[x265] [PATCH] slicetype: fill in missing detail from slicetypeDecide()
Steve Borho
steve at borho.org
Thu Sep 19 18:38:59 CEST 2013
On Thu, Sep 19, 2013 at 4:33 AM, Deepthi Nandakumar <
deepthi at multicorewareinc.com> wrote:
>
>
> On Thu, Sep 19, 2013 at 3:35 AM, Steve Borho <steve at borho.org> wrote:
>
>> # HG changeset patch
>> # User Steve Borho <steve at borho.org>
>> # Date 1378869992 18000
>> # Tue Sep 10 22:26:32 2013 -0500
>> # Node ID a5810b7397ddec0bbec0e885bc5e7fd9bd5a7040
>> # Parent 5bab261d0dd7ac68c4af3837853d48d56017d4d6
>> slicetype: fill in missing detail from slicetypeDecide()
>>
>> This moves slicetypeDecide() to make the file more readable. It also adds
>> stubs
>> for features that we do not support yet including weightp, B-pyramid,
>> intra
>> refresh, user-supplied slice types, etc
>>
>> diff -r 5bab261d0dd7 -r a5810b7397dd source/common/lowres.cpp
>> --- a/source/common/lowres.cpp Wed Sep 18 16:13:33 2013 -0500
>> +++ b/source/common/lowres.cpp Tue Sep 10 22:26:32 2013 -0500
>> @@ -108,11 +108,13 @@
>> // (re) initialize lowres state
>> void Lowres::init(TComPicYuv *orig, int bframes)
>> {
>> - scenecut = 1;
>> + scenecut = true;
>> bIntraCalculated = false;
>> + bLastMiniGopBFrame = false;
>> + keyframe = false; // Not a keyframe unless identified by lookahead
>> + sliceType = X265_TYPE_AUTO;
>> + leadingBframes = 0;
>> memset(costEst, -1, sizeof(costEst));
>> - sliceType = X265_TYPE_AUTO;
>> - keyframe = 0; //Not a keyframe unless identified by lookahead
>> for (int y = 0; y < bframes + 2; y++)
>> {
>> for (int x = 0; x < bframes + 2; x++)
>> diff -r 5bab261d0dd7 -r a5810b7397dd source/common/lowres.h
>> --- a/source/common/lowres.h Wed Sep 18 16:13:33 2013 -0500
>> +++ b/source/common/lowres.h Tue Sep 10 22:26:32 2013 -0500
>> @@ -44,14 +44,17 @@
>> {
>> /* lowres buffers, sizes and strides */
>> pixel *buffer[4];
>> - int width; // width of lowres frame in pixels
>> - int lines; // height of lowres frame in pixel lines
>> + int width; // width of lowres frame in pixels
>> + int lines; // height of lowres frame in pixel lines
>> +
>> + int frameNum; // Presentation frame number
>> + int sliceType; // Slice type decided by lookahead
>> + int leadingBframes; // number of leading B frames for P or I
>> +
>> bool bIntraCalculated;
>> - int frameNum; // Presentation frame number
>> - int scenecut; // Set to zero if the frame cannot possibly be
>> part of a real scenecut.
>> -
>> - int sliceType; // Slice type decided by lookahead
>> - int keyframe;
>> + bool scenecut; // Set to zero if the frame cannot possibly be
>> part of a real scenecut.
>> + bool keyframe;
>> + bool bLastMiniGopBFrame;
>>
>> /* lookahead output data */
>> int costEst[X265_BFRAME_MAX + 2][X265_BFRAME_MAX + 2];
>> diff -r 5bab261d0dd7 -r a5810b7397dd source/encoder/slicetype.cpp
>> --- a/source/encoder/slicetype.cpp Wed Sep 18 16:13:33 2013 -0500
>> +++ b/source/encoder/slicetype.cpp Tue Sep 10 22:26:32 2013 -0500
>> @@ -66,6 +66,8 @@
>> {
>> this->cfg = _cfg;
>> numDecided = 0;
>> + lastKeyframe = -cfg->param.keyframeMax;
>>
>
> _cfg typo here.
>
actually, that is correct, it is initialized to negative keyframeMax, to
force frame 0 to I
see
http://git.videolan.org/?p=x264.git;a=blob;f=encoder/encoder.c;h=78ad56ae3cd96eebe596c79367d428c2e9dbd7bd;hb=9e941d1fabb2de4b15f169057a07dc395307d2ce#l1439
>
>
>> + lastNonB = NULL;
>> predictions = (pixel*)X265_MALLOC(pixel, 35 * 8 * 8);
>> me.setQP(X265_LOOKAHEAD_QP);
>> me.setSearchMethod(X265_HEX_SEARCH);
>> @@ -115,118 +117,6 @@
>> slicetypeDecide();
>> }
>>
>> -void Lookahead::slicetypeDecide()
>> -{
>> - if (numDecided == 0)
>> - {
>> - // Special case for POC 0, send directly to output queue as I
>> slice
>> - TComPic *pic = inputQueue.popFront();
>> - pic->m_lowres.sliceType = X265_TYPE_I;
>> - outputQueue.pushBack(pic);
>> - numDecided++;
>> - lastKeyframe = 0;
>> - pic->m_lowres.keyframe = 1;
>> - frames[0] = &(pic->m_lowres);
>> - return;
>> - }
>> - else if (cfg->param.bFrameAdaptive && cfg->param.lookaheadDepth &&
>> cfg->param.bframes)
>> - {
>> - slicetypeAnalyse(false);
>> -
>> - int dframes;
>> - TComPic* picsAnalysed[X265_LOOKAHEAD_MAX]; //Used for sorting
>> the pics into encode order
>> - int idx = 1;
>> -
>> - for (dframes = 0; (frames[dframes + 1] != NULL) &&
>> (frames[dframes + 1]->sliceType != X265_TYPE_AUTO); dframes++)
>> - {
>> - if (frames[dframes + 1]->sliceType == X265_TYPE_I)
>> - {
>> - frames[dframes + 1]->keyframe = 1;
>> - lastKeyframe = frames[dframes]->frameNum;
>> - if (cfg->param.decodingRefreshType == 2 && dframes > 0)
>> //If an IDR frame following a B
>> - {
>> - frames[dframes]->sliceType = X265_TYPE_P;
>> - dframes--;
>> - }
>> - }
>> - if (!IS_X265_TYPE_B(frames[dframes + 1]->sliceType))
>> - {
>> - dframes++;
>> - break;
>> - }
>> - }
>> -
>> - TComPic *pic = NULL;
>> - for (int i = 1; i <= dframes && !inputQueue.empty(); i++)
>> - {
>> - pic = inputQueue.popFront();
>> - picsAnalysed[idx++] = pic;
>> - }
>> -
>> - picsAnalysed[0] = pic; //Move the P-frame following B-frames to
>> the beginning
>> -
>> - //Push pictures in encode order
>> - for (int i = 0; i < dframes; i++)
>> - {
>> - outputQueue.pushBack(picsAnalysed[i]);
>> - }
>> -
>> - if (pic)
>> - frames[0] = &(pic->m_lowres); // last nonb
>> - return;
>> - }
>> -
>> - // Fixed GOP structures for when B-Adapt and/or lookahead are
>> disabled
>> - if (cfg->param.keyframeMax == 1)
>> - {
>> - TComPic *pic = inputQueue.popFront();
>> -
>> - pic->m_lowres.sliceType = X265_TYPE_I;
>> - pic->m_lowres.keyframe = 1;
>> - outputQueue.pushBack(pic);
>> - numDecided++;
>> - }
>> - else if (cfg->param.bframes == 0 || inputQueue.size() == 1)
>> - {
>> - TComPic *pic = inputQueue.popFront();
>> -
>> - bool forceIntra = (pic->getPOC() % cfg->param.keyframeMax) == 0;
>> - pic->m_lowres.sliceType = forceIntra ? X265_TYPE_I : X265_TYPE_P;
>> - pic->m_lowres.keyframe = forceIntra ? 1 : 0;
>> - outputQueue.pushBack(pic);
>> - numDecided++;
>> - }
>> - else
>> - {
>> - TComPic *picB = inputQueue.popFront();
>> - TComPic *picP = inputQueue.popFront();
>> -
>> - bool forceIntra = (picP->getPOC() % cfg->param.keyframeMax) == 0
>> || (picB->getPOC() % cfg->param.keyframeMax) == 0;
>> - if (forceIntra)
>> - {
>> - picB->m_lowres.sliceType = (picB->getPOC() %
>> cfg->param.keyframeMax) ? X265_TYPE_P : X265_TYPE_I;
>> - picB->m_lowres.keyframe = (picB->getPOC() %
>> cfg->param.keyframeMax) ? 0 : 1;
>> - outputQueue.pushBack(picB);
>> - numDecided++;
>> -
>> - picP->m_lowres.sliceType = (picP->getPOC() %
>> cfg->param.keyframeMax) ? X265_TYPE_P : X265_TYPE_I;
>> - picP->m_lowres.keyframe = (picP->getPOC() %
>> cfg->param.keyframeMax) ? 0 : 1;
>> - outputQueue.pushBack(picP);
>> - numDecided++;
>> - }
>> - else
>> - {
>> - picP->m_lowres.sliceType = X265_TYPE_P;
>> - outputQueue.pushBack(picP);
>> - numDecided++;
>> -
>> - picB->m_lowres.sliceType = X265_TYPE_B;
>> - outputQueue.pushBack(picB);
>> - numDecided++;
>> - }
>> - }
>> -}
>> -
>> // Called by RateControl to get the estimated SATD cost for a given
>> picture.
>> // It assumes dpb->prepareEncode() has already been called for the
>> picture and
>> // all the references are established
>> @@ -482,6 +372,219 @@
>> fenc->lowresCosts[b - p0][p1 - b][cuXY] = (uint16_t)(X265_MIN(bcost,
>> LOWRES_COST_MASK) | (listused << LOWRES_COST_SHIFT));
>> }
>>
>> +void Lookahead::slicetypeDecide()
>> +{
>> + if (cfg->param.bFrameAdaptive && cfg->param.lookaheadDepth &&
>> cfg->param.bframes)
>> + {
>> + slicetypeAnalyse(false);
>> +
>> + TComPic *list[X265_LOOKAHEAD_MAX];
>> + TComList<TComPic*>::iterator iterPic = inputQueue.begin();
>> + int j;
>> + for (j = 0; j < cfg->param.bframes + 2 && iterPic !=
>> inputQueue.end(); iterPic++)
>> + list[j++] = (*iterPic);
>> + list[j] = NULL;
>> +
>> + int bframes, brefs;
>> + for (bframes = 0, brefs = 0 ;; bframes++)
>> + {
>> + Lowres& frm = list[bframes]->m_lowres;
>> +
>> + if (frm.sliceType == X265_TYPE_BREF
>> + /* && h->param.i_bframe_pyramid < X264_B_PYRAMID_NORMAL
>> && brefs == h->param.i_bframe_pyramid*/)
>> + {
>> + frm.sliceType = X265_TYPE_B;
>> + x265_log(&cfg->param, X265_LOG_WARNING, "B-ref is not
>> yet supported\n");
>> + }
>> + /* pyramid with multiple B-refs needs a big enough dpb that
>> the preceding P-frame stays available.
>> + smaller dpb could be supported by smart enough use of
>> mmco, but it's easier just to forbid it.
>> + else if (frm.sliceType == X265_TYPE_BREF &&
>> cfg->param.i_bframe_pyramid == X265_B_PYRAMID_NORMAL &&
>> + brefs && cfg->param.i_frame_reference <= (brefs+3))
>> + {
>> + frm.sliceType = X265_TYPE_B;
>> + x265_log(&cfg->param, X265_LOG_WARNING, "B-ref at frame
>> %d incompatible with B-pyramid %s and %d reference frames\n",
>> + frm.sliceType,
>> x264_b_pyramid_names[h->param.i_bframe_pyramid],
>> h->param.i_frame_reference);
>> + } */
>> +
>> + if (frm.sliceType == X265_TYPE_KEYFRAME)
>> + frm.sliceType = cfg->param.bOpenGOP ? X265_TYPE_I :
>> X265_TYPE_IDR;
>> + if ((/* !cfg->param.intraRefresh || */ frm.frameNum == 0) &&
>> frm.frameNum - lastKeyframe >= cfg->param.keyframeMax)
>> + {
>> + if (frm.sliceType == X265_TYPE_AUTO || frm.sliceType ==
>> X265_TYPE_I)
>> + frm.sliceType = cfg->param.bOpenGOP && lastKeyframe
>> >= 0 ? X265_TYPE_I : X265_TYPE_IDR;
>> + bool warn = frm.sliceType != X265_TYPE_IDR;
>> + if (warn && cfg->param.bOpenGOP)
>> + warn &= frm.sliceType != X265_TYPE_I;
>> + if (warn)
>> + {
>> + x265_log(&cfg->param, X265_LOG_WARNING, "specified
>> frame type (%d) at %d is not compatible with keyframe interval\n",
>> frm.sliceType, frm.frameNum);
>> + frm.sliceType = cfg->param.bOpenGOP && lastKeyframe
>> >= 0 ? X265_TYPE_I : X265_TYPE_IDR;
>> + }
>> + }
>> + if (frm.sliceType == X265_TYPE_I && frm.sliceType -
>> lastKeyframe >= cfg->param.keyframeMin)
>> + {
>> + if (cfg->param.bOpenGOP)
>> + {
>> + lastKeyframe = frm.frameNum; // Use display order
>> + frm.keyframe = true;
>> + }
>> + else
>> + frm.sliceType = X265_TYPE_IDR;
>> + }
>> + if (frm.sliceType == X265_TYPE_IDR)
>> + {
>> + /* Close GOP */
>> + lastKeyframe = frm.frameNum;
>> + frm.keyframe = true;
>> + if (bframes > 0)
>> + {
>> + frames[bframes]->sliceType = X265_TYPE_P;
>> + bframes--;
>> + }
>> + }
>> +
>> + if (bframes == cfg->param.bframes || !list[bframes+1])
>> + {
>> + if (IS_X265_TYPE_B(frm.sliceType))
>> + x265_log(&cfg->param, X265_LOG_WARNING, "specified
>> frame type is not compatible with max B-frames\n");
>> + if (frm.sliceType == X265_TYPE_AUTO ||
>> IS_X265_TYPE_B(frm.sliceType))
>> + frm.sliceType = X265_TYPE_P;
>> + }
>> + if (frm.sliceType == X265_TYPE_BREF)
>> + brefs++;
>> + if (frm.sliceType == X265_TYPE_AUTO)
>> + frm.sliceType = X265_TYPE_B;
>> + else if (!IS_X265_TYPE_B(frm.sliceType))
>> + break;
>> + }
>> + if (bframes)
>> + list[bframes-1]->m_lowres.bLastMiniGopBFrame = true;
>> + list[bframes]->m_lowres.leadingBframes = bframes;
>> + lastNonB = &list[bframes]->m_lowres;
>> +
>> + /* insert a bref into the sequence
>> + if (h->param.i_bframe_pyramid && bframes > 1 && !brefs)
>> + {
>> + h->lookahead->next.list[bframes/2]->i_type = X264_TYPE_BREF;
>> + brefs++;
>> + } */
>> +
>> + /* calculate the frame costs ahead of time for
>> x264_rc_analyse_slice while we still have lowres */
>> + if (cfg->param.rc.rateControlMode != X265_RC_CQP)
>> + {
>> + int p0, p1, b;
>> + p1 = b = bframes + 1;
>> +
>> + frames[0] = lastNonB;
>> + if (IS_X265_TYPE_I(frames[bframes+1]->sliceType))
>> + p0 = bframes + 1;
>> + else // P
>> + p0 = 0;
>> +
>> + estimateFrameCost(p0, p1, b, 0);
>> +
>> + /*
>> + if ((p0 != p1 || bframes) && cfg->param.rc.i_vbv_buffer_size)
>> + {
>> + // We need the intra costs for row SATDs
>> + estimateFrameCost(b, b, b, 0);
>> +
>> + // We need B-frame costs for row SATDs
>> + p0 = 0;
>> + for (b = 1; b <= bframes; b++)
>> + {
>> + if (frames[b]->i_type == X265_TYPE_B)
>> + for (p1 = b; frames[p1]->sliceType ==
>> X265_TYPE_B;)
>> + p1++;
>> + else
>> + p1 = bframes + 1;
>> + estimateFrameCost( p0, p1, b, 0 );
>> + if (frames[b]->sliceType == X265_TYPE_BREF)
>> + p0 = b;
>> + }
>> + } */
>> + }
>> + /* Analyse for weighted P frames
>> + if (!h->param.rc.b_stat_read &&
>> h->lookahead->next.list[bframes]->i_type == X264_TYPE_P
>> + && h->param.analyse.i_weighted_pred >= X264_WEIGHTP_SIMPLE)
>> + {
>> + x265_emms();
>> + x264_weights_analyse(h, h->lookahead->next.list[bframes],
>> h->lookahead->last_nonb, 0);
>> + }*/
>> +
>> + /* add to output queue in encode order */
>> + outputQueue.pushBack(list[bframes]);
>> + inputQueue.popFront();
>> + for (int i = 0; i < bframes; i++)
>> + {
>> + outputQueue.pushBack(list[i]);
>> + inputQueue.popFront();
>> + }
>> + return;
>> + }
>> +
>> + // Fixed GOP structures for when B-Adapt and/or lookahead are
>> disabled
>> + if (numDecided == 0)
>> + {
>> + // Special case for POC 0, send directly to output queue as I
>> slice
>> + TComPic *pic = inputQueue.popFront();
>> + pic->m_lowres.sliceType = X265_TYPE_I;
>> + outputQueue.pushBack(pic);
>> + numDecided++;
>> + lastKeyframe = 0;
>> + pic->m_lowres.keyframe = 1;
>> + return;
>> + }
>> + else if (cfg->param.keyframeMax == 1)
>> + {
>> + TComPic *pic = inputQueue.popFront();
>> +
>> + pic->m_lowres.sliceType = X265_TYPE_I;
>> + pic->m_lowres.keyframe = 1;
>> + outputQueue.pushBack(pic);
>> + numDecided++;
>> + }
>> + else if (cfg->param.bframes == 0 || inputQueue.size() == 1)
>> + {
>> + TComPic *pic = inputQueue.popFront();
>> +
>> + bool forceIntra = (pic->getPOC() % cfg->param.keyframeMax) == 0;
>> + pic->m_lowres.sliceType = forceIntra ? X265_TYPE_I : X265_TYPE_P;
>> + pic->m_lowres.keyframe = forceIntra ? 1 : 0;
>> + outputQueue.pushBack(pic);
>> + numDecided++;
>> + }
>> + else
>> + {
>> + TComPic *picB = inputQueue.popFront();
>> + TComPic *picP = inputQueue.popFront();
>> +
>> + bool forceIntra = (picP->getPOC() % cfg->param.keyframeMax) == 0
>> || (picB->getPOC() % cfg->param.keyframeMax) == 0;
>> + if (forceIntra)
>> + {
>> + picB->m_lowres.sliceType = (picB->getPOC() %
>> cfg->param.keyframeMax) ? X265_TYPE_P : X265_TYPE_I;
>> + picB->m_lowres.keyframe = (picB->getPOC() %
>> cfg->param.keyframeMax) ? 0 : 1;
>> + outputQueue.pushBack(picB);
>> + numDecided++;
>> +
>> + picP->m_lowres.sliceType = (picP->getPOC() %
>> cfg->param.keyframeMax) ? X265_TYPE_P : X265_TYPE_I;
>> + picP->m_lowres.keyframe = (picP->getPOC() %
>> cfg->param.keyframeMax) ? 0 : 1;
>> + outputQueue.pushBack(picP);
>> + numDecided++;
>> + }
>> + else
>> + {
>> + picP->m_lowres.sliceType = X265_TYPE_P;
>> + outputQueue.pushBack(picP);
>> + numDecided++;
>> +
>> + picB->m_lowres.sliceType = X265_TYPE_B;
>> + outputQueue.pushBack(picB);
>> + numDecided++;
>> + }
>> + }
>> +}
>> +
>> void Lookahead::slicetypeAnalyse(bool bKeyframe)
>> {
>> int num_frames, origNumFrames, keyint_limit, framecnt;
>> @@ -489,8 +592,10 @@
>> int cuCount = NUM_CUS;
>> int cost1p0, cost2p0, cost1b1, cost2p1;
>> int reset_start;
>> - int vbv_lookahead = 0;
>>
>> + if (!lastNonB)
>> + return;
>> + frames[0] = lastNonB;
>> TComList<TComPic*>::iterator iterPic = inputQueue.begin();
>> for (framecnt = 0; (framecnt < maxSearch) && (framecnt <
>> (int)inputQueue.size()) && (*iterPic)->m_lowres.sliceType ==
>> X265_TYPE_AUTO; framecnt++)
>> {
>> @@ -499,8 +604,7 @@
>>
>> if (!framecnt)
>> {
>> - frames[1] = &((*iterPic)->m_lowres);
>> - frames[2] = NULL;
>> + // TODO: mb-tree
>> return;
>> }
>>
>> @@ -509,14 +613,7 @@
>> keyint_limit = cfg->param.keyframeMax - frames[1]->frameNum +
>> lastKeyframe;
>> origNumFrames = num_frames = X265_MIN(framecnt, keyint_limit);
>>
>> - /* This is important psy-wise: if we have a non-scenecut keyframe,
>> - * there will be significant visual artifacts if the frames just
>> before
>> - * go down in quality due to being referenced less, despite it being
>> - * more RD-optimal. */
>> -
>> - if (vbv_lookahead)
>> - num_frames = framecnt;
>> - else if (num_frames < framecnt)
>> + if (cfg->param.bOpenGOP && num_frames < framecnt)
>> num_frames++;
>> else if (num_frames == 0)
>> {
>> @@ -641,7 +738,7 @@
>> }
>>
>> // TODO if rc.b_mb_tree Enabled the need to call
>> x264_macroblock_tree currently Ignored the call
>> -
>> + // if (!cfg->param.bIntraRefresh)
>> for (int j = keyint_limit + 1; j <= num_frames; j +=
>> cfg->param.keyframeMax)
>> {
>> frames[j]->sliceType = X265_TYPE_I;
>> diff -r 5bab261d0dd7 -r a5810b7397dd source/encoder/slicetype.h
>> --- a/source/encoder/slicetype.h Wed Sep 18 16:13:33 2013 -0500
>> +++ b/source/encoder/slicetype.h Tue Sep 10 22:26:32 2013 -0500
>> @@ -44,6 +44,7 @@
>> TEncCfg *cfg;
>> pixel *predictions; // buffer for 35 intra predictions
>> Lowres *frames[X265_LOOKAHEAD_MAX];
>> + Lowres *lastNonB;
>> int merange;
>> int numDecided;
>> int lastKeyframe;
>>
>> _______________________________________________
>> x265-devel mailing list
>> x265-devel at videolan.org
>> https://mailman.videolan.org/listinfo/x265-devel
>>
>>
>
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
>
--
Steve Borho
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20130919/f68b3ed5/attachment-0001.html>
More information about the x265-devel
mailing list