FFT3DGPU
From Avisynth wiki
| Abstract | |
|---|---|
| Author | tsp, pinterf |
| Version | v0.8.4 |
| Download | FFT3dGPU-v0.8.4-pfmod.7z |
| Category | Denoisers |
| License | GPLv2 |
| Discussion | Doom9 Forum (original) |
Contents |
Description
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+.
Requirements
- [x86]: AviSynth+ or AviSynth 2.6.0
- [x64]: AviSynth+
- Supported color formats: Y8, YV12, YV16, YV24, YV411; YUV(A)
Syntax and Parameters
FFT3DGPU (clip, float "sigma", float "beta", int "bw", int "bh", int "bt", float "sharpen", int "plane", int "mode", int "bordersize", int "precision", bool "NVPerf", float "degrid", float "scutoff", float "svr", float "smin", float "smax", float "kratio", int "ow", int "oh", int "wintype" , int "interlaced", float "sigma2", float "sigma3", float "sigma4", bool "oldfft")
- clip =
- Input clip
- clip =
- float sigma = 2.0
- Given noise value for all (or highest) frequencies. Only values greater that 0.0
- float sigma = 2.0
- float beta = 1.0
- Noise margin (float>=1.0, default=1.0 for no noise left): control noise amount left in the restoration for Wiener filter, so max filter noise attenuation =
(beta-1)/beta.
- Noise margin (float>=1.0, default=1.0 for no noise left): control noise amount left in the restoration for Wiener filter, so max filter noise attenuation =
- float beta = 1.0
- int bw = 32
- int bh = 32
- Block width and height. It should be a power of 2 ie valid values is 4,8,16,32,64,128,256,512 (note that
bwshould be greater than 4 for best result).
- Block width and height. It should be a power of 2 ie valid values is 4,8,16,32,64,128,256,512 (note that
- int bw = 32
- int bt = 3
- Block temporal size, number of frames.
- -1 : sharpen only (2D)
- 0 : all previous frames (switch Kalman filter mode)
- 1 : only current frame (spatial 2D Wiener filter)
- 2 : previous and current frame (3D Wiener filter)
- 3 : previous, current and next frame (3D Wiener filter)
- 4 : two previous, current and next frame (3D Wiener filter)
- Block temporal size, number of frames.
- int bt = 3
- float sharpen = 0.0
- Sharpening strength, positive values sharpens the image, negative values blurs the image. 0 disables sharpening.
- float sharpen = 0.0
- int plane = 0
- Processed color plane:
- 0 : luma(Y)
- 1 : chroma U
- 2 : chroma V
- 3 : chroma planes U and V
- 4 : both luma and chroma
- For a greyscale clip. 4 means luma (the only plane), 1,2 and 3 gives error message.
- For RGB clips (planar RGB only!) operation is always 4=all planes, regardless of this parameter.
- Processed color plane:
- int plane = 0
- int mode = 1
- Processing mode:
- 0 : only overlaps 1:1. This is faster but produces artifacts with high sigma values.
- 1 : block overlaps 2:1. This is slower but produces fewer artifacts.
- 2 : again 1:1 overlap but with a additional border. This reduces border artifacts seen with mode=0.
- The speed is between mode 0 and 1.
Kalman(bt=0)works well withmode=0.
- Processing mode:
- int mode = 1
- int bordersize = 1
- Only used with <mode>mode</mode> 2. Defines the size of the border.
- int bordersize = 1
- int precision = 0
- 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.
- 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.
- int precision = 0
- bool NVPerf = false
- Enables support for NVPerfHUD (http://developer.nvidia.com/object/nvperfhud_home.html).
- bool NVPerf = false
- float degrid = 1.0
- Enables degriding. Only works well with
mode=1. Doesn't degrid the Kalman filter (but it does degrid the sharpening (if enabled) after kalman filter). default 1.0 formode=1, 0.0 formode=0or 2.
- Enables degriding. Only works well with
- float degrid = 1.0
- float scutoff = 0.3
- Sharpening cutoff frequency, relative to max.
- float scutoff = 0.3
- float svr = 1.0
- Sharpening (and dehalo) vertical ratio (to horizontal) (default=1.0 - same as horizontal, 0 - no vertical sharpening)
- float svr = 1.0
- float smin = 4.0
- float smin = 20.0
- Minimum and maximum limit (approximate noise margin) for sharpening stage.
- float smin = 4.0
- float kratio = 2.0
- Ratio of threshold to sigma to reset Kalman filter. Variation threshold =
sigma*kratio, good value is about from 1.5 to 3.0
- Ratio of threshold to sigma to reset Kalman filter. Variation threshold =
- float kratio = 2.0
- int ow = bw/2
- int oh = bh/2
- Overlap width and height. This only works with
mode=1. This specifies how big the overlap between the blocks are. Overlap size must be less than or equal to half the blocksize.owmust be even.
- Overlap width and height. This only works with
- int ow = bw/2
- int wintype = 0
- weighting windows type:
- 0 : same analysis and synthesis half-cosine window, used in all versions before 1.4.
- 1 : intermediate between 0 and 2
- 2 : flat analysis window, rised cosine (Hanning) synthesis window.
- weighting windows type:
- int wintype = 0
- bool interlaced = false
- Set to true for separate filtering for each field.
- bool interlaced = false
- float singma2 = sigma
- float singma3 = sigma
- float singma4 = sigma
- Given noise value at second, third, and lowest scale level frequencies. Only values greater that 0.0
- float singma2 = sigma
- bool oldfft = 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.
- bool oldfft = false
- bool pframe = false
- Noise pattern frame number.
- bool pframe = false
- int px = 0
- int py = 0
- Noise pattern block horizontal X position and vertical Y position. If
px=py=0, then the pattern block is defined automatically with minimal power spectral density.
- Noise pattern block horizontal X position and vertical Y position. If
- int px = 0
- bool pshow = false
- Show noise pattern block and its properties.
- bool pshow = false
- float pcutoff = 0.1
- Noise pattern cutoff frequency (relative to max).
- float pcutoff = 0.1
- float pfactor = 0.0
- Noise pattern denoise strength (0 to 1.0, default=0, this method disabled).
- float pfactor = 0.0
- float dehalo = 0.0
- Halo removal strength. Only values greater that 0.0
- float dehalo = 0.0
- float hr = 2.0
- Halo approximate radius. Only values greater that 0.0
- float hr = 2.0
- float ht = 50.0
- Halo approximate threshold. Only values greater that 0.0
- float ht = 50.0
- int ncpu = 1
- Max number of CPU threads to use for FFT calculation.
- int ncpu = 1
Changelog
Version Date Changes
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.
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
v0.8.2a 2006/08/24 - Last release by tsp, see included docs for full changelog
Back to External Filters ←
