x264会对每一帧检查它与前面一帧的差异，当差异达到一定阈值时，就认为发生了场景切换，并把该帧置位I帧或IDR帧。在x264_slicetype_analyse函数中，我们可以看到scenecut的调用：
void x264_slicetype_analyse( x264_t *h, int intra_minigop )
{
//省略
//如果frames[1]还未确定类型，且i_scenecut_threshold场景切换的阈值不为0，则进行scenecut判断，若返回非0，则将frames[1]设为I帧
if( IS_X264_TYPE_AUTO_OR_I( frames[1]->i_type ) &&
h->param.i_scenecut_threshold && scenecut( h, &a, frames, 0, 1, 1, orig_num_frames, i_max_search ) )
{
if( frames[1]->i_type == X264_TYPE_AUTO ) //如果是场景切换，则设为I帧
frames[1]->i_type = X264_TYPE_I;
return;
}
//省略
}
下面是scenecut的源码：
static int scenecut( x264_t *h, x264_mb_analysis_t *a, x264_frame_t **frames, int p0, int p1, int real_scenecut, int num_frames, int i_max_search )
{
/* Only do analysis during a normal scenecut check. */
if( real_scenecut && h->param.i_bframe )
{
int origmaxp1 = p0 + 1;
/* Look ahead to avoid coding short flashes as scenecuts. */
if( h->param.i_bframe_adaptive == X264_B_ADAPT_TRELLIS )
/* Don't analyse any more frames than the trellis would have covered. */
origmaxp1 += h->param.i_bframe;
else
origmaxp1++;
int maxp1 = X264_MIN( origmaxp1, num_frames );
/* Where A and B are scenes: AAAAAABBBAAAAAA
* If BBB is shorter than (maxp1-p0), it is detected as a flash
* and not considered a scenecut. */
for( int curp1 = p1; curp1 <= maxp1; curp1++ )
if( !scenecut_internal( h, a, frames, p0, curp1, 0 ) )
/* Any frame in between p0 and cur_p1 cannot be a real scenecut. */
for( int i = curp1; i > p0; i-- )
frames[i]->b_scenecut = 0;
/* Where A-F are scenes: AAAAABBCCDDEEFFFFFF
* If each of BB ... EE are shorter than (maxp1-p0), they are
* detected as flashes and not considered scenecuts.
* Instead, the first F frame becomes a scenecut.
* If the video ends before F, no frame becomes a scenecut. */
for( int curp0 = p0; curp0 <= maxp1; curp0++ )
if( origmaxp1 > i_max_search || (curp0 < maxp1 && scenecut_internal( h, a, frames, curp0, maxp1, 0 )) )
/* If cur_p0 is the p0 of a scenecut, it cannot be the p1 of a scenecut. */
frames[curp0]->b_scenecut = 0;
}
/* Ignore frames that are part of a flash, i.e. cannot be real scenecuts. */
if( !frames[p1]->b_scenecut )
return 0;
return scenecut_internal( h, a, frames, p0, p1, real_scenecut );
}
如源码所示，scenecut主要是比较p0和p1两帧之间的差异，当差异大于某个阈值时，则判断为场景切换，调用它的函数会将p1设为I帧或IDR帧。主要的判断逻辑是在scenecut_internal函数中，我们稍后分析，现在我们先看下scenecut中的两个for循环。
第一个for循环：当存在A和B两种场景，它们出现的顺序像：AAAAAABBBAAAAAA这样的，如果场景B的数目小于(maxp1-p0)，则认为这是一个flash（闪屏或抖动的意思吧...），不认为是场景切换。也就是说，如果一开始是A场景，然后突然来了几帧B场景，但后面又恢复到A场景，则不能认为是一个场景切换，需要把B场景的frames的b_scenecut清空。
第二个for循环：当出现AAAAABBCCDDEEFFFFFF这样的场景，如果BB ... EE这些场景的图像数目小于(maxp1-p0)，则把他们当做是flashes而不能认为是scenecuts。而第一个F则会被当成一个场景切换（因为后面一直是F了，没有恢复到A或变为其他场景）。但如果视频流在F之前结束，那么没有图像被认为是场景切换。
下面我们来看下scenecut_internal源码如何判断场景切换：
static int scenecut_internal( x264_t *h, x264_mb_analysis_t *a, x264_frame_t **frames, int p0, int p1, int real_scenecut )
{
x264_frame_t *frame = frames[p1];
/* Don't do scenecuts on the right view of a frame-packed video. */
if( real_scenecut && h->param.i_frame_packing == 5 && (frame->i_frame&1) )
return 0;
x264_slicetype_frame_cost( h, a, frames, p0, p1, p1 );
int icost = frame->i_cost_est[0][0]; //帧内编码损耗
int pcost = frame->i_cost_est[p1-p0][0]; //帧间编码损耗
float f_bias;
int i_gop_size = frame->i_frame - h->lookahead->i_last_keyframe;
float f_thresh_max = h->param.i_scenecut_threshold / 100.0;
/* magic numbers pulled out of thin air */
float f_thresh_min = f_thresh_max * 0.25;
int res;
//通过i_scenecut_threshold计算出f_bias系数
if( h->param.i_keyint_min == h->param.i_keyint_max )
f_thresh_min = f_thresh_max;
if( i_gop_size <= h->param.i_keyint_min / 4 || h->param.b_intra_refresh )
f_bias = f_thresh_min / 4;
else if( i_gop_size <= h->param.i_keyint_min )
f_bias = f_thresh_min * i_gop_size / h->param.i_keyint_min;
else
{
f_bias = f_thresh_min
+ ( f_thresh_max - f_thresh_min )
* ( i_gop_size - h->param.i_keyint_min )
/ ( h->param.i_keyint_max - h->param.i_keyint_min );
}
res = pcost >= (1.0 - f_bias) * icost; //帧间损耗大于帧内损耗的某个比例，因为帧间损耗越大，说明两帧之间差异越大
if( res && real_scenecut )
{
int imb = frame->i_intra_mbs[p1-p0];
int pmb = NUM_MBS - imb;
x264_log( h, X264_LOG_DEBUG, "scene cut at %d Icost:%d Pcost:%d ratio:%.4f bias:%.4f gop:%d (imb:%d pmb:%d)\n",
frame->i_frame,
icost, pcost, 1. - (double)pcost / icost,
f_bias, i_gop_size, imb, pmb );
}
return res;
}
具体看注释。
网友评论