Nnedi3/nnedi3 rpow2 - Revision history

FranceBB: Added options to set AVX512 in the optimizations settings

2025-08-05T08:48:55Z

Added options to set AVX512 in the optimizations settings

← Older revision		Revision as of 10:48, 5 August 2025
Line 107:		Line 107:
	::::*6 = use FMA3		::::*6 = use FMA3
	::::*7 = use FMA4		::::*7 = use FMA4
			::::*8 = use AVX512
	<br>		<br>
	::{{Par2\|fapprox\|int\|15}}		::{{Par2\|fapprox\|int\|15}}

FranceBB: /* Syntax and Parameters */

2024-07-23T08:05:46Z

Syntax and Parameters

← Older revision		Revision as of 10:05, 23 July 2024
Line 51:		Line 51:
	:::Possible values are 1 or 2. When 1 is selected, a single network is gonna be used, while when 2 is selected, the results of two networks are blended together.		:::Possible values are 1 or 2. When 1 is selected, a single network is gonna be used, while when 2 is selected, the results of two networks are blended together.
	:::In some cases, 2 can provide better results, however the difference is pretty small. The downside is that it can lead to hallucinations.		:::In some cases, 2 can provide better results, however the difference is pretty small. The downside is that it can lead to hallucinations.
	::: The recommended setting is therefore to set {{Template:FuncDef\|qual}} to 1 unless you know what you're doing.		:::The recommended setting is therefore to set {{Template:FuncDef\|qual}} to 1 unless you know what you're doing.
	<br>		<br>
	::{{Par2\|etype\|int\|0}}		::{{Par2\|etype\|int\|0}}

FranceBB: /* Syntax and Parameters */

2024-07-23T08:05:37Z

Syntax and Parameters

← Older revision		Revision as of 10:05, 23 July 2024
Line 47:		Line 47:
	::{{Par2\|qual\|int\|1}}		::{{Par2\|qual\|int\|1}}
	:::Controls the number of different neural network predictions that are blended together to compute the final output value.		:::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.		:::Each neural network was trained on a different set of training data. Blending the results of these different networks should in theory improve 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.		:::In practice, however, blending the results of two networks can lead to hallucinations and therefore introduce artifacts.
	:::~~However~~, the difference is ~~usually~~ pretty small. ~~I would recommend using~~ {{Template:FuncDef\|qual}}>1 ~~for things like single image enlargement~~.		:::Possible values are 1 or 2. When 1 is selected, a single network is gonna be used, while when 2 is selected, the results of two networks are blended together.
			:::In some cases, 2 can provide better results, however the difference is pretty small. The downside is that it can lead to hallucinations.
			::: The recommended setting is therefore to set {{Template:FuncDef\|qual}} to 1 unless you know what you're doing.
	<br>		<br>
	::{{Par2\|etype\|int\|0}}		::{{Par2\|etype\|int\|0}}

FranceBB: /* Syntax and Parameters */

2024-05-12T14:16:04Z

Syntax and Parameters

← Older revision		Revision as of 16:16, 12 May 2024
Line 100:		Line 100:
	::::*1 = use C		::::*1 = use C
	::::*2 = use [[SSE2]]		::::*2 = use [[SSE2]]
			::::*3 = use SSE4.1
			::::*4 = use AVX
			::::*5 = use AVX2
			::::*6 = use FMA3
			::::*7 = use FMA4
	<br>		<br>
	::{{Par2\|fapprox\|int\|15}}		::{{Par2\|fapprox\|int\|15}}
Line 111:		Line 116:
	::::*&12 = 8\|12 = use faster (and more inaccurate) exp function approximation in predictor nn		::::*&12 = 8\|12 = use faster (and more inaccurate) exp function approximation in predictor nn
	<br>		<br>

	== Examples ==		== Examples ==
	*Enlarge image by 4x with default settings (does not correct for center shift):		*Enlarge image by 4x with default settings (does not correct for center shift):

Reel.Deal: /* Examples */

2014-10-12T10:33:49Z

Examples

← Older revision		Revision as of 12:33, 12 October 2014
Line 112:		Line 112:
	<br>		<br>
	== Examples ==		== Examples ==
	*Enlarge image by 4x with default settings (~~identical to the 2nd example~~):		*Enlarge image by 4x with default settings (does not correct for center shift):
	[[AviSource]]("Blah.avi")		[[AviSource]]("Blah.avi")
	nnedi3_rpow2(rfactor=4, nsize=0, nns=3, qual=1, etype=0, pscrn=2, threads=0, opt=0, fapprox=15)		nnedi3_rpow2(rfactor=4, nsize=0, nns=3, qual=1, etype=0, pscrn=2, threads=0, opt=0, fapprox=15)<br>
	<br>		# identical to:
	*Enlarge image by 4x but do not correct for center shift.		# nnedi3_rpow2(rfactor=4)
	~~[[AviSource]]("Blah.avi")~~
	nnedi3_rpow2(rfactor=4)
	<br>		<br>
	*Enlarge image by 2x and correct for center shift using [[Resize#Spline_based_resizers\|Spline36Resize]].		*Enlarge image by 2x and correct for center shift using [[Resize#Spline_based_resizers\|Spline36Resize]].

Reel.Deal: /* Examples */

2014-10-12T10:24:04Z

Examples

← Older revision		Revision as of 12:24, 12 October 2014
Line 122:		Line 122:
	*Enlarge image by 2x and correct for center shift using [[Resize#Spline_based_resizers\|Spline36Resize]].		*Enlarge image by 2x and correct for center shift using [[Resize#Spline_based_resizers\|Spline36Resize]].
	[[AviSource]]("Blah.avi")		[[AviSource]]("Blah.avi")
	nnedi3_rpow2(rfactor=2,cshift="Spline36Resize")		nnedi3_rpow2(rfactor=2,cshift="Spline36Resize")<br>
			# identical to:
			# nnedi3_rpow2(rfactor=2)
			# Spline36Resize([[Clip_properties\|Width()]], [[Clip_properties\|Height()]], src_left=-0.5, src_top=-0.5)

	<br>		<br>
	*Enlarge image by 8x, correct for center shift and downsample from 8x to 7x using [[Resize#LanczosResize\|LanczosResize]] with 5 taps.		*Enlarge image by 8x, correct for center shift and downsample from 8x to 7x using [[Resize#LanczosResize\|LanczosResize]] with 5 taps.
	[[AviSource]]("Blah.avi")		[[AviSource]]("Blah.avi")
	nnedi3_rpow2(rfactor=8,cshift="LanczosResize",fwidth=~~width~~7,fheight=~~height~~7,ep0=5)		nnedi3_rpow2(rfactor=8,cshift="LanczosResize",fwidth=Width()7,fheight=Height()7,ep0=5)<br>
			# identical to:
			# nnedi3_rpow2(rfactor=8)
			# LanczosResize(Width()[[Operators\|/]]8[[Operators\|]]7, Height()[[Operators\|/]]8[[Operators\|]]7, src_left=-0.5, src_top=-0.5, taps=5)
	<br>		<br>
	--------------------------------------------------------		--------------------------------------------------------
	'''Back to [[nnedi3#Filters\|nnedi3]] ←'''		'''Back to [[nnedi3#Filters\|nnedi3]] ←'''

Reel.Deal: minor update

2014-09-21T17:59:05Z

minor update

← Older revision		Revision as of 19:59, 21 September 2014
Line 1:		Line 1:
	~~[[Category:Plugin_Filters]]~~		{{FilterCat4\|External_filters\|Plugin_functions\|Adjustment_filters\|Resizers}}
	~~[[Category:~~Resizers]]
	'''Back to [[nnedi3#Filters\|nnedi3]] ←'''		'''Back to [[nnedi3#Filters\|nnedi3]] ←'''
	--------------------------------------------------		--------------------------------------------------
Line 6:		Line 5:
	== Description ==		== Description ==
	nnedi3_rpow2 is for enlarging images by powers of 2.<br>		nnedi3_rpow2 is for enlarging images by powers of 2.<br>
			<br>
			== Requirements ==
			* [[FAQ_different_types_content#How_do_I_recognize_progressive.2C_interlaced.2C_telecined.2C_hybrid_and_blended_content.3F\|Progressive]] input only
			* Supported color formats: [[RGB24]], [[YUY2]], [[YV12]]
	<br>		<br>
	== [[Script variables\|Syntax and Parameters]] ==		== [[Script variables\|Syntax and Parameters]] ==

Reel.Deal: reformat

2014-09-17T04:27:37Z

reformat

Show changes

Reel.Deal: category

2014-08-23T16:17:31Z

Reel.Deal: /* Description */

2014-04-13T12:13:02Z

Description

← Older revision		Revision as of 14:13, 13 April 2014
Line 4:		Line 4:
	<br>		<br>
	== Description ==		== Description ==
	:~~nnedi3~~ is an intra-field only deinterlacer. It takes in a frame, throws away one field, and then interpolates the missing pixels using only information from the kept field. It has same rate and double rate modes, and works with YV12, YUY2, and RGB24 input.		:nnedi3_rpow2 is for enlarging images by powers of 2.
	<br>		<br>

	== [[Script variables\|Syntax and Parameters]] ==		== [[Script variables\|Syntax and Parameters]] ==
	:{{Template:FuncDef\|nnedi3_rpow2 (''clip'', ''int'' rfactor, ''int'' "nsize", ''int'' "nns", ''int'' "qual", ''int'' "etype", ''int'' "pscrn", ''string'' "cshift", ''int'' "fwidth", ''int'' "fheight", ''float'' "ep0", ''float'' "ep1", ''int'' "threads", ''int'' "opt", ''int'' "fapprox")}}		:{{Template:FuncDef\|nnedi3_rpow2 (''clip'', ''int'' rfactor, ''int'' "nsize", ''int'' "nns", ''int'' "qual", ''int'' "etype", ''int'' "pscrn", ''string'' "cshift", ''int'' "fwidth", ''int'' "fheight", ''float'' "ep0", ''float'' "ep1", ''int'' "threads", ''int'' "opt", ''int'' "fapprox")}}

← Older revision		Revision as of 18:17, 23 August 2014
Line 1:		Line 1:
			[[Category:Plugin_Filters]]
			[[Category:Resizers]]
	'''Back to [[nnedi3#Filters\|nnedi3]] ←'''		'''Back to [[nnedi3#Filters\|nnedi3]] ←'''
	--------------------------------------------------		--------------------------------------------------