From patchwork Tue Sep 29 03:44:34 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mark Reid X-Patchwork-Id: 22654 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 56E51448617 for ; Tue, 29 Sep 2020 06:44:55 +0300 (EEST) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 3ADBE6882E8; Tue, 29 Sep 2020 06:44:55 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mail-pg1-f193.google.com (mail-pg1-f193.google.com [209.85.215.193]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id E9EF1688179 for ; Tue, 29 Sep 2020 06:44:46 +0300 (EEST) Received: by mail-pg1-f193.google.com with SMTP id x16so2758668pgj.3 for ; Mon, 28 Sep 2020 20:44:46 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=Giuyjx5K1l+eBtWHu1vhuiX/iVK8rpXC+M35qj24xXo=; b=e8GtM14SIDupNcB+ZlBml+265uUz5u9QqRGUH/RDRno9VJZOux0bKvDOY3mZCBQitG 2J/S1WSFNu2mEFncp/CvwOSjpokOEeAHAepkVER+7g4MhNtgLAYCtIsqEbhWewsRPBAV RhKrb9nW1GeD3PVieqqi2f8sn7vKPc5I0e/24QQ7d9fNN2KLWAC9VBhnydVrnCnLX6bi N6zff4JsdvAl7GJcNc8Dai1/5RCY53phHB57ZrQdPBQvipb9gjvuLI8AmNoGJcr17MOv iRuWvW5iOgDQg3DQm1T8JKR86J1FeOmyZq0xTW5IZqAeLRUJxK+bp1kBD1fmEdfOlCeN SALg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=Giuyjx5K1l+eBtWHu1vhuiX/iVK8rpXC+M35qj24xXo=; b=AJJvsCnQyH6bqKwOlkiLBEYFKnwzds6k9vGlpEekO//Z0rm+2U/q9RoWj7PzwOvJAk SumV/fnmWyriMliO/nlkqOW8lSQtb6TvfxjWYhe5FE1wPcF7x4IBEHjiajyFAGDb5S25 PjD4P5eIZ2Sg6qGLXdNJqnh80fejayEIG5RGQddDNtWqWfKDQSgouu59KLQf4LDiAqA6 OSzvxCmWsPzD9SuBZbdCmgXpkrvqlz/mGsySfTYjCT3DbV5282eQLWxCawc8soxYaiff sk1gNuNIHqai+amBYcFJ/qrfpZh/gYAHNQnKN7jaihe/Vc803fb9gOE+zJIX78ytpz31 x3Qw== X-Gm-Message-State: AOAM5325YggCnVUfsY/69272G/l0NkTwn9w3EHdGmn9Ghoq1s7sZZJbJ /7DMHNBn6FnOKTcOEhmHirtzRs7TokE= X-Google-Smtp-Source: ABdhPJyimQ0g93Z3mTeKDOAXreDbtYvtVInI1lVIS6E/mWlkh5Kb5ypQYAimJiqhVK2ucAK3+/YhMg== X-Received: by 2002:a17:902:c403:b029:d2:83e9:8f8c with SMTP id k3-20020a170902c403b02900d283e98f8cmr2601422plk.80.1601351084504; Mon, 28 Sep 2020 20:44:44 -0700 (PDT) Received: from localhost.localdomain (S01069050ca607903.vc.shawcable.net. [174.7.236.190]) by smtp.gmail.com with ESMTPSA id z23sm3315927pfj.177.2020.09.28.20.44.43 (version=TLS1_2 cipher=ECDHE-ECDSA-AES128-GCM-SHA256 bits=128/128); Mon, 28 Sep 2020 20:44:43 -0700 (PDT) From: mindmark@gmail.com To: ffmpeg-devel@ffmpeg.org Date: Mon, 28 Sep 2020 20:44:34 -0700 Message-Id: <20200929034434.59110-2-mindmark@gmail.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200929034434.59110-1-mindmark@gmail.com> References: <20200929034434.59110-1-mindmark@gmail.com> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH v2 2/2] libswcale/input: use more accurate rgbf32 yuv conversions 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 Cc: Mark Reid Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" From: Mark Reid --- libswscale/input.c | 12 ++- tests/ref/fate/filter-pixfmts-scale | 8 +- tests/ref/fate/sws-floatimg-cmp | 122 ++++++++++++++-------------- 3 files changed, 70 insertions(+), 72 deletions(-) diff --git a/libswscale/input.c b/libswscale/input.c index 064ed5902f..67a85b0418 100644 --- a/libswscale/input.c +++ b/libswscale/input.c @@ -984,15 +984,14 @@ static av_always_inline void planar_rgbf32_to_uv(uint8_t *_dstU, uint8_t *_dstV, uint16_t *dstV = (uint16_t *)_dstV; int32_t ru = rgb2yuv[RU_IDX], gu = rgb2yuv[GU_IDX], bu = rgb2yuv[BU_IDX]; int32_t rv = rgb2yuv[RV_IDX], gv = rgb2yuv[GV_IDX], bv = rgb2yuv[BV_IDX]; - int bpc = 16; - int shift = 14; + for (i = 0; i < width; i++) { int g = av_clip_uint16(lrintf(65535.0f * rdpx(src[0] + i))); int b = av_clip_uint16(lrintf(65535.0f * rdpx(src[1] + i))); int r = av_clip_uint16(lrintf(65535.0f * rdpx(src[2] + i))); - dstU[i] = (ru*r + gu*g + bu*b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + shift - 14); - dstV[i] = (rv*r + gv*g + bv*b + (257 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + shift - 14); + dstU[i] = (ru*r + gu*g + bu*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; + dstV[i] = (rv*r + gv*g + bv*b + (0x10001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; } } @@ -1003,14 +1002,13 @@ static av_always_inline void planar_rgbf32_to_y(uint8_t *_dst, const uint8_t *_s uint16_t *dst = (uint16_t *)_dst; int32_t ry = rgb2yuv[RY_IDX], gy = rgb2yuv[GY_IDX], by = rgb2yuv[BY_IDX]; - int bpc = 16; - int shift = 14; + for (i = 0; i < width; i++) { int g = av_clip_uint16(lrintf(65535.0f * rdpx(src[0] + i))); int b = av_clip_uint16(lrintf(65535.0f * rdpx(src[1] + i))); int r = av_clip_uint16(lrintf(65535.0f * rdpx(src[2] + i))); - dst[i] = ((ry*r + gy*g + by*b + (33 << (RGB2YUV_SHIFT + bpc - 9))) >> (RGB2YUV_SHIFT + shift - 14)); + dst[i] = (ry*r + gy*g + by*b + (0x2001 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT; } } diff --git a/tests/ref/fate/filter-pixfmts-scale b/tests/ref/fate/filter-pixfmts-scale index d7020ad2c3..30e7cd5b06 100644 --- a/tests/ref/fate/filter-pixfmts-scale +++ b/tests/ref/fate/filter-pixfmts-scale @@ -25,8 +25,8 @@ gbrap12be 1d9b57766ba9c2192403f43967cb9af0 gbrap12le bb1ba1c157717db3dd612a76d38a018e gbrap16be c72b935a6e57a8e1c37bff08c2db55b1 gbrap16le 13eb0e62b1ac9c1c86c81521eaefab5f -gbrapf32be 42e53d9edccbd9e09c4cd78780ba92f3 -gbrapf32le eebf3973ef94c841f0a1ceb1ed61621d +gbrapf32be 366b804d5697276e8c481c4bdf05a00b +gbrapf32le 558a268e6d6b907449d1056afab78f29 gbrp dc3387f925f972c61aae7eb23cdc19f0 gbrp10be 0277d4c3a8498d75e2783fb81379e481 gbrp10le f3d70f8ab845c3c9b8f7452e4a6e285a @@ -38,8 +38,8 @@ gbrp16be 5fc826cfabebfc1442cb793c4b6303e2 gbrp16le 1b3e0b63d47a3e1b6b20931316883bf2 gbrp9be d9c88968001e1452ff31fbc8d16b18a0 gbrp9le 2ccfed0816bf6bd4bb3a5b7591d9603a -gbrpf32be 4614d32e4417f80e0adcc1bdcf6cde42 -gbrpf32le 1366ee77e5559672260bbe51040e28b2 +gbrpf32be f3d0cefdf11c861001880772d817aac8 +gbrpf32le 290468205c1c18a0667edfca45061aee gray 221201cc7cfc4964eacd8b3e426fd276 gray10be 9452756d0b37f4f5c7cae7635e22d747 gray10le 37fd2e1ec6b66410212d39a342e864df diff --git a/tests/ref/fate/sws-floatimg-cmp b/tests/ref/fate/sws-floatimg-cmp index 24204254c4..cf6788fc23 100644 --- a/tests/ref/fate/sws-floatimg-cmp +++ b/tests/ref/fate/sws-floatimg-cmp @@ -1,120 +1,120 @@ gbrpf32le -> yuv444p16le -> gbrpf32le -avg diff: 0.003852 +avg diff: 0.000125 min diff: 0.000000 -max diff: 0.006638 +max diff: 0.000501 gbrpf32le -> yuv444p -> gbrpf32le -avg diff: 0.004316 +avg diff: 0.001804 min diff: 0.000000 -max diff: 0.012704 +max diff: 0.006399 gbrpf32le -> yuv444p9le -> gbrpf32le -avg diff: 0.004053 -min diff: 0.000001 -max diff: 0.009402 +avg diff: 0.000906 +min diff: 0.000000 +max diff: 0.003313 gbrpf32le -> yuv444p10le -> gbrpf32le -avg diff: 0.003960 +avg diff: 0.000467 min diff: 0.000000 -max diff: 0.008123 +max diff: 0.001912 gbrpf32le -> yuv444p12le -> gbrpf32le -avg diff: 0.003878 +avg diff: 0.000166 min diff: 0.000000 -max diff: 0.007011 +max diff: 0.000802 gbrpf32le -> yuv444p14le -> gbrpf32le -avg diff: 0.003868 +avg diff: 0.000127 min diff: 0.000000 -max diff: 0.006729 +max diff: 0.000524 gbrpf32le -> rgb24 -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> bgr24 -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> rgba -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> bgra -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> argb -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> abgr -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> 0rgb -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> 0bgr -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> rgb0 -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> bgr0 -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> rgb48le -> gbrpf32le -avg diff: 0.003851 +avg diff: 0.000249 min diff: 0.000000 -max diff: 0.007076 +max diff: 0.000990 gbrpf32le -> bgr48le -> gbrpf32le -avg diff: 0.003851 +avg diff: 0.000249 min diff: 0.000000 -max diff: 0.007076 +max diff: 0.000990 gbrpf32le -> rgba64le -> gbrpf32le -avg diff: 0.003851 +avg diff: 0.000249 min diff: 0.000000 -max diff: 0.007076 +max diff: 0.000990 gbrpf32le -> bgra64le -> gbrpf32le -avg diff: 0.003851 +avg diff: 0.000249 min diff: 0.000000 -max diff: 0.007076 +max diff: 0.000990 gbrpf32le -> gbrp -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> gbrap -> gbrpf32le -avg diff: 0.004122 +avg diff: 0.001011 min diff: 0.000000 -max diff: 0.008975 +max diff: 0.004229 gbrpf32le -> gbrp9le -> gbrpf32le -avg diff: 0.007737 +avg diff: 0.003917 min diff: 0.000000 -max diff: 0.014009 +max diff: 0.007870 gbrpf32le -> gbrp10le -> gbrpf32le -avg diff: 0.007662 +avg diff: 0.003841 min diff: 0.000000 -max diff: 0.013605 +max diff: 0.007456 gbrpf32le -> gbrap10le -> gbrpf32le -avg diff: 0.007662 +avg diff: 0.003841 min diff: 0.000000 -max diff: 0.013605 +max diff: 0.007456 gbrpf32le -> gbrp12le -> gbrpf32le -avg diff: 0.007622 +avg diff: 0.003796 min diff: 0.000000 -max diff: 0.013335 +max diff: 0.007140 gbrpf32le -> gbrap12le -> gbrpf32le -avg diff: 0.007622 +avg diff: 0.003796 min diff: 0.000000 -max diff: 0.013335 +max diff: 0.007140 gbrpf32le -> gbrp14le -> gbrpf32le -avg diff: 0.007620 +avg diff: 0.003792 min diff: 0.000000 -max diff: 0.013232 +max diff: 0.007034 gbrpf32le -> gbrp16le -> gbrpf32le -avg diff: 0.007680 +avg diff: 0.003853 min diff: 0.000000 -max diff: 0.013275 +max diff: 0.007098 gbrpf32le -> gbrap16le -> gbrpf32le -avg diff: 0.007680 +avg diff: 0.003853 min diff: 0.000000 -max diff: 0.013275 +max diff: 0.007098