From Avisynth wiki
Jump to: navigation, search
Author Chikuzen
Version v0.1.4
Download TMM2-
Category Masking
License GPLv2



TMM2 is a rewrite of TMM for AviSynth2.6/AviSynth+ which written by Kevin Stone (aka tritical).
TMM2 builds a motion-mask for TDeint, which TDeint uses via its 'emask' parameter. TMM2 can use fixed or per-pixel adaptive motion thresholds, as well as any length static period greater than or equal to six fields. It checks backwards, across, and forwards when looking for motion.


*** vcredist_x86.exe is required for TMM2-x86
*** vcredist_x64.exe is required for TMM2-x64

Syntax and Parameters

TMM2 (clip, int "mode", int "order", int "field", int "length", int "mtype", int "ttype", int "mtqL", int "mthL", int "mtqC", int "mthC", int "nt", int "minthresh", int "maxthresh", int "cstr", int "opt")

clip   =
Input clip.

int  mode = 0
Same as the 'mode' parameter of TDeint. This should be set equal to the value of 'mode' given to TDeint. If using mode=2 in TDeint use mode=1 in TMM2.
  • 0 - same rate output
  • 1 - double rate output

int  order = -1
Same as the 'order' parameter of TDeint. This should be set equal to the value of 'order' given to TDeint.
  • -1 - use order from AviSynth
  • 0 - bff
  • 1 - tff

int  field = -1
Same as the 'field' parameter of TDeint. This should be set equal to the value of 'field' given to TDeint. If using mode=1 (double rate output) then this setting does nothing.
  • -1 - set field equal to order
  • 0 - keep bottom field
  • 1 - keep top field

int  length = 10
Sets the number of fields required for declaring pixels as stationary. length=6 means six fields (3 top/3 bottom), length=8 means 8 fields (4 top/4 bottom), etc...
This can be any value greater than or equal to 6 (can be even or odd). A larger value for length will prevent more motion-adaptive related artifacts, but will result in fewer pixels being weaved.

int  mtype = 1
Sets whether or not both vertical neighboring lines in the current field of the line in the opposite parity field attempting to be weaved have to agree on both stationarity and direction. Possible values:
  • 0 - no
  • 1 - no for across, but yes for backwards/forwards
  • 2 - yes
0 will result in the most pixels being weaved, while 2 will have the least artifacts.

int  ttype = 1
Sets how to determine the per-pixel adaptive (quarter pel/half pel) motion thresholds. Possible values:
  • 0 - 4 neighbors, diff to center pixel, compensated
  • 1 - 8 neighbors, diff to center pixel, compensated
  • 2 - 4 neighbors, diff to center pixel, uncompensated
  • 3 - 8 neighbors, diff to center pixel, uncompensated
  • 4 - 4 neighbors, range (max-min) of neighborhood
  • 5 - 8 neighbors, range (max-min) of neighborhood
Compensated means adjusted for distance differences due to field vs frames and chroma downsampling. The compensated versions will always result in thresholds <= to the uncompensated versions.

int  mtqL = -1
int  mthL = -1
int  mtqC = -1
int  mthC = -1
These parameters allow the specification of hard thresholds instead of using per-pixel adaptive motion thresholds. mtqL sets the quarter pel threshold for luma, mthL sets the half pel threshold for luma, mtqC/mthC are the same but for chroma. If these parameters are set to -1 then an adaptive threshold is used. Otherwise, if they are between 0 and 255 (inclusive) then the value of the parameter is used as the threshold for every pixel.

int  nt = 2
int  minthresh = 4
int  maxthresh = 75
nt sets the noise threshold, which will be added to the value of each per-pixel threshold when determining if a pixel is stationary or not. After the addition of 'nt', any threshold less than minthresh will be increased to minthresh and any threshold greater than maxthresh will be decreased to maxthresh.

int  cstr = 4
Sets the number of required neighbor pixels (3x3 neighborhood) in the quarter pel mask, of a pixel marked as moving in the quarter pel mask, but stationary in the half pel mask, marked as stationary for the pixel to be marked as stationary in the combined mask.

int  opt = -1
Controls which cpu optimizations are used for create motion masks.
  • 0 - Use C++ routine.
  • 1 - Use SSE2 routine if possible. When SSE2 can't be used, fallback to 0.
  • others - Use AVX2 routine if possible. When AVX2 can't be used, fallback to 1.


MPEG2Source("source.d2v")  // assume tff source that we want to bob
TDeint(order=1, mode=1, emask=TMM2(order=1, mode=1))
MPEG2Source("source.d2v")  // assume bff source, keep top field, same rate
TDeint(order=0, field=1 ,emask=TMM2(order=0, field=1)


  • TMM2_avx2.dll is compiled with /arch:AVX2. (only applicable to v0.1.1)
  • On AviSynth 2.6, AVX2 is always disabled even if you use TMM2_avx2.dll. (only applicable to v0.1.1)
  • On AviSynth+, TMM2 is set as MT_NICE_FILTER automatically. You don't have to set SetFilterMTMode() yourself for this filter.
  • This filter requires appropriate memory alignments. Thus, if you want to crop the left side of your source clip before this filter, you have to set crop(align=true) (default behaviour in AviSynth+).


Version      Date            Changes
v0.1.4 2020/05/14 - Changes by asd-g - Copy frame properties when available v0.1.3 2020/05/13 - Changes by asd-g - Update to AviSynth+'s v8 interface v0.1.1 2016/07/05 - Update avisynth.h to Avisynth+MT r2005 - CreateMM: allocate temporal buffer at constructor unless the enviroment is not avs+MT. v0.1 2016/05/25 - Use buffer pool on avs+ MT. - Disable AVX2 always on Avisynth2.6. v0.0 2016/05/20 - Initial release

External Links

  • GitHub - Source code repository.
  • GitHub - Source code repository (update).

Back to External Filters

Personal tools