FFT3DGPU - Revision history

Reel.Deal: add precision note and fix a typo

2021-09-18T15:48:53Z

add precision note and fix a typo

← Older revision		Revision as of 17:48, 18 September 2021
Line 70:		Line 70:
	<br>		<br>
	::{{Par2\|bordersize\|int\|1}}		::{{Par2\|bordersize\|int\|1}}
	:::Only used with <~~mode~~>mode</~~mode~~> 2. Defines the size of the border.		:::Only used with <code>mode</code> 2. Defines the size of the border.
	<br>		<br>
	::{{Par2\|precision\|int\|0}}		::{{Par2\|precision\|int\|0}}
	:::*0 : to use 16 bit floats(half precision).		:::*0 : to use 16 bit floats(half precision).
	:::* 1 : to use 32 bit float(single precision) for the fft and 16 bit float for the wienner/kalman and sharpening.		:::*1 : to use 32 bit float(single precision) for the fft and 16 bit float for the wienner/kalman and sharpening.
	:::*2 : always use 32 bit floats.		:::*2 : always use 32 bit floats.
	:::Using 16 bit float increases the performance but reduces precision. With a Geforce 7800GT precision=0 is ~1.5 times faster than than mode 2.		:::Using 16 bit float increases the performance but reduces precision. With a Geforce 7800GT precision=0 is ~1.5 times faster than than mode 2.
			:::Note: in FFT3dGPU v0.8.6, for 10-32 bit clips, the parameter "precision" is ignored and is set to 2: always use 32 bit float internally.
	<br>		<br>
	::{{Par2\|NVPerf\|bool\|false}}		::{{Par2\|NVPerf\|bool\|false}}

Reel.Deal: v0.8.6

2021-02-01T19:40:44Z

v0.8.6

← Older revision		Revision as of 21:40, 1 February 2021
Line 1:		Line 1:
	{{~~FilterCat5~~\|External_filters\|Plugins\|Plugins_x64\|Denoisers\|Spatial-Temporal_Denoisers}}		{{FilterCat6\|External_filters\|Plugins\|Plugins_x64\|Denoisers\|Spatial-Temporal_Denoisers\|Deep_color_tools}}
	{{Filter3		{{Filter3
	\| {{Author/tsp}}, {{Author/pinterf}}		\| {{Author/tsp}}, {{Author/pinterf}}
	\| v0.8.5		\| v0.8.6
	\|[https://github.com/pinterf/FFT3dGPU/releases FFT3dGPU-v0.8.5-pfmod.7z]		\|[https://github.com/pinterf/FFT3dGPU/releases FFT3dGPU-v0.8.6-pfmod.7z]
	\| Denoisers		\| Denoisers
	\| [https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]		\| [https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]
Line 20:		Line 20:
	* [x64]: [[AviSynth+]]		* [x64]: [[AviSynth+]]
	* Supported color formats: [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]]		* Supported color formats: [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]]
	** AviSynth+: ~~all~~ [[planar]] 8-bit Y/YUV~~(A)~~/RGB(A) ~~formats~~ are supported		** AviSynth+: All [[planar]] formats (8/10/12/14/16/32-bit, YUV/RGB with or without alpha) are supported.
	<br>		<br>

Line 121:		Line 121:
	== Changelog ==		== Changelog ==
	Version Date Changes<br>		Version Date Changes<br>
			v0.8.6 2021/01/04 - Support 10-32 bit formats. Parameter ranges (smin, smax, sigmas) are the same for all bit depths.
			- For 10-32 bit clips parameter "precision" is ignored and is set to 2: always use 32 bit float internally
			- Note: for 8 bits "precision" default is 0: calculation is done in 16 bit floating point type (s10m5) internally<br>
	v0.8.5 2020/11/01 - Avisynth+ 3.6 additional support: preserve frame properties		v0.8.5 2020/11/01 - Avisynth+ 3.6 additional support: preserve frame properties
	- Fix bt=0 Kalman mode warning in fft3dgpu.hlsl<br>		- Fix bt=0 Kalman mode warning in fft3dgpu.hlsl<br>

Pinterf: New version 0.8.5

2020-11-01T18:28:21Z

New version 0.8.5

← Older revision		Revision as of 20:28, 1 November 2020
Line 2:		Line 2:
	{{Filter3		{{Filter3
	\| {{Author/tsp}}, {{Author/pinterf}}		\| {{Author/tsp}}, {{Author/pinterf}}
	\| v0.8.4		\| v0.8.5
	\|[https://github.com/pinterf/FFT3dGPU/releases FFT3dGPU-v0.8.4-pfmod.7z]		\|[https://github.com/pinterf/FFT3dGPU/releases FFT3dGPU-v0.8.5-pfmod.7z]
	\| Denoisers		\| Denoisers
	\| [https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]		\| [https://www.gnu.org/licenses/gpl-2.0.txt GPLv2]
Line 121:		Line 121:
	== Changelog ==		== Changelog ==
	Version Date Changes<br>		Version Date Changes<br>
			v0.8.5 2020/11/01 - Avisynth+ 3.6 additional support: preserve frame properties
			- Fix bt=0 Kalman mode warning in fft3dgpu.hlsl<br>
	v0.8.4 2018/11/21 - New color spaces besides YV12 and YUY2:		v0.8.4 2018/11/21 - New color spaces besides YV12 and YUY2:
	Y8, YV16, YV411, YV24 (all 8 bits), 8 bit Planar RGB (AviSynth+)		Y8, YV16, YV411, YV24 (all 8 bits), 8 bit Planar RGB (AviSynth+)

Reel.Deal: minor touchup (add archived links)

2020-06-13T03:10:00Z

minor touchup (add archived links)

← Older revision		Revision as of 05:10, 13 June 2020
Line 12:		Line 12:
	FFT3dGPU is a GPU version of Fizick's [[FFT3DFilter]]. The algorithm (Fast Fourier Transform, denoising) is the same for the most part. Currently the following is not implemented: support for noise pattern.		FFT3dGPU is a GPU version of Fizick's [[FFT3DFilter]]. The algorithm (Fast Fourier Transform, denoising) is the same for the most part. Currently the following is not implemented: support for noise pattern.

	In this version the next frame is processed while waiting for the GPU to end it's work. Meaning the filters before fft3dGPU are working concurrently with it. From v0.8.4, the filter automatically registers itself as MT_SERIALIZED for AviSynth+.		In this version the next frame is processed while waiting for the GPU to end it's work. Meaning the filters before fft3dGPU are working concurrently with it. From v0.8.4, the filter automatically registers itself as [[MT_SERIALIZED]] for AviSynth+.
	<br>		<br>
	<br>		<br>
Line 19:		Line 19:
	* [x86]: [[AviSynth+]] or [https://sourceforge.net/projects/avisynth2/ AviSynth 2.6.0]		* [x86]: [[AviSynth+]] or [https://sourceforge.net/projects/avisynth2/ AviSynth 2.6.0]
	* [x64]: [[AviSynth+]]		* [x64]: [[AviSynth+]]
	* Supported color formats: [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]]; YUV(A)		* Supported color formats: [[Y8]], [[YV12]], [[YV16]], [[YV24]], [[YV411]]
			** AviSynth+: all [[planar]] 8-bit Y/YUV(A)/RGB(A) formats are supported
	<br>		<br>

Line 127:		Line 128:
	- source and project files updated to build under Visual Studio 2017		- source and project files updated to build under Visual Studio 2017
	- (15.9) using DX SDK June 2010		- (15.9) using DX SDK June 2010
	- Moved to Avisynth Interface 6 (v2.6) - using headers from the ~~Avisynth~~+ project		- Moved to Avisynth Interface 6 (v2.6) - using headers from the AviSynth+ project
	- added x64 platform<br>		- added x64 platform<br>
	v0.8.2a 2006/08/24 - Last release by tsp, see included docs for full changelog		v0.8.2a 2006/08/24 - Last release by tsp, see included docs for full changelog
			<br>
			== Archived Downloads ==
			{\| class="wikitable" border="1"; width="600px"
			\|-
			!!width="100px"\| Version
			!!width="150px"\| Download
			!!width="150px"\| Mirror
			\|-
			!v0.8.2a
			\|[https://www.dropbox.com/s/b25nik2r4qkujf9/fft3dgpu0.8.2a.exe?dl=1 fft3dgpu0.8.2a.exe]
			\|[https://web.archive.org/web/20200613030343if_/https://files.videohelp.com/u/223002/fft3dgpu0.8.2a.exe fft3dgpu0.8.2a.exe]
			\|-
			!v0.8.2 (x64)
			\|[https://web.archive.org/web/20200613010756if_/https://files.videohelp.com/u/223002/FFT3DGPU_3-15-2010.rar FFT3DGPU_3-15-2010.rar]
			\|[http://www.mediafire.com/file/2chnt1jkwwm/FFT3DGPU_3-15-2010.rar/file FFT3DGPU_3-15-2010.rar]
			\|-
			!v0.8.2
			\|[http://www.avisynth.nl/users/tsp/fft3dgpu0.8.2.7z fft3dgpu0.8.2.7z] /// [http://www.avisynth.nl/users/tsp/fft3dgpu0.8.2.exe fft3dgpu0.8.2.exe]
			\|[https://web.archive.org/web/20170616074504if_/http://avisynth.nl/users/tsp/fft3dgpu0.8.2.7z fft3dgpu0.8.2.7z] /// [https://web.archive.org/web/20170914090422if_/http://www.avisynth.nl/users/tsp/fft3dgpu0.8.2.exe fft3dgpu0.8.2.exe]
			\|}
			*v0.8.2 x64 version compiled by Joshy D.
			<br>
			== External Links ==
	<br>		<br>
	<br>		<br>
	-----------------------------------------------		-----------------------------------------------
	'''Back to [[External_filters#Spatio-Temporal_Denoisers\|External Filters]] ←'''		'''Back to [[External_filters#Spatio-Temporal_Denoisers\|External Filters]] ←'''

Reel.Deal: /* Changelog */

2020-05-13T17:38:45Z

Changelog

← Older revision		Revision as of 19:38, 13 May 2020
Line 122:		Line 122:
	v0.8.4 2018/11/21 - New color spaces besides YV12 and YUY2:		v0.8.4 2018/11/21 - New color spaces besides YV12 and YUY2:
	Y8, YV16, YV411, YV24 (all 8 bits), 8 bit Planar RGB (AviSynth+)		Y8, YV16, YV411, YV24 (all 8 bits), 8 bit Planar RGB (AviSynth+)
	- FFT3dGPU filter registers itself as MT_SERIALIZED automatically		- FFT3dGPU filter registers itself as [[MT_SERIALIZED]] automatically
	for AviSynth+: no need for SetFilterMTMode.<br>		for AviSynth+: no need for SetFilterMTMode.<br>
	v0.8.3 2018/11/18 - project hosted on https://github.com/pinterf/FFT3dGPU		v0.8.3 2018/11/18 - project hosted on https://github.com/pinterf/FFT3dGPU

Reel.Deal: /* Syntax and Parameters */ remove leftover copy paste parameters

2020-04-13T17:24:10Z

Syntax and Parameters: remove leftover copy paste parameters

@@ Line 116: / Line 116: @@
 ::{{Par2|oldfft|bool|false}}
 :::Set to true to use the old fftcode (used in version 0.6.2 and lower) false to use new fft code. If not defined fft3dgpu will use the fastest code.
 <br>

Reel.Deal: /* Description */

2020-03-25T01:10:25Z

Description

← Older revision		Revision as of 03:10, 25 March 2020
Line 12:		Line 12:
	FFT3dGPU is a GPU version of Fizick's [[FFT3DFilter]]. The algorithm (Fast Fourier Transform, denoising) is the same for the most part. Currently the following is not implemented: support for noise pattern.		FFT3dGPU is a GPU version of Fizick's [[FFT3DFilter]]. The algorithm (Fast Fourier Transform, denoising) is the same for the most part. Currently the following is not implemented: support for noise pattern.

	In this version the next frame is processed while waiting for the GPU to end it's work. Meaning the filters before fft3dGPU are working concurrently with it. ~~The filter is not MT-friendly, from~~ v0.8.4 automatically registers itself as MT_SERIALIZED for AviSynth+.		In this version the next frame is processed while waiting for the GPU to end it's work. Meaning the filters before fft3dGPU are working concurrently with it. From v0.8.4, the filter automatically registers itself as MT_SERIALIZED for AviSynth+.
	<br>		<br>
	<br>		<br>

Reel.Deal: /* Changelog */ add newer versions

2020-03-25T01:09:51Z

Changelog: add newer versions

← Older revision		Revision as of 03:09, 25 March 2020
Line 148:		Line 148:
	== Changelog ==		== Changelog ==
	Version Date Changes<br>		Version Date Changes<br>
	v2.1.1 2006/08/24 - Last release by tsp, see included docs for full changelog		v0.8.4 2018/11/21 - New color spaces besides YV12 and YUY2:
			Y8, YV16, YV411, YV24 (all 8 bits), 8 bit Planar RGB (AviSynth+)
			- FFT3dGPU filter registers itself as MT_SERIALIZED automatically
			for AviSynth+: no need for SetFilterMTMode.<br>
			v0.8.3 2018/11/18 - project hosted on https://github.com/pinterf/FFT3dGPU
			- source and project files updated to build under Visual Studio 2017
			- (15.9) using DX SDK June 2010
			- Moved to Avisynth Interface 6 (v2.6) - using headers from the Avisynth+ project
			- added x64 platform<br>
			v0.8.2a 2006/08/24 - Last release by tsp, see included docs for full changelog

Reel.Deal: /* Syntax and Parameters */ typos

2020-03-25T00:12:25Z

Syntax and Parameters: typos

← Older revision		Revision as of 02:12, 25 March 2020
Line 81:		Line 81:
	<br>		<br>
	::{{Par2\|degrid\|float\|1.0}}		::{{Par2\|degrid\|float\|1.0}}
	:::Enables degriding. Only works well with <~~mode~~>mode=1</~~mode~~>. Doesn't degrid the Kalman filter (but it does degrid the sharpening (if enabled) after kalman filter). default 1.0 for <~~mode~~>mode=1</~~mode~~>, 0.0 for <~~mode~~>mode=0</~~mode~~> or 2.		:::Enables degriding. Only works well with <code>mode=1</code>. Doesn't degrid the Kalman filter (but it does degrid the sharpening (if enabled) after kalman filter). default 1.0 for <code>mode=1</code>, 0.0 for <code>mode=0</code> or 2.
	<br>		<br>
	::{{Par2\|scutoff\|float\|0.3}}		::{{Par2\|scutoff\|float\|0.3}}
Line 98:		Line 98:
	::{{Par2\|ow\|int\|bw/2}}		::{{Par2\|ow\|int\|bw/2}}
	::{{Par2\|oh\|int\|bh/2}}		::{{Par2\|oh\|int\|bh/2}}
	:::Overlap width and height. This only works with <code>mode=1</code>. This specifies how big the overlap between the blocks are. Overlap size must be less than or equal to half the blocksize. <code>Ow<c/~~ode~~> must be even.		:::Overlap width and height. This only works with <code>mode=1</code>. This specifies how big the overlap between the blocks are. Overlap size must be less than or equal to half the blocksize. <code>ow</code> must be even.
	<br>		<br>
	::{{Par2\|wintype\|int\|0}}		::{{Par2\|wintype\|int\|0}}

Reel.Deal: FFT3DGPU: add documentation (WIP)

2020-03-24T23:21:18Z

FFT3DGPU: add documentation (WIP)

Show changes