From patchwork Mon Dec 28 10:41:22 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul B Mahol X-Patchwork-Id: 24665 Return-Path: X-Original-To: patchwork@ffaux-bg.ffmpeg.org Delivered-To: patchwork@ffaux-bg.ffmpeg.org Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by ffaux.localdomain (Postfix) with ESMTP id 383F144AF0F for ; Mon, 28 Dec 2020 13:05:46 +0200 (EET) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 171E4688098; Mon, 28 Dec 2020 13:05:46 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mail-ej1-f41.google.com (mail-ej1-f41.google.com [209.85.218.41]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 5AA2F687FC5 for ; Mon, 28 Dec 2020 13:05:39 +0200 (EET) Received: by mail-ej1-f41.google.com with SMTP id x16so13723677ejj.7 for ; Mon, 28 Dec 2020 03:05:39 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:subject:date:message-id; bh=+bD9DCDuDdehbhoubXu3Yzii1e/lq1kxFkUuK5MfndI=; b=hlh/abVS/MCGp/2nqB3bP+Zfjh4BYq+rTYCnIGLaeiaLKqm70rDY9U4uL6wxbNUKgb hkfULGy/BFbIOGayYdr5AQYuZJykX7sSf5wjGn1MhShnzgpNGRGdHQKljtV63FB/uYD1 y9wC/IMSsPp6V4kAl1Fi/LiD9v32r+/KntE3tnnr2bKdBVgwOLQqlZbjBT5TxAnNlKpU 0OmDlMLsD0YHSQJq1b1RsGd/sYzFoCmIbXtRVlc/iAL6/8P9Oi+7jgfOJyrhUa2xRVla qVUZx337y3dfeRhtAo8EPoSZWjlCGrSF/tYdN6BKR9PbrVZlSTsPQaEk3z5ONlvNCjV7 1LFg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:subject:date:message-id; bh=+bD9DCDuDdehbhoubXu3Yzii1e/lq1kxFkUuK5MfndI=; b=mXSkqDyL0qMkKuQ1zGfeTbHBLMVVAo778yNtQ9Gr6TvNOpM5sHx06KgLJL3OFGSjYB HuFC03uGLr9sEYyBs1A4HRrwliuzrdhE0Q4CpIgy4Pfrc3HjW1RL1IObkxz0fdx6YBOB AGiadd71LH81GRkIPl0T/cSCXZs5yud3ZxxunKg5xjlYDxmiHQdSp1bZZaSesSBw9bSP B97idduiWpHa0bUYtlOESLrRAbPpKVOdmR5bKD4EbVYrbdCCYlV+hJ/VEaUP5ZSgN0LH +A1LE6No3IPk3w/GyssWhHScwifRoCbc5RTEVBEk61JBmRqUGVY3UlA8UCyD91RjTNKv WGCg== X-Gm-Message-State: AOAM533867CwLSxNjxSPJWiJfayns3CNPFxxqyETRBBtGzjerpHpXbLL UtgRNTgDuSuncGs/OhR0ea+Q0MLpt+5wMA== X-Google-Smtp-Source: ABdhPJwymhMLLPwmff52XSjItjo8Rd+A5tb1y/LpWvuc2JJPqaucvazsgVPvtNzYy/lwTPNfgdXYLg== X-Received: by 2002:a17:906:3953:: with SMTP id g19mr40615976eje.429.1609152091256; Mon, 28 Dec 2020 02:41:31 -0800 (PST) Received: from localhost.localdomain ([31.45.254.141]) by smtp.gmail.com with ESMTPSA id r18sm17660093edx.41.2020.12.28.02.41.30 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 28 Dec 2020 02:41:30 -0800 (PST) From: Paul B Mahol To: ffmpeg-devel@ffmpeg.org Date: Mon, 28 Dec 2020 11:41:22 +0100 Message-Id: <20201228104122.21327-1-onemda@gmail.com> X-Mailer: git-send-email 2.17.1 Subject: [FFmpeg-devel] [PATCH] avfilter: add temporal midway equalizer filter X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 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 MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Signed-off-by: Paul B Mahol --- doc/filters.texi | 22 ++ libavfilter/Makefile | 1 + libavfilter/allfilters.c | 1 + libavfilter/vf_tmidequalizer.c | 440 +++++++++++++++++++++++++++++++++ 4 files changed, 464 insertions(+) create mode 100644 libavfilter/vf_tmidequalizer.c diff --git a/doc/filters.texi b/doc/filters.texi index f71d33c602..1a515c2681 100644 --- a/doc/filters.texi +++ b/doc/filters.texi @@ -19319,6 +19319,28 @@ minimum values, and @code{1} maximum values. This filter supports all above options as @ref{commands}, excluding option @code{radius}. +@section tmidequalizer + +Apply Temporal Midway Video Equalization effect. + +Midway Video Equalization adjusts a sequence of video frames to have the same +histograms, while maintaining their dynamics as much as possible. It's +useful for e.g. matching exposures from a video frames sequence. + +This filter accepts the following option: + +@table @option +@item radius +Set filtering radius. Default is @code{5}. Allowed range is from 1 to 127. + +@item sigma +Set filtering sigma. Default is @code{0.5}. This controls strength of filtering. +Setting this option to 0 effectively does nothing. + +@item planes +Set which planes to process. Default is @code{15}, which is all available planes. +@end table + @section tmix Mix successive video frames. diff --git a/libavfilter/Makefile b/libavfilter/Makefile index 5b4831781f..526da8d69e 100644 --- a/libavfilter/Makefile +++ b/libavfilter/Makefile @@ -433,6 +433,7 @@ OBJS-$(CONFIG_TILE_FILTER) += vf_tile.o OBJS-$(CONFIG_TINTERLACE_FILTER) += vf_tinterlace.o OBJS-$(CONFIG_TLUT2_FILTER) += vf_lut2.o framesync.o OBJS-$(CONFIG_TMEDIAN_FILTER) += vf_xmedian.o framesync.o +OBJS-$(CONFIG_TMIDEQUALIZER_FILTER) += vf_tmidequalizer.o OBJS-$(CONFIG_TMIX_FILTER) += vf_mix.o framesync.o OBJS-$(CONFIG_TONEMAP_FILTER) += vf_tonemap.o colorspace.o OBJS-$(CONFIG_TONEMAP_OPENCL_FILTER) += vf_tonemap_opencl.o colorspace.o opencl.o \ diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c index 1b77bf1c81..ce317dfa1c 100644 --- a/libavfilter/allfilters.c +++ b/libavfilter/allfilters.c @@ -413,6 +413,7 @@ extern AVFilter ff_vf_tile; extern AVFilter ff_vf_tinterlace; extern AVFilter ff_vf_tlut2; extern AVFilter ff_vf_tmedian; +extern AVFilter ff_vf_tmidequalizer; extern AVFilter ff_vf_tmix; extern AVFilter ff_vf_tonemap; extern AVFilter ff_vf_tonemap_opencl; diff --git a/libavfilter/vf_tmidequalizer.c b/libavfilter/vf_tmidequalizer.c new file mode 100644 index 0000000000..6e3f3ccce1 --- /dev/null +++ b/libavfilter/vf_tmidequalizer.c @@ -0,0 +1,440 @@ +/* + * Copyright (c) 2020 Paul B Mahol + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "libavutil/imgutils.h" +#include "libavutil/pixdesc.h" +#include "libavutil/opt.h" +#include "avfilter.h" +#include "formats.h" +#include "internal.h" +#include "video.h" + +typedef struct TMidEqualizerContext { + const AVClass *class; + + int planes; + int radius; + float sigma; + + int plane_width[4], plane_height[4]; + int nb_frames; + int depth; + int f_frames; + int l_frames; + int del_frame; + int cur_frame; + int nb_planes; + int histogram_size; + float kernel[127]; + float *histogram[4][256]; + float *change[4]; + + AVFrame **frames; + + void (*compute_histogram)(const uint8_t *ssrc, ptrdiff_t linesize, + int w, int h, float *histogram, size_t hsize); + void (*apply_contrast_change)(const uint8_t *src, ptrdiff_t src_linesize, + uint8_t *dst, ptrdiff_t dst_linesize, + int w, int h, float *change, float *orig); +} TMidEqualizerContext; + +#define OFFSET(x) offsetof(TMidEqualizerContext, x) +#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM + +static const AVOption tmidequalizer_options[] = { + { "radius", "set radius", OFFSET(radius), AV_OPT_TYPE_INT, {.i64=5}, 1, 127, FLAGS }, + { "sigma", "set sigma", OFFSET(sigma), AV_OPT_TYPE_FLOAT, {.dbl=0.5}, 0, 1, FLAGS }, + { "planes", "set planes", OFFSET(planes), AV_OPT_TYPE_INT, {.i64=0xF}, 0, 0xF, FLAGS }, + { NULL } +}; + +AVFILTER_DEFINE_CLASS(tmidequalizer); + +static int query_formats(AVFilterContext *ctx) +{ + static const enum AVPixelFormat pix_fmts[] = { + AV_PIX_FMT_YUVA444P, AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV440P, + AV_PIX_FMT_YUVJ444P, AV_PIX_FMT_YUVJ440P, + AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUV420P, + AV_PIX_FMT_YUVJ422P, AV_PIX_FMT_YUVJ420P, + AV_PIX_FMT_YUVJ411P, AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, + AV_PIX_FMT_GBRP, AV_PIX_FMT_GBRAP, + AV_PIX_FMT_GRAY8, AV_PIX_FMT_GRAY9, AV_PIX_FMT_GRAY10, AV_PIX_FMT_GRAY12, AV_PIX_FMT_GRAY14, + AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9, + AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10, + AV_PIX_FMT_YUV420P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV444P12, + AV_PIX_FMT_YUV420P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV444P14, + AV_PIX_FMT_GBRP9, AV_PIX_FMT_GBRP10, AV_PIX_FMT_GBRP12, AV_PIX_FMT_GBRP14, + AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA444P9, + AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA444P10, + AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA444P12, + AV_PIX_FMT_GBRAP10, AV_PIX_FMT_GBRAP12, + AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16, + AV_PIX_FMT_YUVA420P16, AV_PIX_FMT_YUVA422P16, AV_PIX_FMT_YUVA444P16, + AV_PIX_FMT_GBRP16, AV_PIX_FMT_GBRAP16, + AV_PIX_FMT_GRAY16, + AV_PIX_FMT_NONE + }; + + return ff_set_common_formats(ctx, ff_make_format_list(pix_fmts)); +} + +static void compute_contrast_function(const float *const histograms[256], + const float *const kernel, + int nb_frames, int radius, int hsize, + float *f, int idx) +{ + const float *const h1 = histograms[idx]; + int p2[256] = { 0 }; + + for (int p1 = 0; p1 < hsize; p1++) { + float weight = 1.f; + float sum = p1 * weight; + + for (int j = 0; j < radius; j++) { + const int nidx = ((idx - radius + j) % nb_frames); + const float *const h2 = histograms[nidx < 0 ? nidx + nb_frames: nidx]; + int k = j; + + for (; p2[k] < hsize && h2[p2[k]] < h1[p1]; p2[k]++); + if (p2[k] == hsize) + p2[k]--; + + weight += kernel[j]; + sum += kernel[j] * p2[k]; + } + + for (int j = radius + 1; j < nb_frames; j++) { + const int nidx = (idx - radius + j) % nb_frames; + const float *const h2 = histograms[nidx < 0 ? nidx + nb_frames: nidx]; + int k = j; + + for (; p2[k] < hsize && h2[p2[k]] < h1[p1]; p2[k]++); + if (p2[k] == hsize) + p2[k]--; + + weight += kernel[j - radius - 1]; + sum += kernel[j - radius - 1] * p2[k]; + } + + f[p1] = sum / weight; + } +} + +static void apply_contrast_change8(const uint8_t *src, ptrdiff_t src_linesize, + uint8_t *dst, ptrdiff_t dst_linesize, + int w, int h, float *change, float *orig) +{ + for (int y = 0; y < h; y++) { + for (int x = 0; x < w; x++) + dst[x] = lrintf(change[src[x]]); + + dst += dst_linesize; + src += src_linesize; + } +} + +static void apply_contrast_change16(const uint8_t *ssrc, ptrdiff_t src_linesize, + uint8_t *ddst, ptrdiff_t dst_linesize, + int w, int h, float *change, float *orig) +{ + const uint16_t *src = (const uint16_t *)ssrc; + uint16_t *dst = (uint16_t *)ddst; + + for (int y = 0; y < h; y++) { + for (int x = 0; x < w; x++) + dst[x] = lrintf(change[src[x]]); + + dst += dst_linesize / 2; + src += src_linesize / 2; + } +} + +static int filter_frame(AVFilterLink *inlink, AVFrame *in) +{ + AVFilterContext *ctx = inlink->dst; + TMidEqualizerContext *s = ctx->priv; + AVFilterLink *outlink = ctx->outputs[0]; + AVFrame *out; + int eof = 0; + + if (!in) { + int idx = s->f_frames < s->nb_frames ? s->radius : s->del_frame ? s->del_frame - 1 : s->nb_frames - 1; + + if (s->f_frames < s->nb_frames) { + s->l_frames = s->nb_frames - s->f_frames; + } else { + s->l_frames++; + } + in = av_frame_clone(s->frames[idx]); + if (!in) + return AVERROR(ENOMEM); + eof = 1; + } + + if (s->f_frames < s->nb_frames) { + s->frames[s->f_frames] = in; + + for (int p = 0; p < s->nb_planes; p++) { + s->compute_histogram(in->data[p], in->linesize[p], + s->plane_width[p], s->plane_height[p], + s->histogram[p][s->f_frames], + s->histogram_size); + } + + s->f_frames++; + + while (s->f_frames <= s->radius) { + s->frames[s->f_frames] = av_frame_clone(in); + if (!s->frames[s->f_frames]) + return AVERROR(ENOMEM); + for (int p = 0; p < s->nb_planes; p++) { + memcpy(s->histogram[p][s->f_frames], + s->histogram[p][s->f_frames - 1], + s->histogram_size * sizeof(float)); + } + s->f_frames++; + } + + if (!eof && s->f_frames < s->nb_frames) { + return 0; + } else { + while (s->f_frames < s->nb_frames) { + s->frames[s->f_frames] = av_frame_clone(in); + if (!s->frames[s->f_frames]) + return AVERROR(ENOMEM); + for (int p = 0; p < s->nb_planes; p++) { + memcpy(s->histogram[p][s->f_frames], + s->histogram[p][s->f_frames - 1], + s->histogram_size * sizeof(float)); + } + s->f_frames++; + } + } + s->cur_frame = s->radius; + s->del_frame = 0; + } else { + av_frame_free(&s->frames[s->del_frame]); + s->frames[s->del_frame] = in; + + for (int p = 0; p < s->nb_planes; p++) { + s->compute_histogram(in->data[p], in->linesize[p], + s->plane_width[p], s->plane_height[p], + s->histogram[p][s->del_frame], + s->histogram_size); + } + + s->del_frame++; + if (s->del_frame >= s->nb_frames) + s->del_frame = 0; + } + + if (ctx->is_disabled) { + const int idx = s->cur_frame; + + out = av_frame_clone(s->frames[idx]); + if (!out) + return AVERROR(ENOMEM); + } else { + const int idx = s->cur_frame; + + in = s->frames[idx]; + out = ff_get_video_buffer(outlink, outlink->w, outlink->h); + if (!out) + return AVERROR(ENOMEM); + av_frame_copy_props(out, in); + + for (int p = 0; p < s->nb_planes; p++) { + if (!((1 << p) & s->planes)) { + av_image_copy_plane(out->data[p], out->linesize[p], in->data[p], in->linesize[p], + s->plane_width[p] * (1 + (s->depth > 8)), s->plane_height[p]); + continue; + } + + compute_contrast_function((const float *const *)s->histogram[p], s->kernel, + s->nb_frames, s->radius, s->histogram_size, s->change[p], idx); + + s->apply_contrast_change(in->data[p], in->linesize[p], + out->data[p], out->linesize[p], + s->plane_width[p], s->plane_height[p], + s->change[p], s->histogram[p][idx]); + } + } + + s->cur_frame++; + if (s->cur_frame >= s->nb_frames) + s->cur_frame = 0; + + return ff_filter_frame(outlink, out); +} + +static void compute_histogram8(const uint8_t *src, ptrdiff_t linesize, + int w, int h, float *histogram, size_t hsize) +{ + memset(histogram, 0, hsize * sizeof(*histogram)); + + for (int y = 0; y < h; y++) { + for (int x = 0; x < w; x++) + histogram[src[x]] += 1; + src += linesize; + } + + for (int x = 0; x < hsize; x++) + histogram[x] /= hsize; + + for (int x = 1; x < hsize; x++) + histogram[x] += histogram[x-1]; +} + +static void compute_histogram16(const uint8_t *ssrc, ptrdiff_t linesize, + int w, int h, float *histogram, size_t hsize) +{ + const uint16_t *src = (const uint16_t *)ssrc; + + memset(histogram, 0, hsize * sizeof(*histogram)); + + for (int y = 0; y < h; y++) { + for (int x = 0; x < w; x++) + histogram[src[x]] += 1; + src += linesize / 2; + } + + for (int x = 0; x < hsize; x++) + histogram[x] /= hsize; + + for (int x = 1; x < hsize; x++) + histogram[x] += histogram[x-1]; +} + +static int config_input(AVFilterLink *inlink) +{ + AVFilterContext *ctx = inlink->dst; + TMidEqualizerContext *s = ctx->priv; + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format); + float sigma = s->radius * s->sigma; + int vsub, hsub; + + s->depth = desc->comp[0].depth; + s->nb_frames = s->radius * 2 + 1; + s->nb_planes = av_pix_fmt_count_planes(inlink->format); + + hsub = desc->log2_chroma_w; + vsub = desc->log2_chroma_h; + + s->plane_height[0] = s->plane_height[3] = inlink->h; + s->plane_width[0] = s->plane_width[3] = inlink->w; + s->plane_height[1] = s->plane_height[2] = AV_CEIL_RSHIFT(inlink->h, vsub); + s->plane_width[1] = s->plane_width[2] = AV_CEIL_RSHIFT(inlink->w, hsub); + + s->histogram_size = 1 << s->depth; + + for (int n = 0; n < s->radius; n++) + s->kernel[n] = expf(-0.5 * (n + 1) * (n + 1) / (sigma * sigma)); + + for (int p = 0; p < s->nb_planes; p++) { + for (int n = 0; n < s->nb_frames; n++) { + s->histogram[p][n] = av_calloc(s->histogram_size, sizeof(float)); + if (!s->histogram[p][n]) + return AVERROR(ENOMEM); + } + + s->change[p] = av_calloc(s->histogram_size, sizeof(float)); + if (!s->change[p]) + return AVERROR(ENOMEM); + } + + if (!s->frames) + s->frames = av_calloc(s->nb_frames, sizeof(*s->frames)); + if (!s->frames) + return AVERROR(ENOMEM); + + s->compute_histogram = s->depth <= 8 ? compute_histogram8 : compute_histogram16; + s->apply_contrast_change = s->depth <= 8 ? apply_contrast_change8 : apply_contrast_change16; + + return 0; +} + +static int request_frame(AVFilterLink *outlink) +{ + AVFilterContext *ctx = outlink->src; + TMidEqualizerContext *s = ctx->priv; + int ret; + + ret = ff_request_frame(ctx->inputs[0]); + if (ret == AVERROR_EOF && s->l_frames < s->radius) { + ret = filter_frame(ctx->inputs[0], NULL); + } + + return ret; +} + +static void free_histograms(AVFilterContext *ctx, int x, int nb_frames) +{ + TMidEqualizerContext *s = ctx->priv; + + for (int n = 0; n < nb_frames; n++) + av_freep(&s->histogram[x][n]); + av_freep(&s->change[x]); +} + +static av_cold void uninit(AVFilterContext *ctx) +{ + TMidEqualizerContext *s = ctx->priv; + + free_histograms(ctx, 0, s->nb_frames); + free_histograms(ctx, 1, s->nb_frames); + free_histograms(ctx, 2, s->nb_frames); + free_histograms(ctx, 3, s->nb_frames); + + for (int i = 0; i < s->nb_frames && s->frames; i++) + av_frame_free(&s->frames[i]); + av_freep(&s->frames); +} + +static const AVFilterPad tmidequalizer_inputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .config_props = config_input, + .filter_frame = filter_frame, + }, + { NULL } +}; + +static const AVFilterPad tmidequalizer_outputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .request_frame = request_frame, + }, + { NULL } +}; + +AVFilter ff_vf_tmidequalizer = { + .name = "tmidequalizer", + .description = NULL_IF_CONFIG_SMALL("Apply Temporal Midway Equalization."), + .priv_size = sizeof(TMidEqualizerContext), + .uninit = uninit, + .query_formats = query_formats, + .inputs = tmidequalizer_inputs, + .outputs = tmidequalizer_outputs, + .priv_class = &tmidequalizer_class, + .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_INTERNAL, +};