Problematic interlaced source (missing/dupped fields)

[Dogway]

I spent the last 10 days trying to fix a short clip with no success. It is supposedly interlaced but it shows progressive frames which look just like upsized fields. There are also dupped fields, and missing fields (jumping gaps) on no pattern order.
I made a thread here explaining the problem in detail and exposing some takes and comes out.

17Mb
http://www.mediafire.com/?t26hmac5kyawyqf

My intention is to make this a temporal even spaced 59.94fps.
For that I first drop half the frames (since they are dups), then I nnedi3() the fields to make them suitable for motion interpolation. Interpolate (to create the missing fields gaps) and convert back to fields with separatefields(). Now I need to take all the needed fields (original and interpolated missing fields) in a certain order Top+Bottom or Bottom+Top, and here comes the problem, when I select a field belonging to a missing field (interpolated) the next field I must choose (top or bottom) isn't the original field anymore but the nnedi3 interpolated part of it.

These are the related posts about this issue. Here and here.

This the code I have just before selecting the correct fields.

setmtmode(3,2)
MPEG2Source("source.d2v", cpu=0)
setmtmode(2,2)

Trim(0,926)

separatefields()

#chroma fix
interleave(selectevery(4,0),selectevery(4,1),mergechroma(selectevery(4,2),selectevery(4,-0)),selectevery(4,3))

# drop dups
b=0
selectevery(4,b+2,b+3)

t=trim(774,926).selecteven.TemporalSoften(10,255,255,25,2)
trim(0,773)+interleave(t,t) # Fixing all the crap out of the fields

# back to initial state
weave()

# start from problematic section
trim(252,0)

a=last

# preparing for interpolation
nnedi3(field=-2)
assumefps(10)

# interpolation
chroma=true
superV = MSuper(pel=2)
vbo=MAnalyse(superV,isb=true, overlap=4, delta=1,chroma=chroma,search=3)
vbo=MRecalculate(superV,vbo, overlap=2,blksize=4,chroma=chroma)
vfo=MAnalyse(superV,isb=false, overlap=4,delta=1,chroma=chroma,search=3)
vfo=MRecalculate(superV,vfo, overlap=2,blksize=4,chroma=chroma)
MFlowFps(MSuper(pel=2,levels=1), vbo, vfo, num=20,den=1)

And this the behind thinking/scheme to picture the overall issue.

i=interlaced frame
f=field frame

1) Numbering (useful for selectevery())
2) Interlacing of original source (after trim(252,0))
3) Fields      of original source (after trim(252,0))
4) Fields after Motion Interpolation. (in the form of frames from nnedi3 output)
5) Fields after Motion Interpolation. (same as 4) after separatefields()) Capital letter is top field.
6) The nnedi3 (spatial interpolated) part of original (and motion interpolated) fields (aka garbage) for the 5) line.

1) 0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31   32   33   34   35   36   37   38   39   40   41   42   43   44   45   46   47   48   49   50   51   52   53   54   55   56   57   58   59   60   61   62   63   64   65   66   67   68   69   70   71   72   73   74   75   76   77   78   79   80   81   82   83   84   85   86   87   88   89   90   91  |92   93   94   95   96   97   98   99   100  101  102  103  104  105  106  107  108  109  110  111  112  113  114  115  116  117  118  119  120  121  122  123  124  125  126  127  128  129  130  131  132  133  134  135  136  137  138  139  140  141  142  143  144  145  146  147  148  149  150  151  152  153  154  155  156  157  158  159  160  161  162  163  164  165  166  167  168  169  170  171  172  173  174  175  176  177  178  179  180  181  182  183  184  185  186  187  188  189  190  191  192  193  194  195  196  197  198  199  200  201  202  203  204  205  206  207  208  209  210  211  212  213  214  215  216  217  218  219  220  221  222  223  224  225  226  227  228  229  230  231  232  233  234  235  236  237  238  239  240  241  242  243  244  245  246  247  248  249  250  251  253  253  254  255  256  257  258  259  260  261  262  263  264  265  266  267  268  269  270  271  272  273  274  275  276  277  278  279  280  281  282  283  284  285  286  287  288  289  290  291  292  293  294  295  296  297  298  299  300  301  302  303  304  305  306  307  308  309  310  311  312  313  314  315  316  317  318  319  320  321  322  323  324  325  326  327  328  329  330  331  332  333  334  335  336  337  338  339  340  341  342  343  344  345  346  347  348  349  350  351  352  353  354  355  356  357  358  359  360  361  362  363  364  365  366  367  368  369  370  371  372  373  374  375  376  377  378  379  380  381  382  383  384  385  386  387  388  389  390  391  392  393  394  395  396  397  398  399  400  401  402  403  404  405  406  407  408  409  410  411  412  413  414  415  416  417  418  419  420  421  422  423  424  425  426  427  428  429  430  431  432  433  434  435  436  437  438  439  440  441  442  443  444  445  446  447  448  449  450  451  452  453  454  455  456  457  458  459  460  461  462  463  464  465  466  467  468  469  470  471  472  473  474  475  476  477  478  479  480  481  482  483  484  485  486  487  488  489  490  491  492  493  494  495  496  497  498  499  500  501  502  503  504  505  506  507  508  509  510  511  512  513  514  515  516  517  518  519  520  521  522  523  524  525  526  527  528  529  530  531  532  533  534  535  536  537  538  539  540  541  542
2) f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    i    f    i    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f    i    f    f    i    f    f    i    i    f    f
3) aa  bc   dd   ff   gh   ij   kk    mm  no    pp   rr  st   uu    ww   xy   za   bb   dd  ef    gg   ii  jk   ll    nn  op   qr   ss    uu   vw   xx   zz  ab   cd    ee  fg    hi  jj    ll  mn   oo   qq   rs   tu   vv    xx  yz   aa   cc   de   ff    hh  ij   kl    mm   oo  pq    rr   tt   uv   wx   yy   aa  bc   dd    ff   gh   ii   kk   lm  no    pp   rr   st   uu   ww   xy   zz   bb   cd   ef  gg    ii  jk   ll   nn   op   qq    ss  tu   vw   xx   zz   ab    cc   ee   fg   hi   jj   ll  mn    oo   qq  rs   tt    vv   wx   yz   aa   cc   de  ff    hh   ij   kl   mm   oo   pq   rr   tt   uv   ww   yy   za   bc   dd   ff   gh   ii   kk   lm   nn   pp   qr   st   uu  vv
4) a    x    a    x    b    x    c    x    d    x    d    x    f    x    f    x    g    x    h    x    i    x    j    x    k    x    k    x    m    x    m    x    n    x    o    x    p    x    p    x    r    x    r    x    s    x    t    x    u    x    u    x    w    x    w    x    X    x    y    x    z    x    a    x    b    x    b    x    d    x    d    x    e    x    f    x    g    x    g    x    i    x    i    x    j    x    k    x    l    x    l    x    n    x    n    x    o    x    p    x    q    x    r    x    s    x    s    x    u    x    u    x    v    x    w    x    X    x    X    x    z    x    z    x    a    x    b    x    c    x    d    x    e    x    e    x    f    x    g    x    h    x    i    x    j    x    j    x    l    x    l    x    m    x    n    x    o    x    o    x    q    x    q    x    r    x    s    x    t    x    u    x    v    x    v    x    X    x    X    x    y    x    z    x    a    x    a    x    c    x    c    x    d    x    e    x    f    x    f    x    h    x    h    x    i    x    j    x    k    x    l    x    m    x    m    x    o    x    o    x    p    x    q    x    r    x    r    x    t    x    t    x    u    x    v    x    w    x    X    x    y    x    y    x    a    x    a    x    b    x    c    x    d    x    d    x    f    x    f    x    g    x    h    x    i    x    i    x    k    x    k    x    l    x    m    x    n    x    o    x    p    x    p    x    r    x    r    x    s    x    t    x    u    x    u    x    w    x    w    x    X    x    y    x    z    x    z    x    b    x    b    x    c    x    d    x    e    x    f    x    g    x    g    x    i    x    i    x    j    x    k    x    l    x    l    x    n    x    n    x    o    x    p    x    q    x    q    x    s    x    s    x    t    x    u    x    v    x    w    x    X    x    X    x    z    x    z    x    a    x    b    x    c    x    c    x    e    x    e    x    f    x    g    x    h    x    i    x    j    x    j    x    l    x    l    x    m    x    n    x    o    x    o    x    q    x    q    x    r    x    s    x    t    x    t    x    v    x    v    x    w    x    X    x    y    x    z    x    a    x    a    x    c    x    c    x    d    x    e    x    f    x    f    x    h    x    h    x    i    x    j    x    k    x    l    x    m    x    m    x    o    x    o    x    p    x    q    x    r    x    r    x    t    x    t    x    u    x    v    x    w    x    w    x    y    x    y    x    z    x    a    x    b    x    c    x    d    x    d    x    f    x    f    x    g    x    h    x    i    x    i    x    k    x    k    x    l    x    m    x    n    x    n    x    p    x    p    x    q    x    r    x    s    x    t    x    u    x    u    x    v    x    v
5) a    A    x    x    a    A    x    x    b    B    x    x    c    C    x    x    d    D    x    x    d    D    x    XT   f    F    x    x    f    F    x    x    g    G    x    x    h    H    x    x    i    I    x    x    j    J    x    x    k    K    x    x    k    K    x    XT   m    M
6)      n              n         n              n              n                        n              n                        n              n                        n              n                        n              n                        n             n                         n

Thank you!

[2Bdecided]

I have read your other thread, but honestly, I think you should do this...

selectevery(2,0)
bob()

step through in virtualdub and note the frame numbers of the near-duplicated frames.

Then create this code...
selectevery(2,0)
qtgmc()

adding the mvtools code from this thread...
http://forum.doom9.org/showthread.php?t=152758
...to replace each frame you noted as a bad frame above.

That's it. No need to re-interlace. You already have the frame rate you want, and there wasn't enough motion in the original for smooth 60fps motion anyway. You could create it at the end by using mflowfps to interpolate the whole lot to 60fps and then re-interlace, but using mflowfps that much might introduce artefacts. Your call.

Only process the slow-mo part. Don't touch the rest.

Cheers,
David.

P.S. the frames aren't duplicated - selectevery(2,0) is actually throwing away a fade on the titles which is unique to each frame - but I wouldn't worry about that.

[Dogway]

yes, it's a nice workaround. Not a solution but good enough to do it and end with this. The problem is that not all dups (or near dups) imply there is a time gap, note about ff gh. So the resulting framerate will be lower, around 25-ish I guess, so I will need something like:

assumefps(23.976)
interleave(last,last,last) # to 71fps
selectevery(6,0,1,3,4,5) # to 59.940

Your code is not ideal because some frames are not interlaced, so they are just like dups for the deinterlacer, plus the mentioned time gaps, so in a manner is like deinterlacing animation, something smart bobbers are not good with, better explained in this post from Didée. But I'm very tired of this now and need to keep going, if somebody wants some avisynth self challenge there already are 2 threads of background theorizing on this clip.

About interpolation I don't know which one to use. I have these 2. I remember reading something not good about ReplaceFramesSimple() so I dunno. Also all of them use MFlowInter instead of MFlowFPS:

function morph (clip c, int in, int "out", int "blksize")
{

    Function fill_loop(string s, int stop, int count)
    {
        return (stop == 0) ? s : string("tofill.mflowinter(morph_spr,fill_vb,fill_vf,time=" + string(100*(count - stop)/float(count))) + ",thscd1=255,thscd2=255).selectevery(tofill.framecount(),0)," + fill_loop(s,stop-1,count)
    }

out     = default(out, in+2)
blksize = default(blksize,16)

d=c.trim(in,out)# in-1?

tofill=d.selectevery(c.framecount(),0,c.framecount()-1)
morph_spr=tofill.msuper()
fill_vb=morph_spr.manalyse(truemotion=true,blksize=blksize,isb=true,overlap=blksize/2)
fill_vf=morph_spr.manalyse(truemotion=true,blksize=blksize,isb=false,overlap=blksize/2)
filled=eval("interleave(" + fill_loop("" ,d.framecount()-1,d.framecount()-1) + "tofill.selectevery(tofill.framecount(),tofill.framecount())).assumefps(c.framerate())")
c.trim(0,-(in+1))++filled.trim(1,d.framecount-2)++c.trim(out,0)}

function FixBadFrames(clip c, string frames) {
# Replace each frame from a list of 'bad' frames by using MFlowInter to interpolate
# between the nearest pair of 'good' frames
  c
  sup = MSuper()
  bv = MAnalyse(sup, isb=true, delta=2)
  fv = MAnalyse(sup, isb=false, delta=2)
  candidates = MFlowInter(sup, bv, fv, time=50.0, ml=100).DuplicateFrame(0)
  ReplaceFramesSimple(candidates, mappings=frames)
}

[2Bdecided]

yes, it's a nice workaround. Not a solution but good enough to do it and end with this. The problem is that not all dups (or near dups) imply there is a time gap, note about ff gh. So the resulting framerate will be lower, around 25-ish I guess

I think you must have misunderstood what I wrote, because it doesn't change the frame rate at all, or the number of frames.

Your code is not ideal because some frames are not interlaced, so they are just like dups for the deinterlacer

That doesn't really matter. A deinterlacer can cope with a still picture, and qtgmc is so close to lossless when this happens that it doesn't matter in this context. If the result is half-res because of the bastardised way it was generated (i.e. there's only one good field in the original), so be it. I guess you could use nnedi3 on those individual frames if you really wanted to.

What does matter is that (I think!) some top fields from the original not-slowed-down source have been pasted as bottom fields in the badly-slowed-down output that you have to work with. So there's a slight vertical bounce. But given the rather random zoom in, it hardly matters.

I'm amazed if you can, with 100% certainty, figure out the exact relationship between the original not-slowed-down source and the badly-slowed-down output that you have to work with. I agree you're probably right, but you can't know - which means that any automatically generated output that looks as good is, by definition, as good.

About interpolation I don't know which one to use. I have these 2. I remember reading something not good about ReplaceFramesSimple() so I dunno.

It can't do two adjacent bad frames, but I don't think you have any.

Motion interpolation is rarely perfect. If it looks OK and doesn't crash AVIsynth, be happy.

Also all of them use MFlowInter instead of MFlowFPS

I can't see what's wrong with that. Both use the same motion vectors, but one is easier for generating specific frames and specific times relative to existing ones, the other is easier for generating a string of new equal spaced frames which may each have different time relationships to each existing one. With appropriate calls, values and cutting, you could get the same output from either I think - but depending what you want to do, one is easier than the other.

I don't have as much patience as you though. My solution would take two minutes. It's what I posted above, plus ReplaceFramesSimple("some frame numbers here") - and those frame numbers would take maybe a minute to figure out just by stepping through the video output by my first script and writing down the near-dups.

Cheers,
David.

[Dogway]

Thanks for the feedback.
I think I understood your solution, but if I replace every output dup frame with an interpolated one there will be occasionally strings of 3 close frames, closer in time than the rest. I analyzed the source without all the dups and I only detected time gaps between for example dd ff, but not between ff and gh.

QTGMC always gave me problems if the source has not smooth motion. What I discovered recently and which works very good for animation is that by using the lossless mode you don't get ghosting. It's much slower but the best solution, although that doesn't mean I get ghosting in this case. I was just trying to be "academic" as per Didées post, but for this case exceptionally I don't care anymore.

Thanks for the clarification on MFlow, I never understood why there were 2 functions for the same task and documentation never explained. I always used MFlowFPS because there is nothing it can't do that MFlowInter does, but after some tests I got slightly better results with the latter, probably because the ml and time parameters.

About your solution yes it's fast, I'm just dealing with another matter right now. The second dup has a shift of one pixel upwards, I think always but need to check. So I'm trying to shift all these back to normal with:

str_b = "
0 1 0
0 0 0
0 0 0 1"
mt_edge(trim(21,-1),str_b,0,255,0,255,U=3,V=3)

and depending on which of both dups has better quality, keep this version or the other. With all the camera work involved it's hard to know.
Ok, time to get my hands working.

[Gavino]

About interpolation I don't know which one to use. I have these 2. I remember reading something not good about ReplaceFramesSimple() so I dunno.

It can't do two adjacent bad frames, but I don't think you have any.

Just to clarify this, I'm not aware of any problem with ReplaceFramesSimple itself.
What 2Bdecided is saying (correctly) is that my FixBadFrames function (which uses ReplaceFramesSimple) is only designed to handle isolated bad frames and can't fix adjacent bad frames.
It's strength lies in being simple to use, as you supply all the bad frames in a list in a single call to the function.

[Dogway]

I had something in the back of my mind with remapframessimple, probably regarding multithreading and also some issues pointed when clipclop was released. Anyways I already filtered the clip and it looks nice knowing the garbage where it came from. In the end MFlowInter feels to perform better specially when interpolation is wrong, instead of mangling the pixels it uses a blend approach or so it seems.

45Mb
http://www.mediafire.com/?7i4b9serdq37kka

I proceed to post my final script and procedures:

setmtmode(3,2)
MPEG2Source("source.d2v", cpu=0)
setmtmode(2,2)

Trim(0,926)
separatefields()

#chroma fix
interleave(selectevery(4,0),selectevery(4,1),mergechroma(selectevery(4,2),selectevery(4,-0)),selectevery(4,3))
a=weave.trim(504,0)

# drop dups
b=0
selectevery(4,b+2,b+3)

t=trim(774,926).selecteven.TemporalSoften(10,255,255,25,2)
trim(0,773)+interleave(t,t) # Fixing all the crap out of the fields

# back to initial state
weave()

# start from problematic section
trim(252,0)


# Deinterlace
QTGMC(tr2=3,preset="slower",lsb=true,chromamotion=false,edithreads=2)

# Remove garbage dups (the worst of the pair not replacing any time gap frame)
deleteframe(7, 14, 20, 26, 35, 40, 47, 54, 58, 66, 72, 80, 89, 94, 100, 106, 114, 122, 129, 134, 143, 149, 154, 163, 166, 174, 183, 189, 196, 202, 206, 216, 220, 228, 234, 240, 248, 254, 260, 269)
# Replace time gap dups
fixbadframes("5 12 17 22 29 34 39 46 51 63 68 75 80 85 92 97 104 109 114 121 126 131 138 143 148 155 160 167 172 177 184 189 196 201 206 213 218 223 230")
fixed=last

# Block fix for last part shimmer
trim(231,0)
str_b = "
0 1 0 
0 0 0 
0 0 0 1"
interleave(selectevery(2,0),mt_edge(selectevery(2,1),str_b,0,255,0,255,U=3,V=3))
sharp=last

DeStripeV(1)
TemporalSoften(10,255,255,25,2)
Contrasharpening(last,sharp)

fixed.trim(0,230)+last


# Delete first frame Dup
# Convert to 59.940 and loop section to comply to original duration
trim(1,0)
loop(2,290,330)
interleave(last,last)
assumefps(60000,1001)

And this is how I got the numbers strings:

# 1) Numbering
# 2) fields distribution
# 3) Removed not time concerned dups
# 4) Rearranged 3) to know the numbering of left dups (dups that need to be replaced with a motion interpolated substitute)

# 1) 0  1  2  3  4  5  6  7  8  9  10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 253 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
# 2) a  a  b  c  d  d  f  f  g  h  i  j  k  k  m  m  n  o  p  p  r  r  s  t  u  u  w  w  x  y  z  a  b  b  d  d  e  f  g  g  i  i  j  k  l  l  n  n  o  p  q  r  s  s  u  u  v  w  x  x  z  z  a  b  c  d  e  e  f  g  h  i  j  j  l  l  m  n  o  o  q  q  r  s  t  u  v  v  x  x  y  z  a  a  c  c  d  e  f  f  h   h   i   j   k   l   m   m   o   o   p   q   r   r   t   t   u   v   w   x   y   y   a   a   b   c   d   d   f   f   g   h   i   i   k   k   l   m   n   o   p   p   r   r   s   t   u   u   w   w   x   y   z   z   b   b   c   d   e   f   g   g   i   i   j   k   l   l   n   n   o   p   q   q   s   s   t   u   v   w   x   x   z   z   a   b   c   c   e   e   f   g   h   i   j   j   l   l   m   n   o   o   q   q   r   s   t   t   v   v   w   x   y   z   a   a   c   c   d   e   f   f   h   h   i   j   k   l   m   m   o   o   p   q   r   r   t   t   u   v   w   w   y   y   z   a   b   c   d   d   f   f   g   h   i   i   k   k   l   m   n   n   p   p   q   r   s   t   u   u   v   v
# 3) a  a  b  c  d  d  f     g  h  i  j  k  k     m  n  o  p  p     r  s  t  u  u     w  x  y  z  a  b  b  d     e  f  g  g     i  j  k  l  l  n     o  p  q  r  s  s     u  v  w     x  z  z  a  b  c  d     e  f  g  h  i     j  l  l  m  n  o  o     q  r  s  t  u  v  v  x     y  z  a  a     c  d  e  f  f      h   i   j   k   l       m   o   o   p   q   r   r       t   u   v   w   x   y   y       a   b   c   d   d   f       g   h   i   i       k   l   m   n   o   p   p   r       s   t   u   u   w       x   y   z   z       b   c   d   e   f   g   g   i       j   k       l   n   n   o   p   q   q       s   t   u   v   w   x   x   z       a   b   c   c   e       f   g   h   i   j   j       l   m   n   o   o       q   r   s       t   v   v   w   x   y   z   a   a       c   d   e       f   h   h   i   j   k   l       m   o   o   p   q       r   t   t   u   v       w   y   y   z   a   b   c       d   f   f   g   h       i   k   k   l   m       n   p   p   q   r   s   t   u       v
# 4) a  a  b  c  d  d  f  g  h  i  j  k  k  m  n  o  p  p  r  s  t  u  u  w  x  y  z  a  b  b  d  e  f  g  g  i  j  k  l  l  n  o  p  q  r  s  s  u  v  w  x  z  z  a  b  c  d  e  f  g  h  i  j  l  l  m  n  o  o  q  r  s  t  u  v  v  x  y  z  a  a  c  d  e  f  f  h  i  j  k  l  m  o  o  p  q  r  r  t  u  v   w   x   y   y   a   b   c   d   d   f   g   h   i   i   k   l   m   n   o   p   p   r   s   t   u   u   w   x   y   z   z   b   c   d   e   f   g   g   i   j   k   l   n   n   o   p   q   q   s   t   u   v   w   x   x   z   a   b   c   c   e   f   g   h   i   j   j   l   m   n   o   o   q   r   s   t   v   v   w   x   y   z   a   a   c   d   e   f   h   h   i   j   k   l   m   o   o   p   q   r   t   t   u   v   w   y   y   z   a   b   c   d   f   f   g   h   i   k   k   l   m   n   p   p   q   r   s   t   u   v

I just bookmarked in avspmod the dupped fields/frames, and made a macro to text type the bookmarks so it can used for any kind of filter.

Bookmarks Frame Number to Typed Number.py
http://www.mediafire.com/?5dlxtdpn3phm67h