(4 * (width/4+3) * sizeof(int));
width >>= 2;
height >>= 2;
z = 0;
for( y = 1; y < height; y++ )
{
for( ; z <= y; z++ )
{
XCHG( void*, sum0, sum1 );
for( x = 0; x < width; x+=2 )
pf->ssim_4x4x2_core( &pix1[4*(x+z*stride1)], stride1, &pix2[4*(x+z*stride2)], stride2, &sum0[x] );
}
for( x = 0; x < width-1; x += 4 )
ssim += pf->ssim_end4( sum0+x, sum1+x, X264_MIN(4,width-x-1) );
}
x264_free(sum0);
x264_free(sum1);
return ssim / ((width-1) * (height-1));
}
static void fill_cache( cache_hnd_t *h, int frame )
{
/* shift frames out of the cache as the frame request is beyond the filled cache */
int shift = frame - LAST_FRAME;
/* no frames to shift or no frames left to read */
if( shift <= 0 || h->eof )
return;
/* the next frames to read are either
* A) starting at the end of the current cache, or
* B) starting at a new frame that has the end of the cache at the desired frame
* and proceeding to fill the entire cache */
int cur_frame = X264_MAX( h->first_frame + h->cur_size, frame - h->max_size + 1 );
/* the new starting point is either
* A) the current one shifted the number of frames entering/leaving the cache, or
* B) at a new frame that has the end of the cache at the desired frame. */
h->first_frame = X264_MIN( h->first_frame + shift, cur_frame );
h->cur_size = X264_MAX( h->cur_size - shift, 0 );
while( h->cur_size < h->max_size )
{
cli_pic_t temp;
/* the old front frame is going to shift off, overwrite it with the new frame */
cli_pic_t *cache = h->cache[0];
if( h->prev_filter.get_frame( h->prev_hnd, &temp, cur_frame ) ||
x264_cli_pic_copy( cache, &temp ) ||
h->prev_filter.release_frame( h->prev_hnd, &temp, cur_frame ) )
{
h->eof = cur_frame;
return;
}
/* the read was successful, shift the frame off the front to the end */
x264_frame_push( (void*)h->cache, x264_frame_shift( (void*)h->cache ) );
cur_frame++;
h->cur_size++;
}
}
static void *x264_lookahead_thread( x264_t *h )
{
while( !h->lookahead->b_exit_thread )
{
x264_pthread_mutex_lock( &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_lock( &h->lookahead->next.mutex );
int shift = X264_MIN( h->lookahead->next.i_max_size - h->lookahead->next.i_size, h->lookahead->ifbuf.i_size );
x264_lookahead_shift( &h->lookahead->next, &h->lookahead->ifbuf, shift );
x264_pthread_mutex_unlock( &h->lookahead->next.mutex );
if( h->lookahead->next.i_size <= h->lookahead->i_slicetype_length + h->param.b_vfr_input )
{
while( !h->lookahead->ifbuf.i_size && !h->lookahead->b_exit_thread )
x264_pthread_cond_wait( &h->lookahead->ifbuf.cv_fill, &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
}
else
{
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
x264_lookahead_slicetype_decide( h );
}
} /* end of input frames */
x264_pthread_mutex_lock( &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_lock( &h->lookahead->next.mutex );
x264_lookahead_shift( &h->lookahead->next, &h->lookahead->ifbuf, h->lookahead->ifbuf.i_size );
x264_pthread_mutex_unlock( &h->lookahead->next.mutex );
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
while( h->lookahead->next.i_size )
x264_lookahead_slicetype_decide( h );
x264_pthread_mutex_lock( &h->lookahead->ofbuf.mutex );
h->lookahead->b_thread_active = 0;
x264_pthread_cond_broadcast( &h->lookahead->ofbuf.cv_fill );
x264_pthread_mutex_unlock( &h->lookahead->ofbuf.mutex );
return NULL;
}
int x264_threadpool_init( x264_threadpool_t **p_pool, int threads,
void (*init_func)(void *), void *init_arg )
{
if( threads <= 0 )
return -1;
x264_threadpool_t *pool;
CHECKED_MALLOCZERO( pool, sizeof(x264_threadpool_t) );
*p_pool = pool;
pool->init_func = init_func;
pool->init_arg = init_arg;
pool->threads = X264_MIN( threads, X264_THREAD_MAX );
CHECKED_MALLOC( pool->thread_handle, pool->threads * sizeof(x264_pthread_t) );
if( x264_sync_frame_list_init( &pool->uninit, pool->threads ) ||
x264_sync_frame_list_init( &pool->run, pool->threads ) ||
x264_sync_frame_list_init( &pool->done, pool->threads ) )
goto fail;
for( int i = 0; i < pool->threads; i++ )
{
x264_threadpool_job_t *job;
CHECKED_MALLOC( job, sizeof(x264_threadpool_job_t) );
x264_sync_frame_list_push( &pool->uninit, (void*)job );
}
for( int i = 0; i < pool->threads; i++ )
if( x264_pthread_create( pool->thread_handle+i, NULL, (void*)x264_threadpool_thread, pool ) )
goto fail;
return 0;
fail:
return -1;
}
static void x264_lookahead_thread( x264_t *h )
{
int shift;
#ifdef HAVE_MMX
if( h->param.cpu&X264_CPU_SSE_MISALIGN )
x264_cpu_mask_misalign_sse();
#endif
while( !h->lookahead->b_exit_thread )
{
x264_pthread_mutex_lock( &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_lock( &h->lookahead->next.mutex );
shift = X264_MIN( h->lookahead->next.i_max_size - h->lookahead->next.i_size, h->lookahead->ifbuf.i_size );
x264_lookahead_shift( &h->lookahead->next, &h->lookahead->ifbuf, shift );
x264_pthread_mutex_unlock( &h->lookahead->next.mutex );
if( h->lookahead->next.i_size <= h->lookahead->i_slicetype_length )
{
while( !h->lookahead->ifbuf.i_size && !h->lookahead->b_exit_thread )
x264_pthread_cond_wait( &h->lookahead->ifbuf.cv_fill, &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
}
else
{
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
x264_lookahead_slicetype_decide( h );
}
} /* end of input frames */
#ifdef ERR1
x264_pthread_mutex_lock( &h->lookahead->next.mutex );
x264_pthread_mutex_lock( &h->lookahead->ifbuf.mutex );
#else
x264_pthread_mutex_lock( &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_lock( &h->lookahead->next.mutex );
#endif
x264_lookahead_shift( &h->lookahead->next, &h->lookahead->ifbuf, h->lookahead->ifbuf.i_size );
#ifdef ERR1
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
x264_pthread_mutex_unlock( &h->lookahead->next.mutex );
#else
x264_pthread_mutex_unlock( &h->lookahead->next.mutex );
x264_pthread_mutex_unlock( &h->lookahead->ifbuf.mutex );
#endif
while( h->lookahead->next.i_size )
x264_lookahead_slicetype_decide( h );
x264_pthread_mutex_lock( &h->lookahead->ofbuf.mutex );
h->lookahead->b_thread_active = 0;
x264_pthread_cond_broadcast( &h->lookahead->ofbuf.cv_fill );
x264_pthread_mutex_unlock( &h->lookahead->ofbuf.mutex );
}
void x264_sps_init( x264_sps_t *sps, int i_id, x264_param_t *param )
{
int csp = param->i_csp & X264_CSP_MASK;
sps->i_id = i_id;
sps->i_mb_width = ( param->i_width + 15 ) / 16;
sps->i_mb_height= ( param->i_height + 15 ) / 16;
sps->i_chroma_format_idc = csp >= X264_CSP_I444 ? CHROMA_444 :
csp >= X264_CSP_I422 ? CHROMA_422 : CHROMA_420;
sps->b_qpprime_y_zero_transform_bypass = param->rc.i_rc_method == X264_RC_CQP && param->rc.i_qp_constant == 0;
if( sps->b_qpprime_y_zero_transform_bypass || sps->i_chroma_format_idc == CHROMA_444 )
sps->i_profile_idc = PROFILE_HIGH444_PREDICTIVE;
else if( sps->i_chroma_format_idc == CHROMA_422 )
sps->i_profile_idc = PROFILE_HIGH422;
else if( BIT_DEPTH > 8 )
sps->i_profile_idc = PROFILE_HIGH10;
else if( param->analyse.b_transform_8x8 || param->i_cqm_preset != X264_CQM_FLAT )
sps->i_profile_idc = PROFILE_HIGH;
else if( param->b_cabac || param->i_bframe > 0 || param->b_interlaced || param->b_fake_interlaced || param->analyse.i_weighted_pred > 0 )
sps->i_profile_idc = PROFILE_MAIN;
else
sps->i_profile_idc = PROFILE_BASELINE;
sps->b_constraint_set0 = sps->i_profile_idc == PROFILE_BASELINE;
/* x264 doesn't support the features that are in Baseline and not in Main,
* namely arbitrary_slice_order and slice_groups. */
sps->b_constraint_set1 = sps->i_profile_idc <= PROFILE_MAIN;
/* Never set constraint_set2, it is not necessary and not used in real world. */
sps->b_constraint_set2 = 0;
sps->b_constraint_set3 = 0;
sps->i_level_idc = param->i_level_idc;
if( param->i_level_idc == 9 && ( sps->i_profile_idc == PROFILE_BASELINE || sps->i_profile_idc == PROFILE_MAIN ) )
{
sps->b_constraint_set3 = 1; /* level 1b with Baseline or Main profile is signalled via constraint_set3 */
sps->i_level_idc = 11;
}
/* Intra profiles */
if( param->i_keyint_max == 1 && sps->i_profile_idc > PROFILE_HIGH )
sps->b_constraint_set3 = 1;
sps->vui.i_num_reorder_frames = param->i_bframe_pyramid ? 2 : param->i_bframe ? 1 : 0;
/* extra slot with pyramid so that we don't have to override the
* order of forgetting old pictures */
sps->vui.i_max_dec_frame_buffering =
sps->i_num_ref_frames = X264_MIN(X264_REF_MAX, X264_MAX4(param->i_frame_reference, 1 + sps->vui.i_num_reorder_frames,
param->i_bframe_pyramid ? 4 : 1, param->i_dpb_size));
sps->i_num_ref_frames -= param->i_bframe_pyramid == X264_B_PYRAMID_STRICT;
if( param->i_keyint_max == 1 )
{
sps->i_num_ref_frames = 0;
sps->vui.i_max_dec_frame_buffering = 0;
}
/* number of refs + current frame */
int max_frame_num = sps->vui.i_max_dec_frame_buffering * (!!param->i_bframe_pyramid+1) + 1;
/* Intra refresh cannot write a recovery time greater than max frame num-1 */
if( param->b_intra_refresh )
{
int time_to_recovery = X264_MIN( sps->i_mb_width - 1, param->i_keyint_max ) + param->i_bframe - 1;
max_frame_num = X264_MAX( max_frame_num, time_to_recovery+1 );
}
sps->i_log2_max_frame_num = 4;
while( (1 << sps->i_log2_max_frame_num) <= max_frame_num )
sps->i_log2_max_frame_num++;
sps->i_poc_type = param->i_bframe || param->b_interlaced || param->i_avcintra_class ? 0 : 2;
if( sps->i_poc_type == 0 )
{
int max_delta_poc = (param->i_bframe + 2) * (!!param->i_bframe_pyramid + 1) * 2;
sps->i_log2_max_poc_lsb = 4;
while( (1 << sps->i_log2_max_poc_lsb) <= max_delta_poc * 2 )
sps->i_log2_max_poc_lsb++;
}
sps->b_vui = 1;
sps->b_gaps_in_frame_num_value_allowed = 0;
sps->b_frame_mbs_only = !(param->b_interlaced || param->b_fake_interlaced);
if( !sps->b_frame_mbs_only )
sps->i_mb_height = ( sps->i_mb_height + 1 ) & ~1;
sps->b_mb_adaptive_frame_field = param->b_interlaced;
sps->b_direct8x8_inference = 1;
sps->crop.i_left = param->crop_rect.i_left;
sps->crop.i_top = param->crop_rect.i_top;
sps->crop.i_right = param->crop_rect.i_right + sps->i_mb_width*16 - param->i_width;
sps->crop.i_bottom = (param->crop_rect.i_bottom + sps->i_mb_height*16 - param->i_height) >> !sps->b_frame_mbs_only;
sps->b_crop = sps->crop.i_left || sps->crop.i_top ||
sps->crop.i_right || sps->crop.i_bottom;
sps->vui.b_aspect_ratio_info_present = 0;
if( param->vui.i_sar_width > 0 && param->vui.i_sar_height > 0 )
{
sps->vui.b_aspect_ratio_info_present = 1;
sps->vui.i_sar_width = param->vui.i_sar_width;
sps->vui.i_sar_height= param->vui.i_sar_height;
}
sps->vui.b_overscan_info_present = param->vui.i_overscan > 0 && param->vui.i_overscan <= 2;
if( sps->vui.b_overscan_info_present )
sps->vui.b_overscan_info = ( param->vui.i_overscan == 2 ? 1 : 0 );
sps->vui.b_signal_type_present = 0;
sps->vui.i_vidformat = ( param->vui.i_vidformat >= 0 && param->vui.i_vidformat <= 5 ? param->vui.i_vidformat : 5 );
sps->vui.b_fullrange = ( param->vui.b_fullrange >= 0 && param->vui.b_fullrange <= 1 ? param->vui.b_fullrange :
( csp >= X264_CSP_BGR ? 1 : 0 ) );
sps->vui.b_color_description_present = 0;
sps->vui.i_colorprim = ( param->vui.i_colorprim >= 0 && param->vui.i_colorprim <= 9 ? param->vui.i_colorprim : 2 );
sps->vui.i_transfer = ( param->vui.i_transfer >= 0 && param->vui.i_transfer <= 15 ? param->vui.i_transfer : 2 );
sps->vui.i_colmatrix = ( param->vui.i_colmatrix >= 0 && param->vui.i_colmatrix <= 10 ? param->vui.i_colmatrix :
( csp >= X264_CSP_BGR ? 0 : 2 ) );
if( sps->vui.i_colorprim != 2 ||
sps->vui.i_transfer != 2 ||
sps->vui.i_colmatrix != 2 )
{
sps->vui.b_color_description_present = 1;
}
if( sps->vui.i_vidformat != 5 ||
sps->vui.b_fullrange ||
sps->vui.b_color_description_present )
{
sps->vui.b_signal_type_present = 1;
}
/* FIXME: not sufficient for interlaced video */
sps->vui.b_chroma_loc_info_present = param->vui.i_chroma_loc > 0 && param->vui.i_chroma_loc <= 5 &&
sps->i_chroma_format_idc == CHROMA_420;
if( sps->vui.b_chroma_loc_info_present )
{
sps->vui.i_chroma_loc_top = param->vui.i_chroma_loc;
sps->vui.i_chroma_loc_bottom = param->vui.i_chroma_loc;
}
sps->vui.b_timing_info_present = param->i_timebase_num > 0 && param->i_timebase_den > 0;
if( sps->vui.b_timing_info_present )
{
sps->vui.i_num_units_in_tick = param->i_timebase_num;
sps->vui.i_time_scale = param->i_timebase_den * 2;
sps->vui.b_fixed_frame_rate = !param->b_vfr_input;
}
sps->vui.b_vcl_hrd_parameters_present = 0; // we don't support VCL HRD
sps->vui.b_nal_hrd_parameters_present = !!param->i_nal_hrd;
sps->vui.b_pic_struct_present = param->b_pic_struct;
// NOTE: HRD related parts of the SPS are initialised in x264_ratecontrol_init_reconfigurable
sps->vui.b_bitstream_restriction = param->i_keyint_max > 1;
if( sps->vui.b_bitstream_restriction )
{
sps->vui.b_motion_vectors_over_pic_boundaries = 1;
sps->vui.i_max_bytes_per_pic_denom = 0;
sps->vui.i_max_bits_per_mb_denom = 0;
sps->vui.i_log2_max_mv_length_horizontal =
sps->vui.i_log2_max_mv_length_vertical = (int)log2f( X264_MAX( 1, param->analyse.i_mv_range*4-1 ) ) + 1;
}
}
void x264_sps_init( x264_sps_t *sps, int i_id, x264_param_t *param )
{
sps->i_id = i_id;
sps->b_qpprime_y_zero_transform_bypass = param->rc.i_rc_method == X264_RC_CQP && param->rc.i_qp_constant == 0;
if( sps->b_qpprime_y_zero_transform_bypass )
sps->i_profile_idc = PROFILE_HIGH444;
else if( param->analyse.b_transform_8x8 || param->i_cqm_preset != X264_CQM_FLAT )
sps->i_profile_idc = PROFILE_HIGH;
else if( param->b_cabac || param->i_bframe > 0 )
sps->i_profile_idc = PROFILE_MAIN;
else
sps->i_profile_idc = PROFILE_BASELINE;
sps->i_level_idc = param->i_level_idc;
sps->b_constraint_set0 = sps->i_profile_idc == PROFILE_BASELINE;
/* x264 doesn't support the features that are in Baseline and not in Main,
* namely arbitrary_slice_order and slice_groups. */
sps->b_constraint_set1 = sps->i_profile_idc <= PROFILE_MAIN;
/* Never set constraint_set2, it is not necessary and not used in real world. */
sps->b_constraint_set2 = 0;
sps->i_log2_max_frame_num = 4; /* at least 4 */
while( (1 << sps->i_log2_max_frame_num) <= param->i_keyint_max )
{
sps->i_log2_max_frame_num++;
}
sps->i_log2_max_frame_num++; /* just in case */
sps->i_poc_type = 0;
if( sps->i_poc_type == 0 )
{
sps->i_log2_max_poc_lsb = sps->i_log2_max_frame_num + 1; /* max poc = 2*frame_num */
}
else if( sps->i_poc_type == 1 )
{
int i;
/* FIXME */
sps->b_delta_pic_order_always_zero = 1;
sps->i_offset_for_non_ref_pic = 0;
sps->i_offset_for_top_to_bottom_field = 0;
sps->i_num_ref_frames_in_poc_cycle = 0;
for( i = 0; i < sps->i_num_ref_frames_in_poc_cycle; i++ )
{
sps->i_offset_for_ref_frame[i] = 0;
}
}
sps->b_vui = 1;
sps->b_gaps_in_frame_num_value_allowed = 0;
sps->i_mb_width = ( param->i_width + 15 ) / 16;
sps->i_mb_height= ( param->i_height + 15 ) / 16;
if( param->b_interlaced )
sps->i_mb_height = ( sps->i_mb_height + 1 ) & ~1;
sps->b_frame_mbs_only = ! param->b_interlaced;
sps->b_mb_adaptive_frame_field = param->b_interlaced;
sps->b_direct8x8_inference = param->analyse.i_direct_8x8_inference
|| ! sps->b_frame_mbs_only
|| !(param->analyse.inter & X264_ANALYSE_PSUB8x8);
sps->crop.i_left = 0;
sps->crop.i_top = 0;
sps->crop.i_right = sps->i_mb_width*16 - param->i_width;
sps->crop.i_bottom = (sps->i_mb_height*16 - param->i_height) >> param->b_interlaced;
sps->b_crop = sps->crop.i_left || sps->crop.i_top ||
sps->crop.i_right || sps->crop.i_bottom;
sps->vui.b_aspect_ratio_info_present = 0;
if( param->vui.i_sar_width > 0 && param->vui.i_sar_height > 0 )
{
sps->vui.b_aspect_ratio_info_present = 1;
sps->vui.i_sar_width = param->vui.i_sar_width;
sps->vui.i_sar_height= param->vui.i_sar_height;
}
sps->vui.b_overscan_info_present = ( param->vui.i_overscan ? 1 : 0 );
if( sps->vui.b_overscan_info_present )
sps->vui.b_overscan_info = ( param->vui.i_overscan == 2 ? 1 : 0 );
sps->vui.b_signal_type_present = 0;
sps->vui.i_vidformat = ( param->vui.i_vidformat <= 5 ? param->vui.i_vidformat : 5 );
sps->vui.b_fullrange = ( param->vui.b_fullrange ? 1 : 0 );
sps->vui.b_color_description_present = 0;
sps->vui.i_colorprim = ( param->vui.i_colorprim <= 9 ? param->vui.i_colorprim : 2 );
sps->vui.i_transfer = ( param->vui.i_transfer <= 11 ? param->vui.i_transfer : 2 );
sps->vui.i_colmatrix = ( param->vui.i_colmatrix <= 9 ? param->vui.i_colmatrix : 2 );
if( sps->vui.i_colorprim != 2 ||
sps->vui.i_transfer != 2 ||
sps->vui.i_colmatrix != 2 )
{
sps->vui.b_color_description_present = 1;
}
if( sps->vui.i_vidformat != 5 ||
sps->vui.b_fullrange ||
sps->vui.b_color_description_present )
{
sps->vui.b_signal_type_present = 1;
}
/* FIXME: not sufficient for interlaced video */
sps->vui.b_chroma_loc_info_present = ( param->vui.i_chroma_loc ? 1 : 0 );
if( sps->vui.b_chroma_loc_info_present )
{
sps->vui.i_chroma_loc_top = param->vui.i_chroma_loc;
sps->vui.i_chroma_loc_bottom = param->vui.i_chroma_loc;
}
sps->vui.b_timing_info_present = 0;
if( param->i_fps_num > 0 && param->i_fps_den > 0)
{
sps->vui.b_timing_info_present = 1;
sps->vui.i_num_units_in_tick = param->i_fps_den;
sps->vui.i_time_scale = param->i_fps_num * 2;
sps->vui.b_fixed_frame_rate = 1;
}
sps->vui.i_num_reorder_frames = param->b_bframe_pyramid ? 2 : param->i_bframe ? 1 : 0;
/* extra slot with pyramid so that we don't have to override the
* order of forgetting old pictures */
sps->vui.i_max_dec_frame_buffering =
sps->i_num_ref_frames = X264_MIN(16, X264_MAX(param->i_frame_reference, 1 + sps->vui.i_num_reorder_frames));
sps->vui.b_bitstream_restriction = 1;
if( sps->vui.b_bitstream_restriction )
{
sps->vui.b_motion_vectors_over_pic_boundaries = 1;
sps->vui.i_max_bytes_per_pic_denom = 0;
sps->vui.i_max_bits_per_mb_denom = 0;
sps->vui.i_log2_max_mv_length_horizontal =
sps->vui.i_log2_max_mv_length_vertical = (int)(log(param->analyse.i_mv_range*4-1)/log(2)) + 1;
}
}
void x264_sps_init( x264_sps_t *sps, int i_id, x264_param_t *param )
{
sps->i_id = i_id;
sps->b_qpprime_y_zero_transform_bypass = !param->rc.b_cbr && param->rc.i_qp_constant == 0;
if( sps->b_qpprime_y_zero_transform_bypass )
sps->i_profile_idc = PROFILE_HIGH444;
else if( param->analyse.b_transform_8x8 || param->i_cqm_preset != X264_CQM_FLAT )
sps->i_profile_idc = PROFILE_HIGH;
else if( param->b_cabac || param->i_bframe > 0 )
sps->i_profile_idc = PROFILE_MAIN;
else
sps->i_profile_idc = PROFILE_BASELINE;
sps->i_level_idc = param->i_level_idc;
sps->b_constraint_set0 = 0;
sps->b_constraint_set1 = 0;
sps->b_constraint_set2 = 0;
sps->i_log2_max_frame_num = 4; /* at least 4 */
while( (1 << sps->i_log2_max_frame_num) <= param->i_keyint_max )
{
sps->i_log2_max_frame_num++;
}
sps->i_log2_max_frame_num++; /* just in case */
sps->i_poc_type = 0;
if( sps->i_poc_type == 0 )
{
sps->i_log2_max_poc_lsb = sps->i_log2_max_frame_num + 1; /* max poc = 2*frame_num */
}
else if( sps->i_poc_type == 1 )
{
int i;
/* FIXME */
sps->b_delta_pic_order_always_zero = 1;
sps->i_offset_for_non_ref_pic = 0;
sps->i_offset_for_top_to_bottom_field = 0;
sps->i_num_ref_frames_in_poc_cycle = 0;
for( i = 0; i < sps->i_num_ref_frames_in_poc_cycle; i++ )
{
sps->i_offset_for_ref_frame[i] = 0;
}
}
sps->vui.i_num_reorder_frames = param->b_bframe_pyramid ? 2 : param->i_bframe ? 1 : 0;
sps->vui.i_max_dec_frame_buffering =
sps->i_num_ref_frames = X264_MIN(16, param->i_frame_reference + sps->vui.i_num_reorder_frames);
sps->b_gaps_in_frame_num_value_allowed = 0;
sps->i_mb_width = ( param->i_width + 15 ) / 16;
sps->i_mb_height= ( param->i_height + 15 )/ 16;
sps->b_frame_mbs_only = 1;
sps->b_mb_adaptive_frame_field = 0;
sps->b_direct8x8_inference = 0;
if( sps->b_frame_mbs_only == 0 ||
!(param->analyse.inter & X264_ANALYSE_PSUB8x8) )
{
sps->b_direct8x8_inference = 1;
}
if( param->i_width % 16 != 0 || param->i_height % 16 != 0 )
{
sps->b_crop = 1;
sps->crop.i_left = 0;
sps->crop.i_right = ( 16 - param->i_width % 16)/2;
sps->crop.i_top = 0;
sps->crop.i_bottom = ( 16 - param->i_height % 16)/2;
}
else
{
sps->b_crop = 0;
sps->crop.i_left = 0;
sps->crop.i_right = 0;
sps->crop.i_top = 0;
sps->crop.i_bottom = 0;
}
sps->b_vui = 0;
sps->vui.b_aspect_ratio_info_present = 0;
if( param->vui.i_sar_width > 0 && param->vui.i_sar_height > 0 )
{
sps->vui.b_aspect_ratio_info_present = 1;
sps->vui.i_sar_width = param->vui.i_sar_width;
sps->vui.i_sar_height= param->vui.i_sar_height;
}
sps->b_vui |= sps->vui.b_aspect_ratio_info_present;
if( param->i_fps_num > 0 && param->i_fps_den > 0)
{
sps->vui.b_timing_info_present = 1;
/* The standard is confusing me, but this seems to work best
with other encoders */
sps->vui.i_num_units_in_tick = param->i_fps_den;
sps->vui.i_time_scale = param->i_fps_num;
sps->vui.b_fixed_frame_rate = 1;
}
sps->b_vui |= sps->vui.b_timing_info_present;
sps->vui.b_bitstream_restriction = param->i_bframe > 0;
if( sps->vui.b_bitstream_restriction )
{
sps->vui.b_motion_vectors_over_pic_boundaries = 1;
sps->vui.i_max_bytes_per_pic_denom = 0;
sps->vui.i_max_bits_per_mb_denom = 0;
sps->vui.i_log2_max_mv_length_horizontal =
sps->vui.i_log2_max_mv_length_vertical = (int)(log(param->analyse.i_mv_range*4-1)/log(2)) + 1;
}
sps->b_vui |= sps->vui.b_bitstream_restriction;
}
static int init( hnd_t *handle, cli_vid_filter_t *filter, video_info_t *info, x264_param_t *param, char *opt_string )
{
selvry_hnd_t *h = malloc( sizeof(selvry_hnd_t) );
if( !h )
return -1;
h->pattern_len = 0;
h->step_size = 0;
int offsets[MAX_PATTERN_SIZE];
for( char *tok, *p = opt_string; (tok = strtok( p, "," )); p = NULL )
{
int val = x264_otoi( tok, -1 );
if( p )
{
FAIL_IF_ERROR( val <= 0, "invalid step `%s'\n", tok )
h->step_size = val;
continue;
}
FAIL_IF_ERROR( val < 0 || val >= h->step_size, "invalid offset `%s'\n", tok )
FAIL_IF_ERROR( h->pattern_len >= MAX_PATTERN_SIZE, "max pattern size %d reached\n", MAX_PATTERN_SIZE )
offsets[h->pattern_len++] = val;
}
FAIL_IF_ERROR( !h->step_size, "no step size provided\n" )
FAIL_IF_ERROR( !h->pattern_len, "no offsets supplied\n" )
h->pattern = malloc( h->pattern_len * sizeof(int) );
if( !h->pattern )
return -1;
memcpy( h->pattern, offsets, h->pattern_len * sizeof(int) );
/* determine required cache size to maintain pattern. */
intptr_t max_rewind = 0;
int min = h->step_size;
for( int i = h->pattern_len-1; i >= 0; i-- )
{
min = X264_MIN( min, offsets[i] );
if( i )
max_rewind = X264_MAX( max_rewind, offsets[i-1] - min + 1 );
/* reached maximum rewind size */
if( max_rewind == h->step_size )
break;
}
if( x264_init_vid_filter( "cache", handle, filter, info, param, (void*)max_rewind ) )
return -1;
/* done initing, overwrite properties */
if( h->step_size != h->pattern_len )
{
info->num_frames = (uint64_t)info->num_frames * h->pattern_len / h->step_size;
info->fps_den *= h->step_size;
info->fps_num *= h->pattern_len;
x264_reduce_fraction( &info->fps_num, &info->fps_den );
if( info->vfr )
{
info->timebase_den *= h->pattern_len;
info->timebase_num *= h->step_size;
x264_reduce_fraction( &info->timebase_num, &info->timebase_den );
}
}
h->pts = 0;
h->vfr = info->vfr;
h->prev_filter = *filter;
h->prev_hnd = *handle;
*filter = select_every_filter;
*handle = h;
return 0;
}
void x264_sps_init( x264_sps_t *sps, int i_id, x264_param_t *param )
{
sps->i_id = i_id;
int max_frame_num;
sps->b_qpprime_y_zero_transform_bypass = param->rc.i_rc_method == X264_RC_CQP && param->rc.i_qp_constant == 0;
if( sps->b_qpprime_y_zero_transform_bypass )
sps->i_profile_idc = PROFILE_HIGH444_PREDICTIVE;
else if( BIT_DEPTH > 8 )
sps->i_profile_idc = PROFILE_HIGH10;
else if( param->analyse.b_transform_8x8 || param->i_cqm_preset != X264_CQM_FLAT )
sps->i_profile_idc = PROFILE_HIGH;
else if( param->b_cabac || param->i_bframe > 0 || param->b_interlaced || param->b_fake_interlaced || param->analyse.i_weighted_pred > 0 )
sps->i_profile_idc = PROFILE_MAIN;
else
sps->i_profile_idc = PROFILE_BASELINE;
sps->i_level_idc = param->i_level_idc;
sps->b_constraint_set0 = sps->i_profile_idc == PROFILE_BASELINE;
/* x264 doesn't support the features that are in Baseline and not in Main,
* namely arbitrary_slice_order and slice_groups. */
sps->b_constraint_set1 = sps->i_profile_idc <= PROFILE_MAIN;
/* Never set constraint_set2, it is not necessary and not used in real world. */
sps->b_constraint_set2 = 0;
sps->vui.i_num_reorder_frames = param->i_bframe_pyramid ? 2 : param->i_bframe ? 1 : 0;
/* extra slot with pyramid so that we don't have to override the
* order of forgetting old pictures */
sps->vui.i_max_dec_frame_buffering =
sps->i_num_ref_frames = X264_MIN(X264_REF_MAX, X264_MAX4(param->i_frame_reference, 1 + sps->vui.i_num_reorder_frames,
param->i_bframe_pyramid ? 4 : 1, param->i_dpb_size));
sps->i_num_ref_frames -= param->i_bframe_pyramid == X264_B_PYRAMID_STRICT;
/* number of refs + current frame */
max_frame_num = sps->vui.i_max_dec_frame_buffering * (!!param->i_bframe_pyramid+1) + 1;
sps->i_log2_max_frame_num = 4;
while( (1 << sps->i_log2_max_frame_num) <= max_frame_num )
sps->i_log2_max_frame_num++;
sps->i_poc_type = 0;
if( sps->i_poc_type == 0 )
{
int max_delta_poc = (param->i_bframe + 2) * (!!param->i_bframe_pyramid + 1) * 2;
sps->i_log2_max_poc_lsb = 4;
while( (1 << sps->i_log2_max_poc_lsb) <= max_delta_poc * 2 )
sps->i_log2_max_poc_lsb++;
}
else if( sps->i_poc_type == 1 )
{
int i;
/* FIXME */
sps->b_delta_pic_order_always_zero = 1;
sps->i_offset_for_non_ref_pic = 0;
sps->i_offset_for_top_to_bottom_field = 0;
sps->i_num_ref_frames_in_poc_cycle = 0;
for( i = 0; i < sps->i_num_ref_frames_in_poc_cycle; i++ )
{
sps->i_offset_for_ref_frame[i] = 0;
}
}
sps->b_vui = 1;
sps->b_gaps_in_frame_num_value_allowed = 0;
sps->i_mb_width = ( param->i_width + 15 ) / 16;
sps->i_mb_height= ( param->i_height + 15 ) / 16;
sps->b_frame_mbs_only = !(param->b_interlaced || param->b_fake_interlaced);
if( !sps->b_frame_mbs_only )
sps->i_mb_height = ( sps->i_mb_height + 1 ) & ~1;
sps->b_mb_adaptive_frame_field = param->b_interlaced;
sps->b_direct8x8_inference = 1;
sps->crop.i_left = 0;
sps->crop.i_top = 0;
sps->crop.i_right = sps->i_mb_width*16 - param->i_width;
sps->crop.i_bottom = (sps->i_mb_height*16 - param->i_height) >> !sps->b_frame_mbs_only;
sps->b_crop = sps->crop.i_left || sps->crop.i_top ||
sps->crop.i_right || sps->crop.i_bottom;
sps->vui.b_aspect_ratio_info_present = 0;
if( param->vui.i_sar_width > 0 && param->vui.i_sar_height > 0 )
{
sps->vui.b_aspect_ratio_info_present = 1;
sps->vui.i_sar_width = param->vui.i_sar_width;
sps->vui.i_sar_height= param->vui.i_sar_height;
}
sps->vui.b_overscan_info_present = ( param->vui.i_overscan ? 1 : 0 );
if( sps->vui.b_overscan_info_present )
sps->vui.b_overscan_info = ( param->vui.i_overscan == 2 ? 1 : 0 );
sps->vui.b_signal_type_present = 0;
sps->vui.i_vidformat = ( param->vui.i_vidformat <= 5 ? param->vui.i_vidformat : 5 );
sps->vui.b_fullrange = ( param->vui.b_fullrange ? 1 : 0 );
sps->vui.b_color_description_present = 0;
sps->vui.i_colorprim = ( param->vui.i_colorprim <= 9 ? param->vui.i_colorprim : 2 );
sps->vui.i_transfer = ( param->vui.i_transfer <= 11 ? param->vui.i_transfer : 2 );
sps->vui.i_colmatrix = ( param->vui.i_colmatrix <= 9 ? param->vui.i_colmatrix : 2 );
if( sps->vui.i_colorprim != 2 ||
sps->vui.i_transfer != 2 ||
sps->vui.i_colmatrix != 2 )
{
sps->vui.b_color_description_present = 1;
}
if( sps->vui.i_vidformat != 5 ||
sps->vui.b_fullrange ||
sps->vui.b_color_description_present )
{
sps->vui.b_signal_type_present = 1;
}
/* FIXME: not sufficient for interlaced video */
sps->vui.b_chroma_loc_info_present = ( param->vui.i_chroma_loc ? 1 : 0 );
if( sps->vui.b_chroma_loc_info_present )
{
sps->vui.i_chroma_loc_top = param->vui.i_chroma_loc;
sps->vui.i_chroma_loc_bottom = param->vui.i_chroma_loc;
}
sps->vui.b_timing_info_present = param->i_timebase_num > 0 && param->i_timebase_den > 0;
if( sps->vui.b_timing_info_present )
{
sps->vui.i_num_units_in_tick = param->i_timebase_num;
sps->vui.i_time_scale = param->i_timebase_den * 2;
sps->vui.b_fixed_frame_rate = !param->b_vfr_input;
}
sps->vui.b_vcl_hrd_parameters_present = 0; // we don't support VCL HRD
sps->vui.b_nal_hrd_parameters_present = !!param->i_nal_hrd;
sps->vui.b_pic_struct_present = param->b_pic_struct;
// NOTE: HRD related parts of the SPS are initialised in x264_ratecontrol_init_reconfigurable
sps->vui.b_bitstream_restriction = 1;
if( sps->vui.b_bitstream_restriction )
{
sps->vui.b_motion_vectors_over_pic_boundaries = 1;
sps->vui.i_max_bytes_per_pic_denom = 0;
sps->vui.i_max_bits_per_mb_denom = 0;
sps->vui.i_log2_max_mv_length_horizontal =
sps->vui.i_log2_max_mv_length_vertical = (int)log2f( X264_MAX( 1, param->analyse.i_mv_range*4-1 ) ) + 1;
}
}
/****************************************************************************
* block_residual_write_cavlc:
****************************************************************************/
static void block_residual_write_cavlc( x264_t *h, bs_t *s, int i_idx, int *l, int i_count )
{
int level[16], run[16];
int i_total, i_trailing;
int i_total_zero;
int i_last;
unsigned int i_sign;
int i;
int i_zero_left;
int i_suffix_length;
/* first find i_last */
i_last = i_count - 1;
while( i_last >= 0 && l[i_last] == 0 )
{
i_last--;
}
i_sign = 0;
i_total = 0;
i_trailing = 0;
i_total_zero = i_last + 1;
if( i_last >= 0 )
{
int idx = 0;
/* level and run and total */
while( i_last >= 0 )
{
level[idx] = l[i_last--];
run[idx] = 0;
while( i_last >= 0 && l[i_last] == 0 )
{
run[idx]++;
i_last--;
}
idx++;
}
i_total = idx;
i_total_zero -= idx;
i_trailing = X264_MIN(3, idx);
for( idx = 0; idx < i_trailing; idx++ )
{
if( abs(level[idx]) > 1 )
{
i_trailing = idx;
break;
}
i_sign <<= 1;
i_sign |= level[idx] < 0;
}
}
/* total/trailing */
if( i_idx == BLOCK_INDEX_CHROMA_DC )
{
bs_write_vlc( s, x264_coeff_token[4][i_total*4+i_trailing] );
}
else
{
/* x264_mb_predict_non_zero_code return 0 <-> (16+16+1)>>1 = 16 */
static const int ct_index[17] = {0,0,1,1,2,2,2,2,3,3,3,3,3,3,3,3,3 };
int nC = x264_mb_predict_non_zero_code( h, i_idx == BLOCK_INDEX_LUMA_DC ? 0 : i_idx );
bs_write_vlc( s, x264_coeff_token[ct_index[nC]][i_total*4+i_trailing] );
}
if( i_total <= 0 )
{
return;
}
i_suffix_length = i_total > 10 && i_trailing < 3 ? 1 : 0;
if( i_trailing > 0 )
{
bs_write( s, i_trailing, i_sign );
}
for( i = i_trailing; i < i_total; i++ )
{
unsigned int i_level_code;
/* calculate level code */
if( level[i] < 0 )
{
i_level_code = -2*level[i] - 1;
}
else /* if( level[i] > 0 ) */
{
i_level_code = 2 * level[i] - 2;
}
if( i == i_trailing && i_trailing < 3 )
{
i_level_code -= 2; /* as level[i] can't be 1 for the first one if i_trailing < 3 */
}
if( ( i_level_code >> i_suffix_length ) < 14 )
{
bs_write_vlc( s, x264_level_prefix[i_level_code >> i_suffix_length] );
if( i_suffix_length > 0 )
{
bs_write( s, i_suffix_length, i_level_code );
}
}
else if( i_suffix_length == 0 && i_level_code < 30 )
{
bs_write_vlc( s, x264_level_prefix[14] );
bs_write( s, 4, i_level_code - 14 );
}
else if( i_suffix_length > 0 && ( i_level_code >> i_suffix_length ) == 14 )
{
bs_write_vlc( s, x264_level_prefix[14] );
bs_write( s, i_suffix_length, i_level_code );
}
else
{
void x264_me_search_ref( x264_t *h, x264_me_t *m, int (*mvc)[2], int i_mvc, int *p_fullpel_thresh )
{
const int i_pixel = m->i_pixel;
const int i_me_range = h->param.analyse.i_me_range;
const int b_chroma_me = h->mb.b_chroma_me && i_pixel <= PIXEL_8x8;
int bmx, bmy, bcost;
int omx, omy, pmx, pmy;
uint8_t *p_fref = m->p_fref[0];
int i, j;
int mv_x_min = h->mb.mv_min_fpel[0];
int mv_y_min = h->mb.mv_min_fpel[1];
int mv_x_max = h->mb.mv_max_fpel[0];
int mv_y_max = h->mb.mv_max_fpel[1];
const int16_t *p_cost_mvx = m->p_cost_mv - m->mvp[0];
const int16_t *p_cost_mvy = m->p_cost_mv - m->mvp[1];
if( h->mb.i_me_method == X264_ME_UMH )
{
/* clamp mvp to inside frame+padding, so that we don't have to check it each iteration */
p_cost_mvx = m->p_cost_mv - x264_clip3( m->mvp[0], h->mb.mv_min[0], h->mb.mv_max[0] );
p_cost_mvy = m->p_cost_mv - x264_clip3( m->mvp[1], h->mb.mv_min[1], h->mb.mv_max[1] );
}
bmx = pmx = x264_clip3( ( m->mvp[0] + 2 ) >> 2, mv_x_min, mv_x_max );
bmy = pmy = x264_clip3( ( m->mvp[1] + 2 ) >> 2, mv_y_min, mv_y_max );
bcost = COST_MAX;
COST_MV( bmx, bmy );
/* I don't know why this helps */
bcost -= p_cost_mvx[ bmx<<2 ] + p_cost_mvy[ bmy<<2 ];
/* try extra predictors if provided */
for( i = 0; i < i_mvc; i++ )
{
const int mx = x264_clip3( ( mvc[i][0] + 2 ) >> 2, mv_x_min, mv_x_max );
const int my = x264_clip3( ( mvc[i][1] + 2 ) >> 2, mv_y_min, mv_y_max );
if( mx != bmx || my != bmy )
COST_MV( mx, my );
}
COST_MV( 0, 0 );
mv_x_max += 8;
mv_y_max += 8;
mv_x_min -= 8;
mv_y_min -= 8;
switch( h->mb.i_me_method )
{
case X264_ME_DIA:
/* diamond search, radius 1 */
#define DIA1_ITER(mx, my)\
{\
omx = mx;\
omy = my;\
COST_MV( omx , omy-1 );\
COST_MV( omx , omy+1 );\
COST_MV( omx-1, omy );\
COST_MV( omx+1, omy );\
}
for( i = 0; i < i_me_range; i++ )
{
DIA1_ITER( bmx, bmy );
if( bmx == omx && bmy == omy )
break;
}
break;
case X264_ME_HEX:
/* hexagon search, radius 2 */
#define HEX2_ITER(mx, my)\
{\
omx = mx;\
omy = my;\
COST_MV( omx-2, omy );\
COST_MV( omx-1, omy+2 );\
COST_MV( omx+1, omy+2 );\
COST_MV( omx+2, omy );\
COST_MV( omx+1, omy-2 );\
COST_MV( omx-1, omy-2 );\
}
for( i = 0; i < i_me_range/2; i++ )
{
HEX2_ITER( bmx, bmy );
if( bmx == omx && bmy == omy )
break;
}
/* square refine */
DIA1_ITER( bmx, bmy );
COST_MV( omx-1, omy-1 );
COST_MV( omx-1, omy+1 );
COST_MV( omx+1, omy-1 );
COST_MV( omx+1, omy+1 );
break;
case X264_ME_UMH:
/* Uneven-cross Multi-Hexagon-grid Search
* as in JM, except without early termination */
DIA1_ITER( pmx, pmy );
if( pmx || pmy )
DIA1_ITER( 0, 0 );
DIA1_ITER( bmx, bmy );
if(i_pixel == PIXEL_4x4)
goto umh_small_hex;
/* cross */
omx = bmx; omy = bmy;
for( i = 1; i < i_me_range; i+=2 )
{
if( omx + i <= mv_x_max )
COST_MV( omx + i, omy );
if( omx - i >= mv_x_min )
COST_MV( omx - i, omy );
}
for( i = 1; i < i_me_range/2; i+=2 )
{
if( omy + i <= mv_y_max )
COST_MV( omx, omy + i );
if( omy - i >= mv_y_min )
COST_MV( omx, omy - i );
}
/* 5x5 ESA */
omx = bmx; omy = bmy;
for( i = 0; i < 24; i++ )
{
static const int square2_x[24] = {1,1,0,-1,-1,-1, 0, 1, 2,2,2,2,1,0,-1,-2,-2,-2,-2,-2,-1, 0, 1, 2};
static const int square2_y[24] = {0,1,1, 1, 0,-1,-1,-1,-1,0,1,2,2,2, 2, 2, 1, 0,-1,-2,-2,-2,-2,-2};
COST_MV( omx + square2_x[i], omy + square2_y[i] );
}
/* hexagon grid */
omx = bmx; omy = bmy;
for( i = 1; i <= i_me_range/4; i++ )
{
int bounds_check = 4*i > X264_MIN4( mv_x_max-omx, mv_y_max-omy, omx-mv_x_min, omy-mv_y_min );
for( j = 0; j < 16; j++ )
{
static const int hex4_x[16] = {0,-2,-4,-4,-4,-4,-4,-2, 0, 2, 4, 4,4,4,4,2};
static const int hex4_y[16] = {4, 3, 2, 1, 0,-1,-2,-3,-4,-3,-2,-1,0,1,2,3};
int mx = omx + hex4_x[j]*i;
int my = omy + hex4_y[j]*i;
if( !bounds_check || ( mx >= mv_x_min && mx <= mv_x_max
&& my >= mv_y_min && my <= mv_y_max ) )
COST_MV( mx, my );
}
}
umh_small_hex:
/* iterative search */
for( i = 0; i < i_me_range; i++ )
{
HEX2_ITER( bmx, bmy );
if( bmx == omx && bmy == omy )
break;
}
for( i = 0; i < i_me_range; i++ )
{
DIA1_ITER( bmx, bmy );
if( bmx == omx && bmy == omy )
break;
}
break;
case X264_ME_ESA:
{
const int min_x = X264_MAX( bmx - i_me_range, mv_x_min);
const int min_y = X264_MAX( bmy - i_me_range, mv_y_min);
const int max_x = X264_MIN( bmx + i_me_range, mv_x_max);
const int max_y = X264_MIN( bmy + i_me_range, mv_y_max);
for( omy = min_y; omy <= max_y; omy++ )
for( omx = min_x; omx <= max_x; omx++ )
{
COST_MV( omx, omy );
}
}
break;
}
/* -> qpel mv */
m->mv[0] = bmx << 2;
m->mv[1] = bmy << 2;
/* compute the real cost */
m->cost_mv = p_cost_mvx[ m->mv[0] ] + p_cost_mvy[ m->mv[1] ];
m->cost = h->pixf.mbcmp[i_pixel]( m->p_fenc[0], m->i_stride[0],
&p_fref[bmy * m->i_stride[0] + bmx], m->i_stride[0] )
+ m->cost_mv;
if( b_chroma_me )
{
const int bw = x264_pixel_size[m->i_pixel].w;
const int bh = x264_pixel_size[m->i_pixel].h;
uint8_t pix[8*8*2];
h->mc.mc_chroma( m->p_fref[4], m->i_stride[1], pix, 8, m->mv[0], m->mv[1], bw/2, bh/2 );
h->mc.mc_chroma( m->p_fref[5], m->i_stride[1], pix+8*8, 8, m->mv[0], m->mv[1], bw/2, bh/2 );
m->cost += h->pixf.mbcmp[i_pixel+3]( m->p_fenc[1], m->i_stride[1], pix, 8 )
+ h->pixf.mbcmp[i_pixel+3]( m->p_fenc[2], m->i_stride[1], pix+8*8, 8 );
}
/* subpel refine */
if( h->mb.i_subpel_refine >= 3 )
{
int hpel, qpel;
/* early termination (when examining multiple reference frames)
* FIXME: this can update fullpel_thresh even if the match
* ref is rejected after subpel refinement */
if( p_fullpel_thresh )
{
if( (m->cost*7)>>3 > *p_fullpel_thresh )
return;
else if( m->cost < *p_fullpel_thresh )
*p_fullpel_thresh = m->cost;
}
void x264_speedcontrol_frame( x264_t *h )
{
x264_speedcontrol_t *sc = h->sc;
int64_t t, delta_t, delta_buffer;
int delta_f;
x264_emms();
// update buffer state after encoding and outputting the previous frame(s)
t = x264_mdate();
delta_f = h->i_frame - sc->prev_frame;
delta_t = t - sc->timestamp;
delta_buffer = delta_f * sc->spf / h->param.sc.f_speed - delta_t;
sc->buffer_fill += delta_buffer;
sc->prev_frame = h->i_frame;
sc->timestamp = t;
// update the time predictor
if( delta_f )
{
int cpu_time = h->param.sc.b_alt_timer ? sc->cpu_time : delta_t;
float decay = powf( sc->cplx_decay, delta_f );
sc->cplx_num *= decay;
sc->cplx_den *= decay;
sc->cplx_num += cpu_time / presets[sc->preset].time;
sc->cplx_den += delta_f;
sc->stat.avg_preset += sc->preset * delta_f;
sc->stat.den += delta_f;
}
sc->stat.min_buffer = X264_MIN( sc->buffer_fill, sc->stat.min_buffer );
sc->stat.max_buffer = X264_MAX( sc->buffer_fill, sc->stat.max_buffer );
if( sc->buffer_fill > sc->buffer_size ) // oops, cpu was idle
{
// not really an error, but we'll warn for debugging purposes
static int64_t idle_t = 0, print_interval = 0;
idle_t += sc->buffer_fill - sc->buffer_size;
if( t - print_interval > 1e6 )
{
x264_log( h, X264_LOG_WARNING, "speedcontrol idle (%.6f sec)\n", idle_t/1e6 );
print_interval = t;
idle_t = 0;
}
sc->buffer_fill = sc->buffer_size;
}
else if( sc->buffer_fill < 0 && delta_buffer < 0 ) // oops, we're late
{
// don't clip fullness to 0; we'll hope the real buffer was bigger than
// specified, and maybe we can catch up. if the application had to drop
// frames, then it should override the buffer fullness (FIXME implement this).
x264_log( h, X264_LOG_WARNING, "speedcontrol underflow (%.6f sec)\n", sc->buffer_fill/1e6 );
}
{
// pick the preset that should return the buffer to 3/4-full within a time
// specified by compensation_period
float target = sc->spf / h->param.sc.f_speed
* (sc->buffer_fill + sc->compensation_period)
/ (sc->buffer_size*3/4 + sc->compensation_period);
float cplx = sc->cplx_num / sc->cplx_den;
float set, t0, t1;
float filled = (float) sc->buffer_fill / sc->buffer_size;
int i;
t0 = presets[0].time * cplx;
for( i=1;; i++ )
{
t1 = presets[i].time * cplx;
if( t1 >= target || i == PRESETS-1 )
break;
t0 = t1;
}
// linear interpolation between states
set = i-1 + (target - t0) / (t1 - t0);
// Even if our time estimations in the PRESETS array are off
// this will push us towards our target fullness
set += (20 * (filled-0.75));
set = x264_clip3f(set,0,PRESETS-1);
apply_preset( h, dither( sc, set ) );
// FIXME
if (h->param.i_log_level >= X264_LOG_DEBUG)
{
static float cpu, wall, tgt, den;
float decay = 1-1/100.;
cpu = cpu*decay + sc->cpu_time;
wall = wall*decay + delta_t;
tgt = tgt*decay + target;
den = den*decay + 1;
fprintf( stderr, "speed: %.2f %d[%.5f] (t/c/w: %6.0f/%6.0f/%6.0f = %.4f) fps=%.2f\r",
set, sc->preset, (float)sc->buffer_fill / sc->buffer_size,
tgt/den, cpu/den, wall/den, cpu/wall, 1e6*den/wall );
}
}
}
/* Result must be freed */
x264_param_t *x264_gtk_param_get (X264_Gtk *x264_gtk)
{
x264_param_t *param;
if (!x264_gtk)
return NULL;
param = (x264_param_t *)g_malloc (sizeof (x264_param_t));
if (!param)
return NULL;
x264_param_default (param);
/* rate control */
param->rc.f_ip_factor = 1.0 + (double)x264_gtk->keyframe_boost / 100.0;
param->rc.f_pb_factor = 1.0 + (double)x264_gtk->bframes_reduction / 100.0;
param->rc.f_qcompress = (double)x264_gtk->bitrate_variability / 100.0;
param->rc.i_qp_min = x264_gtk->min_qp;
param->rc.i_qp_max = x264_gtk->max_qp;
param->rc.i_qp_step = x264_gtk->max_qp_step;
param->i_scenecut_threshold = x264_gtk->scene_cut_threshold;
param->i_keyint_min = x264_gtk->min_idr_frame_interval;
param->i_keyint_max = x264_gtk->max_idr_frame_interval;
param->rc.i_vbv_max_bitrate = x264_gtk->vbv_max_bitrate;
param->rc.i_vbv_buffer_size = x264_gtk->vbv_buffer_size;
param->rc.f_vbv_buffer_init = x264_gtk->vbv_buffer_init;
/* mb */
param->analyse.b_transform_8x8 = x264_gtk->transform_8x8;
param->analyse.inter = 0;
if (x264_gtk->pframe_search_8)
param->analyse.inter |= X264_ANALYSE_PSUB16x16;
if (x264_gtk->bframe_search_8)
param->analyse.inter |= X264_ANALYSE_BSUB16x16;
if (x264_gtk->pframe_search_4)
param->analyse.inter |= X264_ANALYSE_PSUB8x8;
if (x264_gtk->inter_search_8)
param->analyse.inter |= X264_ANALYSE_I8x8;
if (x264_gtk->inter_search_4)
param->analyse.inter |= X264_ANALYSE_I4x4;
param->b_bframe_pyramid = x264_gtk->bframe_pyramid && x264_gtk->bframe;
param->analyse.b_bidir_me = x264_gtk->bidir_me;
param->b_bframe_adaptive = x264_gtk->bframe_adaptive;
param->analyse.b_weighted_bipred = x264_gtk->weighted_bipred;
param->i_bframe = x264_gtk->bframe;
param->i_bframe_bias = x264_gtk->bframe_bias;
param->analyse.i_direct_mv_pred = x264_gtk->direct_mode;
/* param->b_bframe_pyramid = param->b_bframe_pyramid && (param->i_bframe > 1); */
/* more */
param->analyse.i_subpel_refine = x264_gtk->partition_decision + 1;
param->analyse.b_bframe_rdo = x264_gtk->bframe_rdo;
param->analyse.i_me_method = x264_gtk->me_method;
param->analyse.i_me_range = x264_gtk->range;
param->analyse.b_chroma_me = x264_gtk->chroma_me;
param->analyse.i_trellis = x264_gtk->trellis;
param->analyse.i_noise_reduction = x264_gtk->noise_reduction;
param->i_frame_reference = x264_gtk->max_ref_frames;
param->analyse.b_mixed_references = x264_gtk->mixed_refs;
param->analyse.b_fast_pskip = x264_gtk->fast_pskip;
param->analyse.b_dct_decimate = x264_gtk->dct_decimate;
/* rdo : RD based mode decision for B-frames. Requires subme 6 */
param->vui.i_sar_width = x264_gtk->sample_ar_x;
param->vui.i_sar_height = x264_gtk->sample_ar_y;
param->i_threads = x264_gtk->threads;
param->b_cabac = x264_gtk->cabac;
param->b_deblocking_filter = x264_gtk->deblocking_filter;
param->i_deblocking_filter_alphac0 = x264_gtk->strength;
param->i_deblocking_filter_beta = x264_gtk->threshold;
param->i_log_level = x264_gtk->debug_method - 1;
/* cqm */
param->i_cqm_preset = x264_gtk->cqm_preset;
if (x264_gtk->cqm_file && (x264_gtk->cqm_file[0] != '\0'))
param->psz_cqm_file = x264_gtk->cqm_file;
memcpy( param->cqm_4iy, x264_gtk->cqm_4iy, 16 );
memcpy( param->cqm_4ic, x264_gtk->cqm_4ic, 16 );
memcpy( param->cqm_4py, x264_gtk->cqm_4py, 16 );
memcpy( param->cqm_4pc, x264_gtk->cqm_4pc, 16 );
memcpy( param->cqm_8iy, x264_gtk->cqm_8iy, 64 );
memcpy( param->cqm_8py, x264_gtk->cqm_8py, 64 );
/* bitrate */
switch (x264_gtk->pass) {
case X264_PASS_SINGLE_BITRATE:
param->rc.i_rc_method = X264_RC_ABR;
param->rc.i_bitrate = x264_gtk->average_bitrate;
break;
case X264_PASS_SINGLE_QUANTIZER:
param->rc.i_rc_method = X264_RC_CQP;
param->rc.i_qp_constant = x264_gtk->quantizer;
break;
case X264_PASS_MULTIPASS_1ST_FAST:
param->analyse.i_subpel_refine = X264_MAX( X264_MIN( 3, param->analyse.i_subpel_refine - 1 ), 1 );
param->i_frame_reference = ( param->i_frame_reference + 1 ) >> 1;
param->analyse.inter &= ( ~X264_ANALYSE_PSUB8x8 );
param->analyse.inter &= ( ~X264_ANALYSE_BSUB16x16 );
case X264_PASS_MULTIPASS_1ST:
param->rc.i_rc_method = X264_RC_ABR;
param->rc.i_bitrate = x264_gtk->average_bitrate;
param->rc.f_rate_tolerance = 4.0;
break;
case X264_PASS_MULTIPASS_NTH:
param->rc.i_rc_method = X264_RC_ABR;
param->rc.i_bitrate = x264_gtk->average_bitrate;
param->rc.f_rate_tolerance = 1.0;
break;
}
param->rc.b_stat_write = x264_gtk->stat_write;
param->rc.b_stat_read = x264_gtk->stat_read;
/* FIXME: potential mem leak... */
param->rc.psz_stat_out = x264_gtk_path (x264_gtk->statsfile_name);
return param;
}
