# Raiders of the Lost Ark AVS - #2

# 'Raiders of the Lost Ark' is almost exactly two hours long, and takes up two
# laserdisc sides, resulting in two capture files.  While I could have made a single
# CD-sized AVI from these two captures, I decided instead to create two AVIs, one
# from each capture, to preserve a higher amount of quality.

# The second AVI comes directly from the second capture.
AVISource("e:\Scratch Space\ROTLA2.avi")

# The telecine pattern breaks very often during the second half of 'Raiders,' and
# in some scenes the five-frame pattern isn't being kept at all!  I suspect that
# certain parts of the movie were given a very slight speed-up during the transfer,
# so that the movie could be fit onto a single laserdisc.  Whatever the reason,
# there are a ton of breaks to find and fix.  Doing so results in several
# noticeable hiccups in the audio and video, but I decided this is an acceptable
# compromise.
#
# These frames must be manipulated to preserve the telecine pattern:
# 4767 repeat this frame
# 4768 repeat this frame
# 24203 delete this frame
# 33755 delete this frame
# 33757 delete this frame
# 33777 repeat this frame
# 33808 repeat this frame
# 33840 repeat this frame
# 33872 repeat this frame
# 33903 repeat this frame
# 33934 repeat this frame
# 33966 repeat this frame
# 33997 repeat this frame
# 34029 repeat this frame
# 34061 repeat this frame
# 34092 repeat this frame
# 34124 repeat this frame
# 36964 delete this frame
# 36965 delete this frame
# 36987 repeat this frame
# 37019 repeat this frame
# 37050 repeat this frame
# 37082 repeat this frame
# 37113 repeat this frame
# 37144 repeat this frame
# 37176 repeat this frame
# 37208 repeat this frame
# 37239 repeat this frame
# 37271 repeat this frame
# 37302 repeat this frame
# 37334 repeat this frame
# 37365 repeat this frame
# 37385 delete this frame
# 38199 delete this frame
# 38204 repeat this frame
# 38210 repeat this frame
# 38217 repeat this frame
# 38223 repeat this frame
# 38230 repeat this frame
# 38236 repeat this frame
# 38243 repeat this frame
# 38250 delete this frame
# 39878 delete this frame
# 39886 repeat this frame
# 39895 repeat this frame
# 39904 repeat this frame
# 39911 repeat this frame
# 39920 repeat this frame
# 39929 repeat this frame
# 42985 delete this frame
# 42987 delete this frame
# 59326 repeat this frame
# 77380 repeat this frame
# 77381 repeat this frame
# 93740 delete this frame
# 96911 delete this frame
# 96912 delete this frame
# 98726 repeat this frame
#
# Note:  AVISynth does not have "soft" line breaks, so it is necessary for all
# this to be on a single line:
Trim(0,4767) ++ Trim(4767,4768) ++ Trim(4768,24202) ++ Trim(24204,33754) ++ Trim(33756,33756) ++ Trim(33758,33777) ++ Trim(33777,33808) ++ Trim(33808,33840) ++ Trim(33840,33872) ++ Trim(33872,33903) ++ Trim(33903,33934) ++ Trim(33934,33966) ++ Trim(33966,33997) ++ Trim(33997,34029) ++ Trim(34029,34061) ++ Trim(34061,34092) ++ Trim(34092,34124) ++ Trim(34124,36963) ++ Trim(36965,36987) ++ Trim(36987,37019) ++ Trim(37019,37050) ++ Trim(37050,37082) ++ Trim(37082,37113) ++ Trim(37113,37144) ++ Trim(37144,37176) ++ Trim(37176,37208) ++ Trim(37208,37239) ++ Trim(37239,37271) ++ Trim(37271,37302) ++ Trim(37302,37334) ++ Trim(37334,37365) ++ Trim(37365,37384) ++ Trim(37386,38198) ++ Trim(38200,38204) ++ Trim(38204,38210) ++ Trim(38210,38217) ++ Trim(38217,38223) ++ Trim(38223,38230) ++ Trim(38230,38236) ++ Trim(38236,38243) ++ Trim(38243,38249) ++ Trim(38251,39877) ++ Trim(39879,39886) ++ Trim(39886,39895) ++ Trim(39895,39904) ++ Trim(39904,39911) ++ Trim(39911,39920) ++ Trim(39920,39929) ++ Trim(39929,42984) ++ Trim(42986,42986) ++ Trim(42988,59326) ++ Trim(59326,77380) ++ Trim(77380,77381) ++ Trim(77381,93739) ++ Trim(93741,96910) ++ Trim(96913,98726) ++ Trim(98726,0)

# Remove the letterboxing bars, keeping resolution numbers divisible by 8.
Crop(0,96,640,280)

# A small amount of visual noise filtering.
TemporalSoften(5,3,6)

# For certain routines, AVISynth internally divides each video frame into two
# "fields."  One field contains the frame's even rows of pixels, while the other
# field has the odd rows.  When untelecining a movie, the appropriate AVISynth
# functions use these fields to stitch the frames containing mixed pictures into
# frames containing single pictures.  However, this requires not only knowing which
# frames contain mixed pictures (defined by the Doubleweave and Pulldown functions
# below), but also knowing which of the two possible ways the fields within those
# frames should be spliced together with surrounding fields to restore the original
# pictures.  AVISynth assumes one way, but sometimes the other way is needed.
#
# To make a long explanation shorter, the AssumeFrameBased and ComplementParity
# functions below allow these captures to be properly untelecined.
AssumeFrameBased.ComplementParity


# Untelecine.
DoubleWeave().Pulldown(0,3)

# To figure out how to make Pulldown match the telecine pattern, define five clips,
# each one using one of the five possible Pulldown routines.  Then, stack these five
# clips on top of each other to compare them, using StackVertical.  To make sure the
# resulting picture isn't too big, crop each individual picture down to its first
# 100 rows of pixels:
# a = Crop(0,0,640,100).Pulldown(0,2)
# b = Crop(0,0,640,100).Pulldown(0,3)
# c = Crop(0,0,640,100).Pulldown(1,3)
# d = Crop(0,0,640,100).Pulldown(1,4)
# e = Crop(0,0,640,100).Pulldown(2,4)
# StackVertical(a,b,c,d,e)
#
# Next, save the AVS script and open it in VirtualDub or other player capable of
# examining AVIs frame by frame.  When advancing the movie by single frames, two
# of the five pictures will look alright.  The rest will show frames that are still
# mixed together.  Either of the two successful Pulldowns can be used.
#
# If all five pictures still show mixed frames, use the AssumeFrameBased and
# ComplementParity functions described above and try again.