MedianBlur2 - Revision history

Pinterf: MedianBlur2: update version

2021-03-09T11:13:33Z

MedianBlur2: update version

Pinterf: /* Performance */ special cases only for 8 bit

2021-03-09T11:12:51Z

‎Performance: special cases only for 8 bit

Pinterf: /* Performance */ Update with TemporalMedianBlur special cases

2021-03-09T11:10:50Z

‎Performance: Update with TemporalMedianBlur special cases

Pinterf: /* Changelog */ fix changelog

2021-03-09T11:05:38Z

‎Changelog: fix changelog

Pinterf: /* Changelog */

2021-03-09T11:04:39Z

‎Changelog

Reel.Deal: add Deep_color_tools category

2020-06-25T20:12:48Z

add Deep_color_tools category

Reel.Deal: v1.0 minor touchup

2020-06-02T23:09:18Z

v1.0 minor touchup

Pinterf: /* Changelog */ update to v1.0

2020-04-02T09:21:28Z

‎Changelog: update to v1.0

Pinterf: /* MedianBlurTemporal */ RGB info, alpha plane info

2020-04-02T09:18:54Z

‎MedianBlurTemporal: RGB info, alpha plane info

Pinterf: /* MedianBlur */ RGB info, alpha plane info

2020-04-02T09:16:38Z

‎MedianBlur: RGB info, alpha plane info

@@ Line 2: / Line 2: @@
 {{Filter3
 | {{Author/tp7}}, {{Author/pinterf}}
-| v1.0
+| v1.1
-| [https://github.com/pinterf/MedianBlur2/releases MedianBlur2_v1.0.7z]
+| [https://github.com/pinterf/MedianBlur2/releases MedianBlur2_v1.1.7z]
 | Blurring
 | [https://opensource.org/licenses/MIT MIT] but binaries are [https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]

@@ Line 16: / Line 16: @@
 Unlike the [[MedianBlur|old MedianBlur]], this implementation has constant runtime complexity, meaning that theoretical performance is the same for any radius. Implementation is based on [http://nomis80.org/ctmf.html this paper], but includes some additional optimizations.
-#When [[SSE2]] is available and radius is 1 or 2, special routines are used. They're a lot faster than the generic one for these radii.
+#When [[SSE2]] is available and radius is 1 or 2, special routines are used, but only for 8 bit videos. They're a lot faster than the generic one for these radii.
-#For 2 < radius < 8, generic approach with 8-bit bin size is used. Using 8-bit bins makes histogram addition/subtraction faster with [[SSE2]] (8 bit videos).
+#For 2 < radius < 8, generic approach with 8-bit bin size is used. Using 8-bit bins makes histogram addition/subtraction faster with [[SSE2]]
 #For large radii, 16-bit bins are used, as described in the paper.
 #TemporalMedianBlur is using 32 bit bins

@@ Line 17: / Line 17: @@
 #When [[SSE2]] is available and radius is 1 or 2, special routines are used. They're a lot faster than the generic one for these radii.
-#For 2 < radius < 8, generic approach with 8-bit bin size is used. Using 8-bit bins makes histogram addition/subtraction faster with [[SSE2]].
+#For 2 < radius < 8, generic approach with 8-bit bin size is used. Using 8-bit bins makes histogram addition/subtraction faster with [[SSE2]] (8 bit videos).
 #For large radii, 16-bit bins are used, as described in the paper.
 In other words, you can expect huge performance drop when going from 1 to 2, not so huge but still large from 2 to 3 and a noticeable slowdown from 7 to 8. Between them the fps should be constant and it actually might get a bit faster with larger radius. Performance with radius > 2 also depends on the actual frame content. Processing [[ColorBars]]() is a lot faster than [[AddGrainC]](10000). It is quite possible that you'll get worse performance for radius = 3 with this plugin compared to the original MedianBlur. All other radii should be (a lot) faster.
@@ Line 30: / Line 32: @@
 <br>
 <br>
 == Requirements ==
 * [x86]: [[AviSynth+]] or [https://sourceforge.net/projects/avisynth2/ AviSynth 2.6]

@@ Line 112: / Line 112: @@
 == Changelog ==
   Version      Date            Changes<br>
-−
-### Change log ###
   v1.1         01/30/2021     (pinterf)
                                - Speed: SSE2 and AVX2 for 10+ bits (generic case, MedianBlur)
@@ Line 125: / Line 124: @@
 : SSE4 (disable AVX2)
 : AVX2
-−
   v1.0         04/02/2020      (pinterf)
                                - add high bitdepth support

@@ Line 112: / Line 112: @@
 == Changelog ==
   Version      Date            Changes<br>
   v1.0         04/02/2020      (pinterf)
                                - add high bitdepth support
@@ Line 130: / Line 144: @@
   v0.9         12/08/2013      - Initial release.
 <br>
 == External Links ==
 *[https://forum.doom9.org/showthread.php?t=169832 Doom9 Forum] - MedianBlur2 discussion, more information [http://forum.doom9.org/showthread.php?p=1656988#post1656988 here].

@@ Line 1: / Line 1: @@
-{{FilterCat5|External_filters|Plugins|Plugins_x64|Adjustment_filters|Blurring}}
+{{FilterCat6|External_filters|Plugins|Plugins_x64|Adjustment_filters|Blurring|Deep_color_tools}}
 {{Filter3
 | {{Author/tp7}}, {{Author/pinterf}}
@@ Line 34: / Line 34: @@
 * [x64]: [[AviSynth+]]
 * Supported color formats: [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]]
-** AviSynth+: All planar formats (8/10/12/14/16/32bit, Y/YUV/RGB with or without alpha) are supported.
+** AviSynth+: All [[planar]] formats (8/10/12/14/16/32-bit, Y, YUV(A), and RGB(A) with or without alpha) are supported.
 <br>
 * [https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads Microsoft Visual C++ 2019 Redistributable Package (x86 / x64)]

@@ Line 1: / Line 1: @@
 {{FilterCat5|External_filters|Plugins|Plugins_x64|Adjustment_filters|Blurring}}
 {{Filter3
-| {{Author/tp7}}, [https://github.com/pinterf pinterf]
+| {{Author/tp7}}, {{Author/pinterf}}
-| v1.0, v0.94
+| v1.0
-| [https://github.com/pinterf/MedianBlur2/releases Releases] [https://github.com/tp7/MedianBlur2/releases Releases]
+| [https://github.com/pinterf/MedianBlur2/releases MedianBlur2_v1.0.7z]
 | Blurring
 | [https://opensource.org/licenses/MIT MIT] but binaries are [https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]
@@ Line 23: / Line 23: @@
 <br>
 <br>
-High bit depth considerations:
+'''High bit depth considerations:'''
 At larger bit depths performance drops because internal histogram tables contain statistical data for 10, 12, 14 and 16 bits precision. Also: higher bit depths need more memory.
@@ Line 29: / Line 29: @@
 Radii 1 and 2 are special case only for 8 bits as of v1.0.
 <br>
+<br>
 == Requirements ==
-* [https://sourceforge.net/projects/avisynth2/ AviSynth 2.6.0] or greater
+* [x86]: [[AviSynth+]] or [https://sourceforge.net/projects/avisynth2/ AviSynth 2.6]
-* x64 version requires [https://forum.doom9.org/showthread.php?t=168856 AviSynth+ r1576] or later
+* [x64]: [[AviSynth+]]
-* Supported color formats: all Y and planar on Avisynth+. [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]] on Aviysnth 2.6
+* Supported color formats: [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]]
 <br>
-* Microsoft Visual C++ Redistributable for Visual Studio 2019
+* [https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads Microsoft Visual C++ 2019 Redistributable Package (x86 / x64)]
-:[https://support.microsoft.com/en-gb/help/2977003/the-latest-supported-visual-c-downloads Latest MS VC++ Redistributables]
+:<span style="color:red">***</span> <tt>vcredist_x86.exe</tt> is required for <tt>MedianBlur2-x86</tt>
-::<span style="color:red">***</span> <tt>vcredist_x86.exe</tt> is required for <tt>x86</tt>
+:<span style="color:red">***</span> <tt>vcredist_x64.exe</tt> is required for <tt>MedianBlur2-x64</tt>
-::<span style="color:red">***</span> <tt>vcredist_x64.exe</tt> is required for <tt>x64</tt>
-:or [https://visualstudio.microsoft.com/downloads/ Visual Studio Downloads]
-::nativate to Other Tools and Frameworks|Microsoft Visual C++ Redistributable for Visual Studio 2019
 <br>
@@ Line 56: / Line 54: @@
 ::{{Par2|radiusv|int|2}}
 :::Spatial radius; YUV planes are independently processed.
-:::for RGB clips radiusy value is used for all planes
+:::For RGB clips <code>radiusy</code> value is used for all planes.
-::::*If set to 0 the corresponding plane is copied from the input clip unaltered.
+:::*If set to 0 the corresponding plane is copied from the input clip unaltered.
-::::*If set between -1 and -255 the corresponding plane will be set to -radius. Works only for 8 bits
+:::*If set between -1 and -255 the corresponding plane will be set to -radius. Works only for 8 bits
-::::*If set to any value less than -255 the corresponding plane will be set to zero.
+:::*If set to any value less than -255 the corresponding plane will be set to zero.
-::::*2 is the default for all planes; maximum value is limited to 127.
+:::*2 is the default for all planes; maximum value is limited to 127.
 <br>
+::If alpha plane is available it will be simply copied.
-If alpha plane is available it will be simply copied.
+::Planar RGB uses the value of <code>radiusy</code> for all its channels. <code>radius</code> u and <code>radius</code> v are ignored.
-Planar RGB is using value for radiusy for all its channels. radius u and radius v are ignored.
 <br>
@@ Line 84: / Line 81: @@
 :::Spatial radius; YUV planes are independently processed.
-:::for RGB clips radiusy value is used for all planes
+:::For RGB clips <code>radiusy</code> value is used for all planes.
-::::*Setting it to 0 will just disable spatial filtering.
-::::*Setting it to -1 will copy the plane from the input clip.
+:::*Setting it to 0 will just disable spatial filtering.
-::::*All other values will set the plane to absolute value of radius. Yes, there is no way to set the plane to -1. Works only for 8 bit clips.
+:::*Setting it to -1 will copy the plane from the input clip.
-::::*If set to any value less than -255 the corresponding plane will be set to zero.
+:::*All other values will set the plane to absolute value of radius. Yes, there is no way to set the plane to -1. Works only for 8 bit clips.
-::::*2 is the default for all planes; maximum value is limited to 127.
+:::*If set to any value less than -255 the corresponding plane will be set to zero.
 <br>
@@ Line 95: / Line 93: @@
 :::Value must be greater than 0; the higher the value the slower it gets.
 <br>
+::If alpha plane is available it will be simply copied.
+::Planar RGB uses the value of <code>radiusy</code> for all its channels. <code>radius</code> u and <code>radius</code> v are ignored.
 If alpha plane is available it will be simply copied.
-Planar RGB is using value for radiusy for all its channels. radius u and radius v are ignored.
 <br>
@@ Line 121: / Line 117: @@
                                - add version resource to DLL
                                - move to Visual Studio 2019, ClangCL, v141_xp and v142 configurations
-                               - project hosted at https://github.com/pinterf/MedianBlur2
+                               - project hosted at https://github.com/pinterf/MedianBlur2<br>
   v0.94        02/10/2014      (tp7)
                                - MedianBlurTemporal now produces more correct result around the clip borders (first and last frames).
                                - Both functions allow radii less than -255, corresponding plane is set to zero.
-                                 For compatibility with [[YAHR]] and possible other functions.<br>
+                                 For compatibility with [[YAHR]] and possible other functions.
   v0.93        12/27/2013      - This fixes a huge memory leak in MedianBlurTemp which made it unusable in complex scripts.
                                  The function is also renamed to MedianBlurTemporal, which is probably the final name.<br>
@@ Line 133: / Line 130: @@
   v0.9         12/08/2013      - Initial release.
 <br>
 == External Links ==
 *[https://forum.doom9.org/showthread.php?t=169832 Doom9 Forum] - MedianBlur2 discussion, more information [http://forum.doom9.org/showthread.php?p=1656988#post1656988 here].
 *[http://nomis80.org/ctmf.html Median Filtering in Constant Time] | [https://web.archive.org/web/20141115173428/http://nomis80.org/ctmf.html Mirror]
 <br>
 <br>
 -----------------------------------------------
 '''Back to [[External_filters#Blurring|External Filters]] &larr;'''