#YV12To422

YV12 to YV16/YUY2 converter for AviSynth2.6

YV12To422 is an avisynth filter plugin which based on YV12ToYUY2(ddcc.dll) written by Kevin Stone(a.k.a tritical) and was written from scratch.

###Info: Convert YV12 clip to YV16/YUY2 using SSE2/AVX2.

version 1.0.2

Requirement:

- avisynth2.60/avisynth+r1576 or later.
- SSE2 capable CPU.
- WindowsVista sp2 or later.
- Visual C++ Redistributable Packages for Visual Studio 2013.

Syntax:

YV12To422(clip, bool "interlaced", int "itype", int "cplace", bool "lshift",
          bool "yuy2", bool "avx2", bool "threads", float "b", float "c")


NOTE: these parameters may be changed later.
      (Sorry, I'm not enthusiastic about keeping backward compatibility.)

interlaced -

  Sets whether or not the input video is interlaced or progressive.

  default:  false

itype -

  Sets interpolation method. Possible settings:

      0 - duplicate (nearest neighbor)
      1 - linear interpolation
      2 - Mitchell-Netravali two-part cubic interpolation
             (adjustable b/c parameters to adjust blurring/ringing)

  default:  2

cplace -

  Specifies vertical chroma placement.  Possible settings:

    progressive input (interlaced=false, progressive upsampling):

        0 - chroma is aligned with top line of each two line pair within the frame

            This would be the case if during 4:2:2 -> 4:2:0 conversion the chroma values of
            odd lines were simply dropped.

        1 - chroma is centered between lines of each two line pair within the frame
            (*** h261, h263, mpeg1, mpeg2, mpeg4, h264 standard progressive conversion)

            This would be the case if during 4:2:2 -> 4:2:0 conversion the chroma from every
            two line pair was averaged, or if an interlaced 4:2:2 -> 4:2:0 conversion was performed
            by using 75/25 averaging of top field pairs and 25/75 averaging of bottom field pairs.

        2 - chroma is aligned with top line of each two line pair within each field

            This would be the case if the 4:2:2 -> 4:2:0 conversion was performed by
            separating the fields, and then doing a 4:2:2 -> 4:2:0 conversion on each field
            by dropping odd line chroma values.

        3 - chroma is centered between lines of each two line pair within each field

            This would be the case if the 4:2:2 -> 4:2:0 conversion was performed by
            separating the fields, and then doing a 4:2:2 -> 4:2:0 conversion on each field
            by averaging chroma from every two line pair.

    interlaced input (interlaced=true, interlaced upsampling):

        0 - chroma is aligned with top line of each two line pair within each field

            This would be the case if the 4:2:2 -> 4:2:0 conversion was performed by
            separating the fields, and then doing a 4:2:2 -> 4:2:0 conversion on each field
            by dropping odd line chroma values.

        1 - chroma is centered between lines of each two line pair within each field

            This would be the case if the 4:2:2 -> 4:2:0 conversion was performed by
            separating the fields, and then doing a 4:2:2 -> 4:2:0 conversion on each field
            by averaging chroma from every two line pair.

        2 - top field chroma is 1/4 pixel below even lines in the top field, and
            bottom field chroma is 1/4 pixel above odd lines in the bottom field.
            (*** mpeg2, mpeg4, h264 standard interlaced conversion)

            This would be the case if the 4:2:2 -> 4:2:0 conversion was performed by
            averaging top field pairs using 75/25 weighting, and averaging bottom field
            pairs using 25/75 weighting.  This results in the same chroma placement as
            progressive cplace option 1.

        3 - U is aligned with top line of each two line pair within each field, and
            V is aligned with bottom line of each pair within each field.
            (*** DV-PAL standerd interlaced conversion)

    ** Progressive option 1 and interlaced option 2 are actually the same in terms of
       chroma placement.  If a progressive frame was converted to yv12 using interlaced
       method 2, then it is safe (actually better) to convert it to yuy2 using progressive
       method 1.  However, if a progressive frame was converted to yv12 using a different
       type of interlaced sampling, resulting in different chroma placement (such as those
       described by interlaced options 0 or 1), then it is best to convert it to yv16/yuy2
       using progressive upsampling, but with the cplace option that correctly specifies
       the positioning of the chroma.

  default:  1 (if interlaced = false)
            2 (if interlaced = true)

lshift -

  If set this to true, chroma placement will shift to 1/4 sample to the left.

    default: false

yuy2 -

  Sets whether or not the output video format is packed(YUY2) or planar(YV16).

  defaullt: true (YUY2 output)

avx2 -

  Sets whether AVX2 is used or not.

  default: false (use SSE2)

    ** Currentry, avisynth2.60 can't make memory alignment anything but 16bytes.
       Thus, if you use avisynth2.60, you shouldn't to set this true.
       avisynth+ has no problem.
       see https://github.com/AviSynth/AviSynthPlus/commit/ab4ea303b4ca78620c2ef90fdaad184bc18b7708

threads -

  When sets this to true, V-plain is processed with a different thread at the
  same time with U-plain.
  However, processing doesn't always become speedy by this.

  default: false (use single thread)

b / c -

  Adjusts properties of cubic interpolation (itype=2).  Same as Avisynth's BicubicResize filter.

  default:  0.0,0.75

Lisence:

GPLv2 or later

Source code:

https://github.com/chikuzen/YV12To422/