NNEDI3CL - Revision history

Asd: NNEDI3CL 1.0.7

2023-05-18T07:39:33Z

NNEDI3CL 1.0.7

Reel.Deal: /* Requirements */

2022-11-23T06:29:36Z

‎Requirements

Reel.Deal: /* Requirements */ nnedi3_weights.bin download info

2022-11-21T19:14:10Z

‎Requirements: nnedi3_weights.bin download info

Reel.Deal: v1.0.5 update

2022-11-20T02:56:05Z

v1.0.5 update

Reel.Deal: add field note

2022-09-05T06:33:55Z

add field note

Reel.Deal: v1.0.3

2022-08-16T13:35:31Z

v1.0.3

Reel.Deal: /* Examples */ add examples

2022-07-04T13:49:32Z

‎Examples: add examples

Reel.Deal: NNEDI3CL documentation

2022-06-29T00:17:34Z

NNEDI3CL documentation

New page

{{FilterCat6|External_filters|Plugins|Plugins_x64|Restoration_filters|Deinterlacers|Deep_color_tools}}
{{Filter3
|1={{Author/Asd-g}}
|2=1.0.1
|3=[https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/ NNEDI3CL-1.0.1.7z]
|4=Deinterlacing
|5=[https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]
|6=}}

== Description ==
[[NNEDI3CL]] is an [https://en.wikipedia.org/wiki/OpenCL OpenCL] implementation of [[nnedi3]]. It is an intra-field only deinterlacer. It takes a frame, throws away one field, and then interpolates the missing pixels using only information from the remaining field. It has same rate and double rate modes. NNEDI3 is also very good for enlarging images by powers of 2.

This is [https://github.com/HomeOfVapourSynthEvolution/VapourSynth-NNEDI3CL a port of the VapourSynth plugin NNEDI3CL].
 
 
== Requirements ==
* [x86] / [x64]: [[AviSynth+]]
* Supported color formats: All [[planar]] formats (8/10/12/14/16/32-bit) are supported.
* The file <code>nnedi3_weights.bin</code> is required. It must be located in the same folder as NNEDI3CL.
 
== [[Script variables|Syntax and Parameters]] ==
:{{Template:FuncDef|NNEDI3CL (clip input, int field, bool "dh", bool "dw", int[] "planes", int "nsize", int "nns", int "qual", int "etype", int "pscrn", int "device", bool "list_device", bool "info")}}
 
::{{Par2|input|clip|}}
:::Input clip; it must be in 8..32-bit planar format.
 
::{{Par2|field|int|}}
:::Controls the mode of operation (double vs same rate) and which field is kept.
:::*0 = same rate, keep bottom field
:::*1 = same rate, keep top field
:::*2 = double rate (alternates each frame), starts with bottom
:::*3 = double rate (alternates each frame), starts with top
 
::{{Par2|dh|bool|false}}
:::Doubles the height of the input. Each line of the input is copied to every other line of the output and the missing lines are interpolated.
:::If {{Template:FuncDef|field}}=0, the input is copied to the odd lines of the output. If {{Template:FuncDef|field}}=1, the input is copied to the even lines of the output.
:::{{Template:FuncDef|field}} must be set to either 0 or 1 when using {{Template:FuncDef|dh}}=true.
 
::{{Par2|dw|bool|false}}
::: Doubles the width of the input.
:::{{Template:FuncDef|field}} must be set to either 0 or 1 when using {{Template:FuncDef|dh}}=true.
 
::{{Par2|planes|int array|[0,1,2]}}
:::Sets which planes will be processed.
:::Planes that are not processed will contain uninitialized memory.
 
::{{Par2|nsize|int|6}}
:::Sets the size of the local neighborhood around each pixel that is used by the predictor neural network.
:::For image enlargement it is recommended to use 0 or 4. Larger y_diameter settings will result in sharper output. For deinterlacing, larger x_diameter settings will allow connecting lines of smaller slope.
:::However, what setting to use really depends on the amount of aliasing (lost information) in the source. If the source was heavily low-pass filtered before interlacing then aliasing will be low and a large x_diameter setting wont be needed, and vice versa.
:::Possible settings (x_diameter x y_diameter):
:::*0 - 8x6
:::*1 - 16x6
:::*2 - 32x6
:::*3 - 48x6
:::*4 - 8x4
:::*5 - 16x4
:::*6 - 32x4
 
::{{Par2|nns|int|1}}
:::Sets the number of neurons in the predictor neural network.
:::0 is fastest. 4 is slowest, but should give the best quality.
:::This is a quality vs speed option; however, differences are usually small. The difference in speed will become larger as '{{Template:FuncDef|qual}}' is increased.
:::*0 - 16
:::*1 - 32
:::*2 - 64
:::*3 - 128
:::*4 - 256
 
::{{Par2|qual|int|1}}
:::Controls the number of different neural network predictions that are blended together to compute the final output value.
:::Each neural network was trained on a different set of training data. Blending the results of these different networks improves generalization to unseen data.
:::Possible values are 1 or 2. Essentially this is a quality vs speed option. Larger values will result in more processing time, but should give better results.
:::However, the difference is usually pretty small. I would recommend using {{Template:FuncDef|qual}}>1 for things like single image enlargement.
 
::{{Par2|etype|int|0}}
:::Controls which set of weights to use in the predictor nn.
:::*0 - weights trained to minimize absolute error
:::*1 - weights trained to minimize squared error
 
::{{Par2|pscrn|int|2}}
:::Controls whether or not the prescreener neural network is used to decide which pixels should be processed by the predictor neural network and which can be handled by simple cubic interpolation.
:::The prescreener is trained to know whether cubic interpolation will be sufficient for a pixel or whether it should be predicted by the predictor nn. The computational complexity of the prescreener nn is much less than that of the predictor nn.
:::Since most pixels can be handled by cubic interpolation, using the prescreener generally results in much faster processing. The prescreener is pretty accurate, so the difference between using it and not using it is almost always unnoticeable.
:::The new prescreener is faster than the old one, and it also causes more pixels to be handled by cubic interpolation.
:::*1 - old prescreener
:::*2 - new prescreener (unavailable with float input)
 
::{{Par2|device|int|}}
:::Sets target OpenCL device.
:::Use list_device to get the index of the available devices.
:::By default the default device is selected
 
::{{Par2|list_device|bool|false}}
::: Whether to draw the devices list on the frame.
 
::{{Par2|info|bool|false}}
::: Whether to draw the OpenCL-related info on the frame.
 
==Examples==
NNEDI3CL with default settings:
[[AviSource]]("Blah.avi")
NNEDI3CL(2, dh=false, dw=false, planes=[0,1,2], nsize=6, nns=1, qual=1, etype=0, pscrn=2, list_device=false, info=false)
 
== Changelog ==
Version Date Changes 
v1.0.1 2022/06/26 - Properly handle the clips with one plane only.
v1.0.0 2022/06/26 - Initial release.
 
== External Links ==
*[https://github.com/Asd-g/AviSynthPlus-NNEDI3CL GitHub] - Source code repository.
 
 
-----------------------------------------------
'''Back to [[External_filters#Deinterlacing|External Filters]] ←'''

@@ Line 2: / Line 2: @@
 {{Filter3
 |1={{Author/Asd-g}}
-|2=1.0.5
+|2=1.0.7
-|3=[https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/ NNEDI3CL-1.0.5.7z]
+|3=[https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/ NNEDI3CL-1.0.7.7z]
 |4=Deinterlacing
 |5=[https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]
@@ Line 21: / Line 21: @@
 == [[Script variables|Syntax and Parameters]] ==
-:{{Template:FuncDef|NNEDI3CL (clip input, int "field", bool "dh", bool "dw", int_array "planes", int "nsize", int "nns", int "qual", int "etype", int "pscrn", int "device", bool "list_device", bool "info", bool "st")}}
+:{{Template:FuncDef|NNEDI3CL (clip input, int "field", bool "dh", bool "dw", int_array "planes", int "nsize", int "nns", int "qual", int "etype", int "pscrn", int "device", bool "list_device", bool "info", bool "st", bool "luma")}}
 <br>
 ::{{Par2|input|clip|}}
@@ Line 114: / Line 114: @@
 :::In some cases using <code>NNEDI3CL</code> and <code>prefetch()</code> could cause very high cpu usage. In these cases <code>st=true</code> could help without forcing <code>[[MT_SERIALIZED]]</code> mode.
 :::Default: Auto determined by device properties.
 <br>
 ==Examples==
@@ Line 130: / Line 135: @@
 == Changelog ==
   Version      Date            Changes<br>
   v1.0.5       2022/11/18      - Changed the check of field parity.
                                - Add support for alpha plane.

@@ Line 17: / Line 17: @@
 * [x64]: [[AviSynth+]]
 * Supported color formats: all [[planar]] formats (8/10/12/14/16/32bit, Y/YUV/RGB with or without alpha) are supported.
-* The file <code>nnedi3_weights.bin</code> is required. It must be located in the same folder as NNEDI3CL. <br>Download it from [https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/tag/1.0.3 NNEDI3CL 1.0.3] or directly from here: [https://github.com/dubhater/vapoursynth-nnedi3/raw/master/src/nnedi3_weights.bin nnedi3_weights.bin]
+* The file <code>nnedi3_weights.bin</code> is required. It must be located in the same folder as NNEDI3CL. <br>Download it from [https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/tag/1.0.3 NNEDI3CL 1.0.3] or directly from here: [https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/raw/main/src/nnedi3_weights.bin nnedi3_weights.bin]
 <br>

@@ Line 17: / Line 17: @@
 * [x64]: [[AviSynth+]]
 * Supported color formats: all [[planar]] formats (8/10/12/14/16/32bit, Y/YUV/RGB with or without alpha) are supported.
-* The file <code>nnedi3_weights.bin</code> is required. It must be located in the same folder as NNEDI3CL.
+* The file <code>nnedi3_weights.bin</code> is required. It must be located in the same folder as NNEDI3CL. <br>Download it from [https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/tag/1.0.3 NNEDI3CL 1.0.3] or directly from here: [https://github.com/dubhater/vapoursynth-nnedi3/raw/master/src/nnedi3_weights.bin nnedi3_weights.bin]
 <br>
 == [[Script variables|Syntax and Parameters]] ==
 :{{Template:FuncDef|NNEDI3CL (clip input, int "field", bool "dh", bool "dw", int_array "planes", int "nsize", int "nns", int "qual", int "etype", int "pscrn", int "device", bool "list_device", bool "info", bool "st")}}

@@ Line 31: / Line 31: @@
 :::*2 = double rate (alternates each frame), starts with bottom
 :::*3 = double rate (alternates each frame), starts with top
 <br>
 ::{{Par2|dh|bool|false}}
@@ Line 106: / Line 107: @@
 Deinterlace:
   [[AviSource]]("Blah.avi")
-  NNEDI3CL(1, dh=true) # single rate TFF
+  NNEDI3CL(1) # single rate TFF
-  #NNEDI3CL(3, dh=true) # double rate TFF
+  #NNEDI3CL(3) # double rate TFF
 Enlarge by 2x and correct for center shift using Spline36Resize:

@@ Line 2: / Line 2: @@
 {{Filter3
 |1={{Author/Asd-g}}
-|2=1.0.3
+|2=1.0.5
-|3=[https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/ NNEDI3CL-1.0.3.7z]
+|3=[https://github.com/Asd-g/AviSynthPlus-NNEDI3CL/releases/ NNEDI3CL-1.0.5.7z]
 |4=Deinterlacing
 |5=[https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]
@@ Line 15: / Line 15: @@
 <br>
 == Requirements ==
-* [x86] / [x64]: [[AviSynth+]]
+* [x64]: [[AviSynth+]]
-* Supported color formats: All [[planar]] formats (8/10/12/14/16/32-bit) are supported.
+* Supported color formats: all [[planar]] formats (8/10/12/14/16/32bit, Y/YUV/RGB with or without alpha) are supported.
 * The file <code>nnedi3_weights.bin</code> is required. It must be located in the same folder as NNEDI3CL.
 <br>
 == [[Script variables|Syntax and Parameters]] ==
-:{{Template:FuncDef|NNEDI3CL (clip input, int field, bool "dh", bool "dw", int[] "planes", int "nsize", int "nns", int "qual", int "etype", int "pscrn", int "device", bool "list_device", bool "info", bool "st")}}
+:{{Template:FuncDef|NNEDI3CL (clip input, int "field", bool "dh", bool "dw", int_array "planes", int "nsize", int "nns", int "qual", int "etype", int "pscrn", int "device", bool "list_device", bool "info", bool "st")}}
 <br>
 ::{{Par2|input|clip|}}
@@ Line 27: / Line 27: @@
 ::{{Par2|field|int|}}
 :::Controls the mode of operation (double vs same rate) and which field is kept.
 :::*0 = same rate, keep bottom field
 :::*1 = same rate, keep top field
 :::*2 = double rate (alternates each frame), starts with bottom
 :::*3 = double rate (alternates each frame), starts with top
-:::Note: If <code>_FieldBased</code> property (if exists) is > 0, this is internally set to - <code>_FieldBased=1</code> => <code>field=0</code> and <code>_FieldBased=2</code> => <code>field=1</code>.
+:::Default: -1
 <br>
 ::{{Par2|dh|bool|false}}
@@ Line 37: / Line 39: @@
 :::If {{Template:FuncDef|field}}=0, the input is copied to the odd lines of the output. If {{Template:FuncDef|field}}=1, the input is copied to the even lines of the output.
 :::{{Template:FuncDef|field}} must be set to either 0 or 1 when using {{Template:FuncDef|dh}}=true.
 <br>
 ::{{Par2|dw|bool|false}}
 ::: Doubles the width of the input.
 :::{{Template:FuncDef|field}} must be set to either 0 or 1 when using {{Template:FuncDef|dh}}=true.
 <br>
-::{{Par2|planes|int array|[0,1,2]}}
+::{{Par2|planes|int_array|[0,1,2,3]}}
 :::Sets which planes will be processed.
 :::Planes that are not processed will contain uninitialized memory.
 <br>
 ::{{Par2|nsize|int|6}}
@@ Line 58: / Line 63: @@
 :::*5 - 16x4
 :::*6 - 32x4
 <br>
 ::{{Par2|nns|int|1}}
@@ Line 68: / Line 74: @@
 :::*3 - 128
 :::*4 - 256
 <br>
 ::{{Par2|qual|int|1}}
@@ Line 74: / Line 81: @@
 :::Possible values are 1 or 2. Essentially this is a quality vs speed option. Larger values will result in more processing time, but should give better results.
 :::However, the difference is usually pretty small. I would recommend using {{Template:FuncDef|qual}}>1 for things like single image enlargement.
 <br>
 ::{{Par2|etype|int|0}}
@@ Line 79: / Line 87: @@
 :::*0 - weights trained to minimize absolute error
 :::*1 - weights trained to minimize squared error
 <br>
 ::{{Par2|pscrn|int|2}}
@@ Line 87: / Line 96: @@
 :::*1 - old prescreener
 :::*2 - new prescreener (unavailable with float input)
 <br>
 ::{{Par2|device|int|}}
@@ Line 100: / Line 110: @@
 <br>
 ::{{Par2|st|bool|false}}
-:::Whether to run always in single thread even if [[SetFilterMTMode#Prefetch|prefetch()]] is used.
+:::Whether to run always in single thread even if [[SetFilterMTMode#Prefetch|Prefetch()]] is used.
-:::In some cases using <code>NNEDI3CL</code> and <code>prefetch()</code> could cause very high cpu usage. In these cases <code>st=true</code> could help without forcing <code>MT_SERIALIZED</code> mode.
+:::In some cases using <code>NNEDI3CL</code> and <code>prefetch()</code> could cause very high cpu usage. In these cases <code>st=true</code> could help without forcing <code>[[MT_SERIALIZED]]</code> mode.
-:::Default: false.
+:::Default: Auto determined by device properties.
 <br>
 ==Examples==
@@ Line 119: / Line 129: @@
 == Changelog ==
   Version      Date            Changes<br>
   v1.0.3       2022/08/15      - Added parameter st.
   v1.0.2       2022/07/13      - Fixed memory leak when list_device=true/info=true.