TEMmod

Revision as of 15:56, 4 August 2014 (view source) Reel.Deal (Talk | contribs) m (→‎Syntax and Parameters) ← Older edit		Revision as of 16:17, 4 August 2014 (view source) Reel.Deal (Talk | contribs) m (→‎Syntax and Parameters) Newer edit →
Line 57:		Line 57:
	::::*0 : do not process		::::*0 : do not process
	::::*1 : process		::::*1 : process
−	::::*3 : fill with zeroes	+	::::*2 : fill with zeroes
	<br>		<br>
	::{{Par2|invert|bool|false}}		::{{Par2|invert|bool|false}}

---

Revision as of 16:17, 4 August 2014

Contents 1 Description 2 Requirements 3 Syntax and Parameters 4 Examples 5 Changelog 6 Archived Downloads 7 External Links

Description

TEMmod creates an edge mask using gradient vector magnitude. It's based on TEdgeMask written by Kevin Stone (a.k.a. tritical) but rewritten from scratch exclusively for AviSynth 2.6.

Requirements

- AviSynth 2.6.0 Alpha5 or greater

- Supported color formats: Y8, YV12, YV16, YV24, YV411

- SSE2 capable CPU

- Microsoft Visual C++ 2010 SP1 Redistributable Package (x86)

Syntax and Parameters

TEMmod (clip, float "threshY", float "threshC", int "type", int "link", int "chroma", bool "invert", bool "preblur", float "scale")

clip   =

Input clip; only planar formats are supported.

float  threshY = 8.0

Set the magnitude thresholds for Y-plane.

If over this value then a sample will be considered an edge, and the output mask will be binarized to by 0 and 255.

Setting this to 0 means output a magnitude mask instead of binary mask.

float  threshC = 8.0

Same as threshY but sets the magnitude thresholds for U-plane and V-plane.

int  type = 4

Sets the type of first order partial derivative approximation that is used.

Possible values:

• 1 : 2 pixel (floating point arithmetic, almost same as original)
• 2 : 4 pixel (floating point arithmetic, almost same as original)
• 3 : 2 pixel (type1 with SSE2 integer arithmetic, a bit incorrect but faster)
• 4 : 4 pixel (type2 with SSE2 integer arithmetic, a bit incorrect but faster)
• 5 : 6 pixel (Sobel operator with SSE2 integer arithmetic)

int  link = 1

Specifies whether luma to chroma linking, no linking, or linking of every plane to every other plane is used.

• 0 : no linking
• 1 : luma to chroma linking
• 2 : every plane to every other plane

int  chroma = 1

This control how chroma (U-plane and V-plane) is processed:

• 0 : do not process
• 1 : process
• 2 : fill with zeroes

bool  invert = false

If this is set to True, the output mask will be inverted.

bool  preblur = false

Indicates whether to apply a 3x3 guassian blur to the input image prior to generating the edge map.

float  scale = 0.0

If output is a magnitude mask(threshY=0), it is scaled by this value.

Setting this to 0 means output is adjusted automatically and scaled onto 0-255 range.

Examples

TEMmod with default settings:

AviSource("blah.avi")
TEMmod (threshY=8.0, threshC=8.0, type=4, link=1, chroma=1, invert=false, preblur=false, scale=0.0)

Changelog

Version      Date            Changes

v0.2.0       08/12/2013      - Add new calculatre map method (generic Sobel edge detection) as type5.
                             - Add new argument 'scale'. 

v0.1.1       08/09/2013      - Fix a bug

v0.1.0       08/09/2013      - Add two types of calculate map method (almost same as original)
                             - More SSE2 optimization
    
v0.0.1       08/08/2013      - Initial release

Archived Downloads

Version	Download	Source code	Mirror
v0.2.0	TEMmod-0.2.0.zip	TEMmod-master.zip

External Links

• GitHub - Source code repository.
• Doom9 Forum - TEMmod discussion.

---

Back to External Filters ←