Yet another inverse telecine filter by Zoltan Hidvegi <mplayer@hzoli.2y.net>. Also heavily MMX centric.

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11601 b3059339-0415-0410-9bf9-f77b7e298cf2

DOCS/man/en/mplayer.1

@@ -2917,6 +2917,80 @@ during 30 fps progressive sequences or stills, it will output more
 than 24 fps, meaning you'll have trouble if you use it with mencoder
 \-ofps 23.976.
 .TP
+.B filmdint[=options]
+Inverse telecine filter, similar to the pullup filter above.
+It is designed to handle any pulldown pattern, including mixed soft and
+hard telecine and limited support for movies that are slowed down or sped
+up from their original framerate for TV.
+Only the luma plane is used to find the frame breaks.
+If a field has no match, it is deinterlaced with simple linear
+approximation.
+If the source is MPEG-2, libmpeg2 must be used for decoding (not
+ffmpeg2!), and this must be the first filter to allow access to the
+field-flags set by the MPEG-2 decoder.
+Depending on the source mpeg, you may be fine ignoring this advice, as
+long as you do not see lots of "Bottom-first field" warnings.
+With no options it does normal inverse telecine, and should be used
+together with mencoder \-fps 29.97 \-ofps 23.976.
+When this filter is used with mplayer, it will result in an uneven
+framerate during playback, but it is still generally better than using
+pp=lb or no deinterlacing at all.
+Multiple options can be specified separated by /.
+.RSs
+.IPs crop=w:h:x:y
+Just like the crop filter, but faster, and works on mixed hard and soft
+telecined content as well as when y is not a multiple of 4.
+If x or y would require cropping fractional pixels from the chroma
+planes, the crop area is extended.
+This usually means that x and y must be even.
+.IPs io=ifps:ofps
+For each ifps input frames the filter will output ofps frames.
+The ratio of ifps/ofps should match the \-fps/\-ofps ratio.
+This could be used to filter movies that are broadcast on TV at a frame
+rate different from their original frame rate.
+.IPs luma_only=n
+If n is nonzero, the chroma plane is copied unchanged.
+This is useful for YV12 sampled TV, which discards one of the chroma
+fields.
+.IPs mmx2=n
+On x86, if n=1, use MMX2 optimized functions, if n=2, use 3DNow!
+optimized functions, othewise, use plain C.
+If this option is not specified, MMX2 and 3DNow! are auto-detected, use
+this option to override auto-detection.
+.IPs fast=n
+The larger n will speed up the filter at the expense of accuracy.
+The default value is n=3.
+If n is odd, a frame immediately following a frame marked with the
+REPEAT_FIRST_FIELD mpeg flag is assumed to be progressive, thus filter
+will not spend any time on soft-telecined MPEG-2 content.
+This is the only effect of this flag if MMX2 or 3DNow! is available.
+Without MMX2 and 3DNow, if n=0 or 1, the same calculations will be used
+as with MMX2.
+If n=2 or 3, the number of luma levels used to find the frame breaks is
+reduced from 256 to 128, which results in a faster filter without losing
+much accuracy.
+If n=4 or 5, a faster, but much less accurate metrics will be used to
+find the frame breaks, which is more likely to misdetect high vertical
+detail as interlaced content.
+.IPs verbose=n
+If n is nonzero, print the detailed metrics for each frame.
+Useful for debugging.
+.IPs dint_thres=n
+Deinterlace threshold.
+Used during de-interlacing of unmatched frames.
+Larger value means less deinterlacing, use n=256 to completely turn off
+deinterlacing.
+Default is n=8.
+.IPs comb_thres=n
+Threshold for comparing a top and bottom fields.
+Defaults to 128.
+.IPs diff_thres=n
+Threshold to detect temporal change of a field.
+Default is 128.
+.IPs sad_thres=n
+Sum of Absolute Difference threshold, default is 64.
+.RE
+.TP
 .B softpulldown
 This filter works only correct with MEncoder and acts on the MPEG2 flags
 used for soft 3:2 pulldown (soft telecine).

libmpcodecs/Makefile

@@ -14,7 +14,7 @@ VIDEO_SRCS_NAT=vd_null.c vd_cinepak.c vd_qtrpza.c vd_raw.c vd_hmblck.c vd_msvidc
 VIDEO_SRCS_OPT=vd_realvid.c vd_ffmpeg.c vd_dshow.c vd_dmo.c vd_vfw.c vd_vfwex.c vd_odivx.c vd_divx4.c vd_zrmjpeg.c vd_xanim.c vd_xvid.c vd_xvid4.c vd_libdv.c vd_qtvideo.c vd_theora.c
 VIDEO_SRCS=dec_video.c vd.c $(VIDEO_SRCS_NAT) $(VIDEO_SRCS_LIB) $(VIDEO_SRCS_OPT)
 
-VFILTER_SRCS=vf.c vf_vo.c vf_crop.c vf_expand.c vf_scale.c vf_format.c vf_yuy2.c vf_flip.c vf_rgb2bgr.c vf_rotate.c vf_mirror.c vf_palette.c vf_lavc.c vf_dvbscale.c vf_cropdetect.c vf_test.c vf_noise.c vf_yvu9.c vf_rectangle.c vf_lavcdeint.c vf_eq.c vf_eq2.c vf_halfpack.c vf_dint.c vf_1bpp.c vf_bmovl.c vf_2xsai.c vf_unsharp.c vf_swapuv.c vf_il.c vf_boxblur.c vf_sab.c vf_smartblur.c vf_perspective.c vf_down3dright.c vf_field.c vf_denoise3d.c vf_hqdn3d.c vf_detc.c vf_telecine.c vf_tfields.c vf_ivtc.c vf_ilpack.c vf_dsize.c vf_decimate.c vf_softpulldown.c vf_tinterlace.c vf_pullup.c pullup.c vf_framestep.c vf_tile.c vf_delogo.c vf_fil.c vf_hue.c vf_spp.c vf_yuvcsp.c
+VFILTER_SRCS=vf.c vf_vo.c vf_crop.c vf_expand.c vf_scale.c vf_format.c vf_yuy2.c vf_flip.c vf_rgb2bgr.c vf_rotate.c vf_mirror.c vf_palette.c vf_lavc.c vf_dvbscale.c vf_cropdetect.c vf_test.c vf_noise.c vf_yvu9.c vf_rectangle.c vf_lavcdeint.c vf_eq.c vf_eq2.c vf_halfpack.c vf_dint.c vf_1bpp.c vf_bmovl.c vf_2xsai.c vf_unsharp.c vf_swapuv.c vf_il.c vf_boxblur.c vf_sab.c vf_smartblur.c vf_perspective.c vf_down3dright.c vf_field.c vf_denoise3d.c vf_hqdn3d.c vf_detc.c vf_telecine.c vf_tfields.c vf_ivtc.c vf_ilpack.c vf_dsize.c vf_decimate.c vf_softpulldown.c vf_tinterlace.c vf_pullup.c pullup.c vf_framestep.c vf_tile.c vf_delogo.c vf_fil.c vf_hue.c vf_spp.c vf_yuvcsp.c vf_filmdint.c
 ifeq ($(HAVE_FFPOSTPROCESS),yes)
 VFILTER_SRCS += vf_pp.c
 endif

libmpcodecs/cmmx.h

@@ -0,0 +1,186 @@
+/*
+ * x86 MMX and MMX2 packed byte operations in portable C.
+ * Extra instructions: pdiffub, pcmpzb, psumbw, pcmpgtub
+ * Author: Zoltan Hidvegi
+ */
+
+#ifndef __CMMX_H
+#define __CMMX_H
+
+typedef unsigned long cmmx_t;
+
+#define ONE_BYTES (~(cmmx_t)0 / 255)
+#define SIGN_BITS (ONE_BYTES << 7)
+#define LOWBW_MASK (~(cmmx_t)0 / 257)
+
+static inline cmmx_t
+paddb(cmmx_t a, cmmx_t b)
+{
+    return ((a & ~SIGN_BITS) + (b & ~SIGN_BITS)) ^ ((a^b) & SIGN_BITS);
+}
+
+static inline cmmx_t
+psubb(cmmx_t a, cmmx_t b)
+{
+    return ((a | SIGN_BITS) - (b & ~SIGN_BITS)) ^ (~(a^b) & SIGN_BITS);
+}
+
+static inline cmmx_t
+paddusb(cmmx_t a, cmmx_t b)
+{
+    cmmx_t s = (a & ~SIGN_BITS) + (b & ~SIGN_BITS);
+    cmmx_t abs = (a | b) & SIGN_BITS;
+    cmmx_t c = abs & (s | (a & b));
+    return s | abs | (abs - (c >> 7));
+}
+
+static inline cmmx_t
+paddusb_s(cmmx_t a, cmmx_t b)
+{
+    cmmx_t sum = a+b;
+    cmmx_t ov = sum & SIGN_BITS;
+    return sum + (sum ^ (ov - (ov>>7)));
+}
+
+static inline cmmx_t
+psubusb(cmmx_t a, cmmx_t b)
+{
+    cmmx_t s = (a | SIGN_BITS) - (b & ~SIGN_BITS);
+    cmmx_t anb = a & ~b;
+    cmmx_t c = (anb | (s & ~(a^b))) & SIGN_BITS;
+    return s & ((c & anb) | (c - (c >> 7)));
+}
+
+static inline cmmx_t
+psubusb_s(cmmx_t a, cmmx_t b)
+{
+    cmmx_t d = (a|SIGN_BITS) - b;
+    cmmx_t m = d & SIGN_BITS;
+    return d & (m - (m>>7));
+}
+
+static inline cmmx_t
+pcmpgtub(cmmx_t b, cmmx_t a)
+{
+    cmmx_t s = (a | SIGN_BITS) - (b & ~SIGN_BITS);
+    cmmx_t ret = ((~a & b) | (~s & ~(a ^ b))) & SIGN_BITS;
+    return ret | (ret - (ret >> 7));
+}
+
+static inline cmmx_t
+pdiffub(cmmx_t a, cmmx_t b)
+{
+    cmmx_t xs = (~a ^ b) & SIGN_BITS;
+    cmmx_t s = ((a | SIGN_BITS) - (b & ~SIGN_BITS)) ^ xs;
+    cmmx_t gt = ((~a & b) | (s & xs)) & SIGN_BITS;
+    cmmx_t gt7 = gt >> 7;
+    return (s ^ gt ^ (gt - gt7)) + gt7;
+}
+
+static inline cmmx_t
+pdiffub_s(cmmx_t a, cmmx_t b)
+{
+    cmmx_t d = (a|SIGN_BITS) - b;
+    cmmx_t g = (~d & SIGN_BITS) >> 7;
+    return (d ^ (SIGN_BITS-g)) + g;
+}
+
+static inline cmmx_t
+pmaxub(cmmx_t a, cmmx_t b)
+{
+    return psubusb(a,b) + b;
+}
+
+static inline cmmx_t
+pminub(cmmx_t a, cmmx_t b)
+{
+    return paddusb(a,~b) - ~b;
+}
+
+static inline cmmx_t
+pminub_s(cmmx_t a, cmmx_t b)
+{
+    cmmx_t d = (a|SIGN_BITS) - b;
+    cmmx_t m = ~SIGN_BITS + ((d&SIGN_BITS)>>7);
+    return ((d&m) + b) & ~SIGN_BITS;
+}
+
+static inline cmmx_t
+pavgb(cmmx_t a, cmmx_t b)
+{
+    cmmx_t ao = a & ONE_BYTES;
+    cmmx_t bo = b & ONE_BYTES;
+    return ((a^ao)>>1) + ((b^bo)>>1) + (ao|bo);
+}
+
+static inline cmmx_t
+pavgb_s(cmmx_t a, cmmx_t b)
+{
+    return ((a+b+ONE_BYTES)>>1) & ~SIGN_BITS;
+}
+
+static inline cmmx_t
+p31avgb(cmmx_t a, cmmx_t b)
+{
+    cmmx_t ao = a & (3*ONE_BYTES);
+    cmmx_t bo = b & (3*ONE_BYTES);
+    return 3*((a^ao)>>2) + ((b^bo)>>2) +
+	(((3*ao+bo+2*ONE_BYTES)>>2) & (3*ONE_BYTES));
+}
+
+static inline cmmx_t
+p31avgb_s(cmmx_t a, cmmx_t b)
+{
+    cmmx_t avg = ((a+b)>>1) & ~SIGN_BITS;
+    return pavgb_s(avg, a);
+}
+
+static inline unsigned long
+psumbw(cmmx_t a)
+{
+    cmmx_t t = (a & LOWBW_MASK) + ((a>>8) & LOWBW_MASK);
+    unsigned long ret =
+	(unsigned long)t + (unsigned long)(t >> (4*sizeof(cmmx_t)));
+    if (sizeof(cmmx_t) > 4)
+	ret += ret >> 16;
+    return ret & 0xffff;
+}
+
+static inline unsigned long
+psumbw_s(cmmx_t a)
+{
+    unsigned long ret =
+	(unsigned long)a + (unsigned long)(a >> (4*sizeof(cmmx_t)));
+    if (sizeof(cmmx_t) <= 4)
+	return (ret & 0xff) + ((ret>>8) & 0xff);
+    ret = (ret & 0xff00ff) + ((ret>>8) & 0xff00ff);
+    ret += ret >> 16;
+    return ret & 0xffff;
+}
+
+static inline unsigned long
+psadbw(cmmx_t a, cmmx_t b)
+{
+    return psumbw(pdiffub(a,b));
+}
+
+static inline unsigned long
+psadbw_s(cmmx_t a, cmmx_t b)
+{
+    return psumbw_s(pdiffub_s(a,b));
+}
+
+static inline cmmx_t
+pcmpzb(cmmx_t a)
+{
+    cmmx_t ret = (((a | SIGN_BITS) - ONE_BYTES) | a) & SIGN_BITS;
+    return ~(ret | (ret - (ret >> 7)));
+}
+
+static inline cmmx_t
+pcmpeqb(cmmx_t a, cmmx_t b)
+{
+    return pcmpzb(a ^ b);
+}
+
+#endif

libmpcodecs/vf.c

@@ -75,6 +75,7 @@ extern vf_info_t vf_info_dsize;
 extern vf_info_t vf_info_decimate;
 extern vf_info_t vf_info_softpulldown;
 extern vf_info_t vf_info_pullup;
+extern vf_info_t vf_info_filmdint;
 extern vf_info_t vf_info_framestep;
 extern vf_info_t vf_info_tile;
 extern vf_info_t vf_info_delogo;
@@ -143,6 +144,7 @@ static vf_info_t* filter_list[]={
     &vf_info_decimate,
     &vf_info_softpulldown,
     &vf_info_pullup,
+    &vf_info_filmdint,
     &vf_info_framestep,
     &vf_info_tile,
     &vf_info_delogo,