From patchwork Sun Aug 7 14:13:40 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andreas Rheinhardt X-Patchwork-Id: 37172 Delivered-To: ffmpegpatchwork2@gmail.com Received: by 2002:a05:6a20:92a9:b0:8d:7f3b:94fc with SMTP id q41csp1319646pzg; Sun, 7 Aug 2022 07:14:17 -0700 (PDT) X-Google-Smtp-Source: AA6agR6fKiyJDLEVJGt6b17aRdBZoD8Z/gKSPxd9kdD2yWSZsAmeAmXF1f9mLULcnjZFV8ig9raC X-Received: by 2002:a17:907:1c87:b0:6f0:29ea:cc01 with SMTP id nb7-20020a1709071c8700b006f029eacc01mr11005841ejc.671.1659881657286; Sun, 07 Aug 2022 07:14:17 -0700 (PDT) Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id d7-20020a170906640700b0072f7debe7d3si6717896ejm.352.2022.08.07.07.14.13; Sun, 07 Aug 2022 07:14:17 -0700 (PDT) Received-SPF: pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) client-ip=79.124.17.100; Authentication-Results: mx.google.com; dkim=neutral (body hash did not verify) header.i=@outlook.com header.s=selector1 header.b=V3PpdUD6; arc=fail (body hash mismatch); spf=pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) smtp.mailfrom=ffmpeg-devel-bounces@ffmpeg.org; dmarc=fail (p=NONE sp=QUARANTINE dis=NONE) header.from=outlook.com Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 4CD0568B658; Sun, 7 Aug 2022 17:14:09 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from EUR04-HE1-obe.outbound.protection.outlook.com (mail-oln040092073032.outbound.protection.outlook.com [40.92.73.32]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 1EE1468B1AE for ; Sun, 7 Aug 2022 17:14:03 +0300 (EEST) ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=Rt3JpOsp4s1JrgCbOBuynJvvT75rNljrbATupFtrVk0d6QRpAl5JKs+CGEyyDFSwQRZVjaFdeRRLc5oKx/GtI+KPXYWXgpVrwB69BL8btC98gdLbw0u+OcMNkOzVk+SOD+hYdwfFYnb4hh2LUqP69F1mgATElC6wF3Xqd5eQshGaNqn22tp2FE7d8n57ZGxqV3gRoZLVuQXJXAZnJ8aqci7uS22Ac5sJTXsHBTo0qbvXTR78prSDSSlmkCSopNZP3tqAyIx8AsymVuloM1Zt5SpXraLut8i0Jk5FZF8ZTiactrgzqNGnXw1B1gkEPSNUlXM5QzCqm5TOeORUIDFujA== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=4G7ybowHkLmLYIyYQ21nYGPKeHxalyXJgbYLZgsWeuQ=; b=mnOxOE+baLyRB4Be+ziIaPTzJROiFbj6EMK3CVciOd7D0Oil/FhqFwLwjv7+4thHm62ach8MFfrH+G2lT0aa4SqsV7y+YZzOGSeyJCd5+JJ9niKhr8t+2AhdFti+sAT+Mgq4W8eQ4RcFpTJy36BjgdGmdxjFPte/UtCiNQjFRxqO5f+w9GwPQE6l3rZDv5ycLp0CgxOyzTO9XQsEXAV+M9qogWEw4Nm5gBYAF2wlRVI+a59MoGZEQZ1s3lJ3To43fAWeEV9CBpuKE2xyUuXIAVPmKk9Lic+2GIS4iUq21tyfMtXlDbFTtrr/qZUsdJeT4rxKl3KhIkgupps+dYuJzw== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=none; dmarc=none; dkim=none; arc=none DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=outlook.com; s=selector1; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=4G7ybowHkLmLYIyYQ21nYGPKeHxalyXJgbYLZgsWeuQ=; b=V3PpdUD6k/3yU7m4gCkKpT9r73u25xpC6WW8BQtFh6bNvPaHpx0UOdJLMV6r9GGGEVqLIIl0ajbs4KMHuJgqAQZV0m+IbMw/oRZYuG4xt+0a+GlF57niuGohe84GoMPih/RuObZYAcxFqhy210Mo5JhU6aXlqji1AS3zo982oFDwFdzTbvGr73waVksKoQOwJw3pBE2b2BH0HH06VOj7LPIbjnB1cg1/xk6OsMC3HaidqxbOc0dUzHOR2mkme+GKthrONMqo5/25e3B6b9mIrFOnJkzJWMh75mJ0K8KcQ/6iOc5J8Dc5ROye8iGJWKt4Rl9stEwbJ3k8ro7a0FPYTg== Received: from DB6PR0101MB2214.eurprd01.prod.exchangelabs.com (2603:10a6:4:42::27) by DB7PR01MB4380.eurprd01.prod.exchangelabs.com (2603:10a6:5:36::32) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.5504.17; Sun, 7 Aug 2022 14:14:00 +0000 Received: from DB6PR0101MB2214.eurprd01.prod.exchangelabs.com ([fe80::210e:b627:bcc9:8c46]) by DB6PR0101MB2214.eurprd01.prod.exchangelabs.com ([fe80::210e:b627:bcc9:8c46%11]) with mapi id 15.20.5504.019; Sun, 7 Aug 2022 14:14:00 +0000 From: Andreas Rheinhardt To: ffmpeg-devel@ffmpeg.org Date: Sun, 7 Aug 2022 16:13:40 +0200 Message-ID: X-Mailer: git-send-email 2.34.1 X-TMN: [Nq0C1gMZ5xPjg+uvyPrM9MuZBuxJ8Tf+] X-ClientProxiedBy: AS8PR04CA0136.eurprd04.prod.outlook.com (2603:10a6:20b:127::21) To DB6PR0101MB2214.eurprd01.prod.exchangelabs.com (2603:10a6:4:42::27) X-Microsoft-Original-Message-ID: <20220807141340.1170873-1-andreas.rheinhardt@outlook.com> MIME-Version: 1.0 X-MS-Exchange-MessageSentRepresentingType: 1 X-MS-PublicTrafficType: Email X-MS-Office365-Filtering-Correlation-Id: 6cd6e24e-a758-47c8-139a-08da787f12d1 X-MS-Exchange-SLBlob-MailProps: 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 X-MS-TrafficTypeDiagnostic: DB7PR01MB4380:EE_ X-Microsoft-Antispam: BCL:0; X-Microsoft-Antispam-Message-Info: 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 X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1 X-MS-Exchange-AntiSpam-MessageData-0: 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 X-OriginatorOrg: outlook.com X-MS-Exchange-CrossTenant-Network-Message-Id: 6cd6e24e-a758-47c8-139a-08da787f12d1 X-MS-Exchange-CrossTenant-AuthSource: DB6PR0101MB2214.eurprd01.prod.exchangelabs.com X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-OriginalArrivalTime: 07 Aug 2022 14:14:00.2592 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-CrossTenant-Id: 84df9e7f-e9f6-40af-b435-aaaaaaaaaaaa X-MS-Exchange-CrossTenant-RMS-PersistedConsumerOrg: 00000000-0000-0000-0000-000000000000 X-MS-Exchange-Transport-CrossTenantHeadersStamped: DB7PR01MB4380 Subject: [FFmpeg-devel] [PATCH] avcodec/mpegpicture: Move mb_var, mc_mb_var and mb_mean to MpegEncCtx X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Cc: Andreas Rheinhardt Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" X-TUID: 0+LcucaX+zRd These tables are only used by encoders and only for the current picture; ergo they need not be put into the picture at all, but rather into the encoder's context. They also don't need to be refcounted, because there is only one owner. In contrast to this, the earlier code refcounts them which incurs unnecessary overhead. These references are not unreferenced in ff_mpeg_unref_picture() (they are kept in order to have something like a buffer pool), so that several buffers are kept at the same time, although only one is needed, thereby wasting memory. The code also propagates references to other pictures not part of the pictures array (namely the copy of the current/next/last picture in the MpegEncContext which get references of their own). These references are not unreferenced in ff_mpeg_unref_picture() (the buffers are probably kept in order to have something like a pool), yet if the current picture is a B-frame, it gets unreferenced at the end of ff_mpv_encode_picture() and its slot in the picture array will therefore be reused the next time; but the copy of the current picture also still has its references and therefore these buffers will be made duplicated in order to make them writable in the next call to ff_mpv_encode_picture(). This is of course unnecessary. Finally, ff_find_unused_picture() is supposed to just return any unused picture and the code is supposed to work with it; yet for the vsynth*-mpeg4-adap tests the result depends upon the content of these buffers; given that this patchset changes the content of these buffers (the initial content is now the state of these buffers after encoding the last frame; before this patch the buffers used came from the last picture that occupied the same slot in the picture array) their ref-files needed to be changed. This points to a bug somewhere (if one removes the initialization, one gets uninitialized reads in adaptive_quantization in ratecontrol.c). Signed-off-by: Andreas Rheinhardt --- The decoder-part of the vsynth*-h263-obmc tests also has problems with returning random pictures in ff_find_unused_picture(). Probably a bug in preview_obmc() or so. Anyway, next thing is to add proper buffer-pools for these buffers to remove the needs_realloc-hack. libavcodec/motion_est.c | 14 ++++++------ libavcodec/mpegpicture.c | 30 ++----------------------- libavcodec/mpegpicture.h | 9 -------- libavcodec/mpegvideo.h | 6 +++++ libavcodec/mpegvideo_enc.c | 17 ++++++++++---- libavcodec/ratecontrol.c | 7 +++--- libavcodec/svq1enc.c | 6 ++--- tests/ref/seek/vsynth_lena-mpeg4-adap | 28 +++++++++++------------ tests/ref/vsynth/vsynth1-mpeg4-adap | 8 +++---- tests/ref/vsynth/vsynth2-mpeg4-adap | 8 +++---- tests/ref/vsynth/vsynth3-mpeg4-adap | 8 +++---- tests/ref/vsynth/vsynth_lena-mpeg4-adap | 8 +++---- 12 files changed, 64 insertions(+), 85 deletions(-) diff --git a/libavcodec/motion_est.c b/libavcodec/motion_est.c index 29ab41dc8c..d17ffe42b4 100644 --- a/libavcodec/motion_est.c +++ b/libavcodec/motion_est.c @@ -895,7 +895,6 @@ void ff_estimate_p_frame_motion(MpegEncContext * s, int P[10][2]; const int shift= 1+s->quarter_sample; int mb_type=0; - Picture * const pic= &s->current_picture; init_ref(c, s->new_picture->data, s->last_picture.f->data, NULL, 16*mb_x, 16*mb_y, 0); @@ -917,8 +916,8 @@ void ff_estimate_p_frame_motion(MpegEncContext * s, varc = s->mpvencdsp.pix_norm1(pix, s->linesize) - (((unsigned) sum * sum) >> 8) + 500; - pic->mb_mean[s->mb_stride * mb_y + mb_x] = (sum+128)>>8; - pic->mb_var [s->mb_stride * mb_y + mb_x] = (varc+128)>>8; + s->mb_mean[s->mb_stride * mb_y + mb_x] = (sum+128)>>8; + s->mb_var [s->mb_stride * mb_y + mb_x] = (varc+128)>>8; c->mb_var_sum_temp += (varc+128)>>8; if (s->motion_est != FF_ME_ZERO) { @@ -965,7 +964,7 @@ void ff_estimate_p_frame_motion(MpegEncContext * s, vard = s->mecc.sse[0](NULL, pix, ppix, s->linesize, 16); - pic->mc_mb_var[s->mb_stride * mb_y + mb_x] = (vard+128)>>8; + s->mc_mb_var[s->mb_stride * mb_y + mb_x] = (vard+128)>>8; c->mc_mb_var_sum_temp += (vard+128)>>8; if (c->avctx->mb_decision > FF_MB_DECISION_SIMPLE) { @@ -1509,7 +1508,7 @@ void ff_estimate_b_frame_motion(MpegEncContext * s, score= ((unsigned)(score*score + 128*256))>>16; c->mc_mb_var_sum_temp += score; - s->current_picture.mc_mb_var[mb_y*s->mb_stride + mb_x] = score; //FIXME use SSE + s->mc_mb_var[mb_y*s->mb_stride + mb_x] = score; //FIXME use SSE s->mb_type[mb_y*s->mb_stride + mb_x]= CANDIDATE_MB_TYPE_DIRECT0; return; @@ -1574,7 +1573,7 @@ void ff_estimate_b_frame_motion(MpegEncContext * s, score= ((unsigned)(score*score + 128*256))>>16; c->mc_mb_var_sum_temp += score; - s->current_picture.mc_mb_var[mb_y*s->mb_stride + mb_x] = score; //FIXME use SSE + s->mc_mb_var[mb_y*s->mb_stride + mb_x] = score; //FIXME use SSE } if(c->avctx->mb_decision > FF_MB_DECISION_SIMPLE){ @@ -1629,7 +1628,8 @@ int ff_get_best_fcode(MpegEncContext * s, const int16_t (*mv_table)[2], int type continue; for(j=0; jpict_type==AV_PICTURE_TYPE_B || s->current_picture.mc_mb_var[xy] < s->current_picture.mb_var[xy]) + if (s->pict_type == AV_PICTURE_TYPE_B || + s->mc_mb_var[xy] < s->mb_var[xy]) score[j]-= 170; } } diff --git a/libavcodec/mpegpicture.c b/libavcodec/mpegpicture.c index aaa1df0bd8..f32f8d061b 100644 --- a/libavcodec/mpegpicture.c +++ b/libavcodec/mpegpicture.c @@ -37,9 +37,6 @@ static void av_noinline free_picture_tables(Picture *pic) pic->alloc_mb_width = pic->alloc_mb_height = 0; - av_buffer_unref(&pic->mb_var_buf); - av_buffer_unref(&pic->mc_mb_var_buf); - av_buffer_unref(&pic->mb_mean_buf); av_buffer_unref(&pic->mbskip_table_buf); av_buffer_unref(&pic->qscale_table_buf); av_buffer_unref(&pic->mb_type_buf); @@ -60,9 +57,6 @@ do {\ return ret;\ } while (0) - MAKE_WRITABLE(mb_var_buf); - MAKE_WRITABLE(mc_mb_var_buf); - MAKE_WRITABLE(mb_mean_buf); MAKE_WRITABLE(mbskip_table_buf); MAKE_WRITABLE(qscale_table_buf); MAKE_WRITABLE(mb_type_buf); @@ -218,14 +212,6 @@ static int alloc_picture_tables(AVCodecContext *avctx, Picture *pic, int encodin if (!pic->mbskip_table_buf || !pic->qscale_table_buf || !pic->mb_type_buf) return AVERROR(ENOMEM); - if (encoding) { - pic->mb_var_buf = av_buffer_allocz(mb_array_size * sizeof(int16_t)); - pic->mc_mb_var_buf = av_buffer_allocz(mb_array_size * sizeof(int16_t)); - pic->mb_mean_buf = av_buffer_allocz(mb_array_size); - if (!pic->mb_var_buf || !pic->mc_mb_var_buf || !pic->mb_mean_buf) - return AVERROR(ENOMEM); - } - if (out_format == FMT_H263 || encoding || (avctx->export_side_data & AV_CODEC_EXPORT_DATA_MVS)) { int mv_size = 2 * (b8_array_size + 4) * sizeof(int16_t); @@ -285,12 +271,6 @@ int ff_alloc_picture(AVCodecContext *avctx, Picture *pic, MotionEstContext *me, if (ret < 0) goto fail; - if (encoding) { - pic->mb_var = (uint16_t*)pic->mb_var_buf->data; - pic->mc_mb_var = (uint16_t*)pic->mc_mb_var_buf->data; - pic->mb_mean = pic->mb_mean_buf->data; - } - pic->mbskip_table = pic->mbskip_table_buf->data; pic->qscale_table = pic->qscale_table_buf->data + 2 * mb_stride + 1; pic->mb_type = (uint32_t*)pic->mb_type_buf->data + 2 * mb_stride + 1; @@ -316,7 +296,7 @@ fail: */ void ff_mpeg_unref_picture(AVCodecContext *avctx, Picture *pic) { - int off = offsetof(Picture, mb_mean) + sizeof(pic->mb_mean); + int off = offsetof(Picture, hwaccel_priv_buf) + sizeof(pic->hwaccel_priv_buf); pic->tf.f = pic->f; /* WM Image / Screen codecs allocate internal buffers with different @@ -340,10 +320,7 @@ int ff_update_picture_tables(Picture *dst, const Picture *src) { int i, ret; - ret = av_buffer_replace(&dst->mb_var_buf, src->mb_var_buf); - ret |= av_buffer_replace(&dst->mc_mb_var_buf, src->mc_mb_var_buf); - ret |= av_buffer_replace(&dst->mb_mean_buf, src->mb_mean_buf); - ret |= av_buffer_replace(&dst->mbskip_table_buf, src->mbskip_table_buf); + ret = av_buffer_replace(&dst->mbskip_table_buf, src->mbskip_table_buf); ret |= av_buffer_replace(&dst->qscale_table_buf, src->qscale_table_buf); ret |= av_buffer_replace(&dst->mb_type_buf, src->mb_type_buf); for (i = 0; i < 2; i++) { @@ -356,9 +333,6 @@ int ff_update_picture_tables(Picture *dst, const Picture *src) return ret; } - dst->mb_var = src->mb_var; - dst->mc_mb_var = src->mc_mb_var; - dst->mb_mean = src->mb_mean; dst->mbskip_table = src->mbskip_table; dst->qscale_table = src->qscale_table; dst->mb_type = src->mb_type; diff --git a/libavcodec/mpegpicture.h b/libavcodec/mpegpicture.h index 0671bbae8e..62589595d0 100644 --- a/libavcodec/mpegpicture.h +++ b/libavcodec/mpegpicture.h @@ -62,19 +62,10 @@ typedef struct Picture { AVBufferRef *ref_index_buf[2]; int8_t *ref_index[2]; - AVBufferRef *mb_var_buf; - uint16_t *mb_var; ///< Table for MB variances - - AVBufferRef *mc_mb_var_buf; - uint16_t *mc_mb_var; ///< Table for motion compensated MB variances - int alloc_mb_width; ///< mb_width used to allocate tables int alloc_mb_height; ///< mb_height used to allocate tables int alloc_mb_stride; ///< mb_stride used to allocate tables - AVBufferRef *mb_mean_buf; - uint8_t *mb_mean; ///< Table for MB luminance - AVBufferRef *hwaccel_priv_buf; void *hwaccel_picture_private; ///< Hardware accelerator private data diff --git a/libavcodec/mpegvideo.h b/libavcodec/mpegvideo.h index b1b2727535..47619c1976 100644 --- a/libavcodec/mpegvideo.h +++ b/libavcodec/mpegvideo.h @@ -235,6 +235,12 @@ typedef struct MpegEncContext { int16_t (*b_field_mv_table[2][2][2])[2];///< MV table (4MV per MB) interlaced B-frame encoding uint8_t (*p_field_select_table[2]); ///< Only the first element is allocated uint8_t (*b_field_select_table[2][2]); ///< Only the first element is allocated + + /* The following three arrays are encoder-only */ + uint16_t *mb_var; ///< Table for MB variances + uint16_t *mc_mb_var; ///< Table for motion compensated MB variances + uint8_t *mb_mean; ///< Table for MB luminance + int motion_est; ///< ME algorithm int me_penalty_compensation; int me_pre; ///< prepass for motion estimation diff --git a/libavcodec/mpegvideo_enc.c b/libavcodec/mpegvideo_enc.c index 0ca004ee39..a34fb66eac 100644 --- a/libavcodec/mpegvideo_enc.c +++ b/libavcodec/mpegvideo_enc.c @@ -313,6 +313,7 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx) MpegEncContext *s = avctx->priv_data; AVCPBProperties *cpb_props; int i, ret; + int mb_array_size; mpv_encode_defaults(s); @@ -823,6 +824,12 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx) !FF_ALLOCZ_TYPED_ARRAY(s->reordered_input_picture, MAX_PICTURE_COUNT)) return AVERROR(ENOMEM); + mb_array_size = s->mb_stride * s->mb_height; + if (!FF_ALLOCZ_TYPED_ARRAY(s->mc_mb_var, mb_array_size) || + !FF_ALLOCZ_TYPED_ARRAY(s->mb_var, mb_array_size) || + !(s->mb_mean = av_mallocz(mb_array_size))) + return AVERROR(ENOMEM); + if (s->noise_reduction) { if (!FF_ALLOCZ_TYPED_ARRAY(s->dct_offset, 2)) return AVERROR(ENOMEM); @@ -949,6 +956,9 @@ av_cold int ff_mpv_encode_end(AVCodecContext *avctx) av_freep(&s->input_picture); av_freep(&s->reordered_input_picture); av_freep(&s->dct_offset); + av_freep(&s->mb_var); + av_freep(&s->mc_mb_var); + av_freep(&s->mb_mean); return 0; } @@ -2235,8 +2245,7 @@ static av_always_inline void encode_mb_internal(MpegEncContext *s, } } /* pre quantization */ - if (s->current_picture.mc_mb_var[s->mb_stride * mb_y + mb_x] < - 2 * s->qscale * s->qscale) { + if (s->mc_mb_var[s->mb_stride * mb_y + mb_x] < 2 * s->qscale * s->qscale) { // FIXME optimize if (s->mecc.sad[1](NULL, ptr_y, dest_y, wrap_y, 8) < 20 * s->qscale) skip_dct[0] = 1; @@ -2648,8 +2657,8 @@ static int mb_var_thread(AVCodecContext *c, void *arg){ varc = (s->mpvencdsp.pix_norm1(pix, s->linesize) - (((unsigned) sum * sum) >> 8) + 500 + 128) >> 8; - s->current_picture.mb_var [s->mb_stride * mb_y + mb_x] = varc; - s->current_picture.mb_mean[s->mb_stride * mb_y + mb_x] = (sum+128)>>8; + s->mb_var [s->mb_stride * mb_y + mb_x] = varc; + s->mb_mean[s->mb_stride * mb_y + mb_x] = (sum+128)>>8; s->me.mb_var_sum_temp += varc; } } diff --git a/libavcodec/ratecontrol.c b/libavcodec/ratecontrol.c index a7614d75e2..53930f147b 100644 --- a/libavcodec/ratecontrol.c +++ b/libavcodec/ratecontrol.c @@ -751,15 +751,14 @@ static void adaptive_quantization(MpegEncContext *s, double q) float *bits_tab = s->bits_tab; const int qmin = s->avctx->mb_lmin; const int qmax = s->avctx->mb_lmax; - Picture *const pic = &s->current_picture; const int mb_width = s->mb_width; const int mb_height = s->mb_height; for (i = 0; i < s->mb_num; i++) { const int mb_xy = s->mb_index2xy[i]; - float temp_cplx = sqrt(pic->mc_mb_var[mb_xy]); // FIXME merge in pow() - float spat_cplx = sqrt(pic->mb_var[mb_xy]); - const int lumi = pic->mb_mean[mb_xy]; + float temp_cplx = sqrt(s->mc_mb_var[mb_xy]); // FIXME merge in pow() + float spat_cplx = sqrt(s->mb_var[mb_xy]); + const int lumi = s->mb_mean[mb_xy]; float bits, cplx, factor; int mb_x = mb_xy % s->mb_stride; int mb_y = mb_xy / s->mb_stride; diff --git a/libavcodec/svq1enc.c b/libavcodec/svq1enc.c index e5216a99cd..36bc214d76 100644 --- a/libavcodec/svq1enc.c +++ b/libavcodec/svq1enc.c @@ -313,9 +313,9 @@ static int svq1_encode_plane(SVQ1EncContext *s, int plane, s->m.mb_type = s->mb_type; // dummies, to avoid segfaults - s->m.current_picture.mb_mean = (uint8_t *)s->dummy; - s->m.current_picture.mb_var = (uint16_t *)s->dummy; - s->m.current_picture.mc_mb_var = (uint16_t *)s->dummy; + s->m.mb_mean = (uint8_t *)s->dummy; + s->m.mb_var = (uint16_t *)s->dummy; + s->m.mc_mb_var = (uint16_t *)s->dummy; s->m.current_picture.mb_type = s->dummy; s->m.current_picture.motion_val[0] = s->motion_val8[plane] + 2; diff --git a/tests/ref/seek/vsynth_lena-mpeg4-adap b/tests/ref/seek/vsynth_lena-mpeg4-adap index 0611ee4511..fe841ef973 100644 --- a/tests/ref/seek/vsynth_lena-mpeg4-adap +++ b/tests/ref/seek/vsynth_lena-mpeg4-adap @@ -2,45 +2,45 @@ ret: 0 st: 0 flags:1 dts: 0.000000 pts: NOPTS pos: 5652 size: 6855 ret: 0 st:-1 flags:0 ts:-1.000000 ret: 0 st: 0 flags:1 dts: 0.000000 pts: NOPTS pos: 5652 size: 6855 ret: 0 st:-1 flags:1 ts: 1.894167 -ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 174446 size: 16883 +ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 161318 size: 19176 ret: 0 st: 0 flags:0 ts: 0.800000 -ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 98220 size: 17063 +ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 75056 size: 19178 ret:-1 st: 0 flags:1 ts:-0.320000 ret:-1 st:-1 flags:0 ts: 2.576668 ret: 0 st:-1 flags:1 ts: 1.470835 -ret: 0 st: 0 flags:1 dts: 1.360000 pts: NOPTS pos: 135582 size: 17525 +ret: 0 st: 0 flags:1 dts: 1.360000 pts: NOPTS pos: 118696 size: 20018 ret: 0 st: 0 flags:0 ts: 0.360000 -ret: 0 st: 0 flags:1 dts: 0.400000 pts: NOPTS pos: 59446 size: 17261 +ret: 0 st: 0 flags:1 dts: 0.400000 pts: NOPTS pos: 35800 size: 17261 ret:-1 st: 0 flags:1 ts:-0.760000 ret:-1 st:-1 flags:0 ts: 2.153336 ret: 0 st:-1 flags:1 ts: 1.047503 -ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 98220 size: 17063 +ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 75056 size: 19178 ret: 0 st: 0 flags:0 ts:-0.040000 ret: 0 st: 0 flags:1 dts: 0.000000 pts: NOPTS pos: 5652 size: 6855 ret: 0 st: 0 flags:1 ts: 2.840000 -ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 174446 size: 16883 +ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 161318 size: 19176 ret: 0 st:-1 flags:0 ts: 1.730004 -ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 174446 size: 16883 +ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 161318 size: 19176 ret: 0 st:-1 flags:1 ts: 0.624171 -ret: 0 st: 0 flags:1 dts: 0.400000 pts: NOPTS pos: 59446 size: 17261 +ret: 0 st: 0 flags:1 dts: 0.400000 pts: NOPTS pos: 35800 size: 17261 ret: 0 st: 0 flags:0 ts:-0.480000 ret: 0 st: 0 flags:1 dts: 0.000000 pts: NOPTS pos: 5652 size: 6855 ret: 0 st: 0 flags:1 ts: 2.400000 -ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 174446 size: 16883 +ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 161318 size: 19176 ret: 0 st:-1 flags:0 ts: 1.306672 -ret: 0 st: 0 flags:1 dts: 1.360000 pts: NOPTS pos: 135582 size: 17525 +ret: 0 st: 0 flags:1 dts: 1.360000 pts: NOPTS pos: 118696 size: 20018 ret: 0 st:-1 flags:1 ts: 0.200839 ret: 0 st: 0 flags:1 dts: 0.000000 pts: NOPTS pos: 5652 size: 6855 ret: 0 st: 0 flags:0 ts:-0.920000 ret: 0 st: 0 flags:1 dts: 0.000000 pts: NOPTS pos: 5652 size: 6855 ret: 0 st: 0 flags:1 ts: 2.000000 -ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 174446 size: 16883 +ret: 0 st: 0 flags:1 dts: 1.840000 pts: NOPTS pos: 161318 size: 19176 ret: 0 st:-1 flags:0 ts: 0.883340 -ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 98220 size: 17063 +ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 75056 size: 19178 ret:-1 st:-1 flags:1 ts:-0.222493 ret:-1 st: 0 flags:0 ts: 2.680000 ret: 0 st: 0 flags:1 ts: 1.560000 -ret: 0 st: 0 flags:1 dts: 1.360000 pts: NOPTS pos: 135582 size: 17525 +ret: 0 st: 0 flags:1 dts: 1.360000 pts: NOPTS pos: 118696 size: 20018 ret: 0 st:-1 flags:0 ts: 0.460008 -ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 98220 size: 17063 +ret: 0 st: 0 flags:1 dts: 0.880000 pts: NOPTS pos: 75056 size: 19178 ret:-1 st:-1 flags:1 ts:-0.645825 diff --git a/tests/ref/vsynth/vsynth1-mpeg4-adap b/tests/ref/vsynth/vsynth1-mpeg4-adap index e6bd6a59bd..67705f0f43 100644 --- a/tests/ref/vsynth/vsynth1-mpeg4-adap +++ b/tests/ref/vsynth/vsynth1-mpeg4-adap @@ -1,4 +1,4 @@ -f120f0bf976bb510c5b5305fe7d8159a *tests/data/fate/vsynth1-mpeg4-adap.avi -403436 tests/data/fate/vsynth1-mpeg4-adap.avi -fad0b9dc08fe4a95b297af1a7411c1e9 *tests/data/fate/vsynth1-mpeg4-adap.out.rawvideo -stddev: 14.05 PSNR: 25.17 MAXDIFF: 184 bytes: 7603200/ 7603200 +0f1cbbdc3f9b91f2d9ac3d1fc2cf7d4e *tests/data/fate/vsynth1-mpeg4-adap.avi +325518 tests/data/fate/vsynth1-mpeg4-adap.avi +1e6c596f9f491fbf15920ef1bace7fb8 *tests/data/fate/vsynth1-mpeg4-adap.out.rawvideo +stddev: 14.12 PSNR: 25.13 MAXDIFF: 184 bytes: 7603200/ 7603200 diff --git a/tests/ref/vsynth/vsynth2-mpeg4-adap b/tests/ref/vsynth/vsynth2-mpeg4-adap index a3223f6363..35b2b6aac9 100644 --- a/tests/ref/vsynth/vsynth2-mpeg4-adap +++ b/tests/ref/vsynth/vsynth2-mpeg4-adap @@ -1,4 +1,4 @@ -4bff98da2342836476da817428594403 *tests/data/fate/vsynth2-mpeg4-adap.avi -213508 tests/data/fate/vsynth2-mpeg4-adap.avi -0c709f2b81f4593eaa29490332c2cb39 *tests/data/fate/vsynth2-mpeg4-adap.out.rawvideo -stddev: 4.87 PSNR: 34.36 MAXDIFF: 86 bytes: 7603200/ 7603200 +06a397fe43dab7b6cf56870410fbbbaf *tests/data/fate/vsynth2-mpeg4-adap.avi +203000 tests/data/fate/vsynth2-mpeg4-adap.avi +686565d42d8ba5aea790824b04fa0a18 *tests/data/fate/vsynth2-mpeg4-adap.out.rawvideo +stddev: 4.55 PSNR: 34.95 MAXDIFF: 84 bytes: 7603200/ 7603200 diff --git a/tests/ref/vsynth/vsynth3-mpeg4-adap b/tests/ref/vsynth/vsynth3-mpeg4-adap index 0942f83015..6386f13466 100644 --- a/tests/ref/vsynth/vsynth3-mpeg4-adap +++ b/tests/ref/vsynth/vsynth3-mpeg4-adap @@ -1,4 +1,4 @@ -c16e5c2436ca9953517eadba562768e9 *tests/data/fate/vsynth3-mpeg4-adap.avi -43706 tests/data/fate/vsynth3-mpeg4-adap.avi -b42b614e19e7c4859fca1af6d4e36eae *tests/data/fate/vsynth3-mpeg4-adap.out.rawvideo -stddev: 5.48 PSNR: 33.34 MAXDIFF: 53 bytes: 86700/ 86700 +6b2f641f2e68b11b992fd6ba1ed66a21 *tests/data/fate/vsynth3-mpeg4-adap.avi +41012 tests/data/fate/vsynth3-mpeg4-adap.avi +3483a2032cb02c3a37f5e43b128e59ed *tests/data/fate/vsynth3-mpeg4-adap.out.rawvideo +stddev: 5.79 PSNR: 32.87 MAXDIFF: 49 bytes: 86700/ 86700 diff --git a/tests/ref/vsynth/vsynth_lena-mpeg4-adap b/tests/ref/vsynth/vsynth_lena-mpeg4-adap index 10af80b397..91edafe9b0 100644 --- a/tests/ref/vsynth/vsynth_lena-mpeg4-adap +++ b/tests/ref/vsynth/vsynth_lena-mpeg4-adap @@ -1,4 +1,4 @@ -c6108621b1202d32dac68b1944c5b8c2 *tests/data/fate/vsynth_lena-mpeg4-adap.avi -198500 tests/data/fate/vsynth_lena-mpeg4-adap.avi -87b6dbe98d276137fceaae2fa672eced *tests/data/fate/vsynth_lena-mpeg4-adap.out.rawvideo -stddev: 3.75 PSNR: 36.65 MAXDIFF: 71 bytes: 7603200/ 7603200 +633da125f46391eef33bb031cd728f4b *tests/data/fate/vsynth_lena-mpeg4-adap.avi +187598 tests/data/fate/vsynth_lena-mpeg4-adap.avi +21312bfcb28c40299fb27a5b03477f8c *tests/data/fate/vsynth_lena-mpeg4-adap.out.rawvideo +stddev: 3.63 PSNR: 36.92 MAXDIFF: 71 bytes: 7603200/ 7603200