transcode(filter) 3th February 2008 transcode(filter)NAME
transcode - filter modules collection
SYNOPSIS
transcode -J filters [other options]
COPYRIGHT
transcode is Copyright (C) 2001-2003 by Thomas Oestreich, 2003-2004
Tilmann Bitterberg, 2004-2008 Transcode Team
MODULE OPTIONS
The syntax for filter options is simple. A filter is specified with
-J filter=optionstring
The optionstring can contain multiple options which are separated by
colons `:´
-J filter=option1:option2:option3
An option can have an argument or non (bool). For options with an argu‐
ment, the format in which the argument has to be given to the option is
specified in a printf(1) like string. The most common case is `%d'
which simply means a number. The argument has to be seperated from the
option by a `=´.
-J filter=bool1:option1=15:option2=20x30
Most filters try to do the right thing with the default options. You
should play with various parameters if you are not satisfied with the
default behaviour. If you have no idea what a filter does, its very
likely that you don't need it.
If a filter takes (for eg.) a filename as an argument, make sure that
the filename does not contain a `:' or a `='. Its a limitation of the
parser. A comma `,' is possible but must be extra quoted. For the text
filter that is
-J text=string="Hello\, World"
FILTER MODULES
To see what filter modules are avalaible for your transcode installa‐
tion, do a
ls -1 $( tcmodinfo -p )/filter*.so
A complete transcode installation has the following filter modules:
29to23 - frame rate conversion filter (interpolating 29 to 23)
29to23 was written by Max Alekseyev, Tilmann Bitterberg. The ver‐
sion documented here is v0.3 (2003-07-18). This is a video filter.
It can handle RGB and YUV mode. It is a pre-processing only filter.
32detect - 3:2 pulldown / interlace detection plugin
32detect was written by Thomas. The version documented here is
v0.2.4 (2003-07-22). This is a video filter. It can handle RGB and
YUV mode. It supports multiple instances and can run as a pre-pro‐
cessing and/or as a post-processing filter.
· threshold = %d [default 9]
Interlace detection threshold
· chromathres = %d [default 4]
Interlace detection chroma threshold
· equal = %d [default 10]
threshold for equal colors
· chromaeq = %d [default 5]
threshold for equal chroma
· diff = %d [default 30]
threshold for different colors
· chromadi = %d [default 15]
threshold for different chroma
· force_mode = %d [default 0]
set internal force de-interlace flag with mode -I N
· pre = %d [default 1]
run as pre filter
· verbose (bool)
show results
This filter checks for interlaced video frames. Subsequent
de-interlacing with transcode can be enforced with
'force_mode' option
32drop - 3:2 inverse telecine removal plugin
32drop was written by Thomas Oestreich. The version documented here
is v0.4 (2003-02-01). This is a video filter. It can handle RGB and
YUV mode. It is a pre-processing only filter.
aclip - generate audio clips from source
aclip was written by Thomas Oestreich. The version documented here
is v0.1.1 (2003-09-04). This is a audio filter. It is a pre-pro‐
cessing only filter.
· level = %d [default 10]
The audio must be under this level to be skipped
· range = %d [default 25]
Number of samples over level will be keyframes
astat - audio statistics filter plugin
astat was written by Thomas Oestreich. The version documented here
is v0.1.3 (2003-09-04). This is a audio filter. It is a pre-pro‐
cessing only filter.
· file = %s
File to save the calculated volume rescale number to
compare - compare with other image to find a pattern
compare was written by Antonio Beamud. The version documented here
is v0.1.2 (2003-08-29). This is a video filter. It can handle RGB
and YUV mode. It supports multiple instances. It is a post-process‐
ing only filter.
· pattern = %s
Pattern image file path
· results = %s
Results file path
· delta = %f [default 45.000000]
Delta error
Generate a file in with information about the times, frame,
etc the pattern defined in the image parameter is observed.
The format of the command file is framenumber followed by at least
one whitespace followed by the command followed by at least one
whitespace followed by arguments for the command. Empty lines and
lines starting with a `#' are ignored. The frame numbers must be
sorted ascending.
# Example file
# At frame 10 load the smooth filter
10 load smooth
# reconfigure at 20
20 configure smooth=strength=0.9
99 disable smooth
cpaudio - copy one audio channel to the other channel filter plugin
cpaudio was written by William H Wittig. The version documented
here is v0.1 (2003-04-30). This is a audio filter. It is a post-
processing only filter.
· source = %c [default l]
Source channel (l=left, r=right)
Copies audio from one channel to another
decimate - NTSC decimation plugin
decimate was written by Thanassis Tsiodras. The version documented
here is v0.4 (2003-04-22). This is a video filter. It can handle
YUV mode only. It is a post-processing only filter.
· verbose (bool)
print verbose information
see /docs/README.Inverse.Telecine.txt
denoise3d - High speed 3D Denoiser
denoise3d was written by Daniel Moreno & A'rpi. The version docu‐
mented here is v1.0.6 (2003-12-20). This is a video filter. It can
handle YUV mode only. It supports multiple instances. It can be
used as a pre-processing or as a post-processing filter.
· luma = %f [default 4.000000]
spatial luma strength
· chroma = %f [default 3.000000]
spatial chroma strength
· luma_strength = %f [default 6.000000]
temporal luma strength
· chroma_strength = %f [default 4.000000]
temporal chroma strength
· pre = %d [default 0]
run as a pre filter
What: The denoise3d filter from mplayer (sibling of hqdn3d).
Works very crude and simple but also very fast. In fact it
is even faster than the original from mplayer as I managed
to tweak some things (a.o. zero frame copying).
Who: Everyone who wants to have their captured frames thor‐
oughly denoised (i.e. who want to encode to mpeg or mjpeg)
but do not have enough processing power to real-time encode
AND use hqdn3d (better quality but a lot slower) or dnr (yet
slower), not to mention the other denoisers that are even
slower. Quality is really good for static scenes (if fed
with the right parameters), moving objects may show a little
ghost-image (also depends on parameters) though. Your milage
may vary.
How: Parameters are the same as the hqdn3d module, although
in practice you'll not end up with exactly the same values.
Just experiment. Particular for this version of the filter
is that if you supply -1 to either component's parameters
(luma/chroma), that component will not have the filter
applied to. If you're still short on CPU cycles, try dis‐
abling the luma filter, this will not make much difference
in the effectiveness of the filter!
detectclipping - detect clipping parameters (-j or -Y)
detectclipping was written by Tilmann Bitterberg, A'rpi. The ver‐
sion documented here is v0.1.0 (2003-11-01). This is a video fil‐
ter. It can handle RGB and YUV mode. It supports multiple instances
and can run as a pre-processing and/or as a post-processing filter.
· range = %u-%u/%d [default 0-4294967295/1]
apply filter to [start-end]/step frames
· limit = %d [default 24]
the sum of a line must be below this limit to be considered as
black
· post (bool)
run as a POST filter (calc -Y instead of the default -j)
Detect black regions on top, bottom, left and right of an
image. It is suggested that the filter is run for around
100 frames. It will print its detected parameters every
frame. If you don't notice any change in the printout for a
while, the filter probably won't find any other values. The
filter converges, meaning it will learn.
detectsilence - audio silence detection with tcmp3cut commandline gen‐
eration
detectsilence was written by Tilmann Bitterberg. The version docu‐
mented here is v0.0.1 (2003-07-26). This is a audio filter. It is a
pre-processing only filter.
divxkey - check for DivX 4.xx / OpenDivX / DivX;-) keyframe
divxkey was written by Thomas Oestreich. The version documented
here is v0.1 (2002-01-15). This is a video filter. It is a pre-pro‐
cessing only filter.
dnr - dynamic noise reduction
dnr was written by Gerhard Monzel. The version documented here is
v0.2 (2003-01-21). This is a video filter. It can handle RGB and
YUV mode. It is a post-processing only filter.
· lt = %d [default 10]
Threshold to blend luma/red
· ll = %d [default 4]
Threshold to lock luma/red
· ct = %d [default 16]
Threshold to blend croma/green+blue
· cl = %d [default 8]
Threshold to lock croma/green+blue
· sc = %d [default 30]
Percentage of picture difference (scene change)
see /docs/filter_dnr.txt (german only)
doublefps - double frame rate by deinterlacing fields into frames
doublefps was written by Andrew Church. The version documented here
is v1.1 (2006-05-14). This is a video and audio filter. It can han‐
dle YUV and YUV422 modes. It is a pre-processing only filter.
· topfirst = %d [default 0 when height=480, else 1]
Assume top field is displayed first
· fullheight = %d [default 0]
Output full-height frames
Doubles the frame rate of interlaced video by separating each field
into a separate frame. The fields can either be left as is (giving
a progessive video with half the height of the original) or re-
interlaced into their original height (at the doubled frame rate)
for the application of a separate deinterlacing filter.
Note that due to transcode limitations, it is currently necessary
to use the -Z option to specify the output frame size when using
half-height mode (this does not slow the program down if no actual
zooming is done).
extsub - DVD subtitle overlay plugin
extsub was written by Thomas Oestreich. The version documented here
is 0.3.5 (2003-10-15). This is a video filter. It can handle RGB
and YUV mode. It can be used as a pre-processing or as a post-pro‐
cessing filter.
· track = %d [default 0]
Subtitle track to render
· vertshift = %d [default 0]
offset of subtitle with respect to bottom of frame in rows
· timeshift = %d [default 0]
global display start time correction in msec
· antialias = %d [default 1]
anti-aliasing the rendered text (0=off,1=on)
· pre = %d [default 1]
Run as a pre filter
· color1 = %d [default 0]
Make a subtitle color visible with given intensity
· color2 = %d [default 0]
Make a subtitle color visible with given intensity
· ca = %d [default 0]
Shuffle the color assignment by choosing another subtitle color
· cb = %d [default 0]
Shuffle the color assignment by choosing another subtitle color
fieldanalysis - Field analysis for detecting interlace and telecine
fieldanalysis was written by Matthias Hopf. The version documented
here is v1.0 pl1 (2004-08-13). This is a video filter. It can han‐
dle RGB,YUV and YUV422 mode. It is a pre-processing only filter.
· interlacediff = %f [default 1.1]
Minimum temporal inter-field difference for detecting interlaced
video
· unknowndiff = %f [default 1.5]
Maximum inter-frame change vs. detail differences for neglecting
interlaced video
· progressivediff = %f [default 8]
Minimum inter-frame change vs. detail differences for detecting
progressive video
· progressivechange = %f [default 0.2]
Minimum temporal change needed for detecting progressive video
· changedifmore = %f [default 10]
Minimum temporal change for detecting truly changed frames
· forcetelecinedetect = %d [default 0]
Detect telecine even on non-NTSC (29.97fps) video
· verbose = %d [default 0]
Output analysis for every frame
· outdiff = %d [default 0]
Output internal debug frames as luminance of YUV video (see
source)
fields - Field adjustment plugin
fields was written by Alex Stewart. The version documented here is
v0.1.1 (2003-01-21). This is a video filter. It can handle RGB and
YUV mode. It is a pre-processing only filter.
· flip (bool)
Exchange the top field and bottom field of each frame
· shift (bool)
Shift the video by one field
· flip_first (bool)
Normally shifting is performed before flipping, this option
reverses that
The 'fields' filter is designed to shift, reorder, and gen‐
erally rearrange independent fields of an interlaced video
input. Input retrieved from broadcast (PAL, NTSC, etc)
video sources generally comes in an interlaced form where
each pass from top to bottom of the screen displays every
other scanline, and then the next pass displays the lines
between the lines from the first pass. Each pass is known
as a "field" (there are generally two fields per frame).
When this form of video is captured and manipulated digi‐
tally, the two fields of each frame are usually merged
together into one flat (planar) image per frame. This usu‐
ally produces reasonable results, however there are condi‐
tions which can cause this merging to be performed incor‐
rectly or less-than-optimally, which is where this filter
can help.
The following options are supported for this filter (they
can be separated by colons):
shift - Shift the video by one field (half a frame),
changing frame boundaries appropriately. This is
useful if a video capture started grabbing video
half a frame (one field) off from where frame
boundaries were actually intended to be.
flip - Exchange the top field and bottom field of each
frame. This can be useful if the video signal was
sent "bottom field first" (which can happen
sometimes with PAL video sources) or other
oddities occurred which caused the frame
boundaries to be at the right place, but the
scanlines to be swapped.
flip_first
- Normally shifting is performed before flipping if
both are specified. This option reverses that
behavior. You should not normally need to use
this unless you have some extremely odd input
material, it is here mainly for completeness.
help - Print this text.
Note: the 'shift' function may produce slight color discrep‐
ancies if YV12 is used as the internal transcode video for‐
mat (default). This is because YV12 does not contain enough
information to do field shifting cleanly. For best (but
slower) results, use RGB mode for field shifting.
fps - convert video frame rate, gets defaults from -f and --export_fps
fps was written by Christopher Cramer. The version documented here
is v1.1 (2004-05-01). This is a video filter. It can handle RGB and
YUV mode. It can be used as a pre-processing or as a post-process‐
ing filter.
options: <input fps>:<output fps> example: -J fps=25:29.97
will convert from PAL to NTSC If no options are given,
defaults or -f/--export_fps/--export_frc will be used. Some
examples:
-J fps=10:5:pre convert from 10 fps to 5 fps,
preprocess -J fps=10:post:12 convert from 10 to 12,
postprocess -J fps=pre=1:7:5 convert from 7 to 5,
postprocess -J fps=9:3.1:post=-0x7 convert from 9 to
3.1, postprocess
If that last one is confusing you, remember that 0 is false
and everything else is true. Of course, octal and hexadeci‐
mal numbers are supported too. This is intended to be back‐
ward compatible with the old format.
hqdn3d - High Quality 3D Denoiser
hqdn3d was written by Daniel Moreno & A'rpi. The version documented
here is v1.0.2 (2003-08-15). This is a video filter. It can handle
YUV mode only. It supports multiple instances. It can be used as a
pre-processing or as a post-processing filter.
· luma = %f [default 4.000000]
spatial luma strength
· chroma = %f [default 3.000000]
spatial chroma strength
· luma_strength = %f [default 6.000000]
temporal luma strength
· chroma_strength = %f [default 4.500000]
temporal chroma strength
· pre = %d [default 0]
run as a pre filter
This filter aims to reduce image noise producing smooth
images and making still images really still (This should
enhance compressibility).
invert - invert the image
invert was written by Tilmann Bitterberg. The version documented
here is v0.1.4 (2003-10-12). This is a video filter. It can handle
RGB,YUV and YUV422 mode. It is a post-processing only filter.
· range = %u-%u/%d [default 0-4294967295/1]
apply filter to [start-end]/step frames
ivtc - NTSC inverse telecine plugin
ivtc was written by Thanassis Tsiodras. The version documented here
is v0.4.1 (2004-06-01). This is a video filter. It can handle YUV
mode only. It is a pre-processing only filter.
· verbose (bool)
print verbose information
· field = %d [default 0]
which field to replace (0=top 1=bottom)
· magic = %d [default 0]
perform magic? (0=no 1=yes)
see /docs/README.Inverse.Telecine.txt
levels - Luminosity level scaler
levels was written by Bryan Mayland. The version documented here is
v1.0.0 (2004-06-09). This is a video filter. It can handle YUV mode
only. It supports multiple instances. It is a post-processing only
filter.
· input = %d-%d [default 0-255]
input luma range (black-white)
· gamma = %f [default 1.000000]
input luma gamma
· output = %d-%d [default 0-255]
output luma range (black-white)
logo - render image in videostream
logo was written by Tilmann Bitterberg. The version documented here
is v0.10 (2003-10-16). This is a video filter. It can handle RGB
and YUV mode. It is a post-processing only filter.
· file = %s
Image filename
· posdef = %d [default 0]
Position (0=None, 1=TopL, 2=TopR, 3=BotL, 4=BotR, 5=Center)
· pos = %dx%d [default 0x0]
Position (0-width x 0-height)
· range = %u-%u [default 0-0]
Restrict rendering to framerange
· ignoredelay (bool)
Ignore delay specified in animations
· rgbswap (bool)
Swap red/blue colors
· grayout (bool)
YUV only: don't write Cb and Cr, makes a nice effect
· flip (bool)
Mirror image
This filter renders an user specified image into the video.
Any image format ImageMagick can read is accepted. Trans‐
parent images are also supported. Image origin is at the
very top left.
see /docs/filter_logo.txt
logoaway - remove an image from the video
logoaway was written by Thomas Wehrspann <thomas@wehrspann.de>. The
version documented here is v0.5 (2004-03-07). This is a video fil‐
ter. It can handle RGB and YUV mode. It is a post-processing only
filter. It supports multiple instances.
· range = %d-%d [default 0-4294967295]
Frame Range
· pos = %dx%d [default -1x-1]
Position of logo
· size = %dx%d [default -2x-2]
Size of logo
· mode = %d [default 0]
Filter Mode (0=none,1=solid,2=xy,3=shape)
· border (bool)
Visible Border
· dump (bool)
Dump filterarea to file
· xweight = %d [default 50]
X-Y Weight(0%-100%)
· fill = %2x%2x%2x [default 000]
Solid Fill Color(RGB)
· file = %s
Image with alpha/shape information
This filter removes an image in a user specified area from
the video. You can choose from different methods.
see /docs/filter_logoaway.txt
lowpass - High and low pass filter
lowpass was written by Tilmann Bitterberg. The version documented
here is v0.1.0 (2002-02-26). This is a audio filter. It is a pre-
processing only filter.
· taps = %d [default 30]
strength (may be negative)
mask - Filter through a rectangular Mask
mask was written by Thomas Oestreich, Chad Page. The version docu‐
mented here is v0.2.3 (2003-10-12). This is a video filter. It can
handle RGB,YUV and YUV422 mode. It is a pre-processing only filter.
· lefttop = %dx%d [default 0x0]
Upper left corner of the box
· rightbot = %dx%d [default 32x32]
Lower right corner of the box
This filter applies an rectangular mask to the video.
Everything outside the mask is set to black.
modfps - plugin to modify framerate
modfps was written by Marrq. The version documented here is v0.10
(2003-08-18). This is a video filter. It can handle RGB and YUV
mode. It is a pre-processing only filter.
· mode = %d [default 1]
mode of operation
· infps = %f [default 25.000000]
Original fps
· infrc = %d [default 0]
Original frc
· buffer = %d [default 5]
How many frames to buffer
· subsample = %d [default 32]
How many pixels to subsample
· clonetype = %d [default 0]
How to clone frames
· verbose = %d [default 1]
run in verbose mode
This filter aims to allow transcode to alter the fps of
video. While one can reduce the fps to any amount, one can
only increase the fps to at most twice the original fps.
There are two modes of operation, buffered and unbuffered,
unbuffered is quick, but buffered, especially when dropping
frames should look better.
For most users, modfps will need either no options, or just
mode=1
see /docs/README.filter.modfps
msharpen - VirtualDub's MSharpen Filter
msharpen was written by Donald Graft, William Hawkins. The version
documented here is (1.0) (2003-07-17). This is a video filter. It
can handle RGB and YUV mode. It is a post-processing only filter.
· strength = %d [default 100]
How much of the effect
· threshold = %d [default 10]
How close a pixel must be to the brightest or dimmest pixel to
be mapped
· highq = %d [default 1]
Tradeoff speed for quality of detail detection
· mask = %d [default 0]
Areas to be sharpened are shown in white
This plugin implements an unusual concept in spatial sharp‐
ening. Although designed specifically for anime, it also
works well with normal video. The filter is very effective
at sharpening important edges without amplifying noise.
* Strength 'strength' (0-255) [100]
This is the strength of the sharpening to be applied to
the edge detail areas. It is applied only to the edge detail
areas as determined by the 'threshold' parameter. Strength
255 is the strongest sharpening.
* Threshold 'threshold' (0-255) [10]
This parameter determines what is detected as edge
detail and thus sharpened. To see what edge detail areas
will be sharpened, use the 'mask' parameter.
* Mask 'mask' (0-1) [0]
When set to true, the areas to be sharpened are shown in
white against a black background. Use this to set the level
of detail to be sharpened. This function also makes a basic
edge detection filter.
* HighQ 'highq' (0-1) [1]
This parameter lets you tradeoff speed for quality of
detail detection. Set it to true for the best detail detec‐
tion. Set it to false for maximum speed.
normalize - Volume normalizer
normalize was written by pl, Tilmann Bitterberg. The version docu‐
mented here is v0.1.1 (2002-06-18). This is a audio filter. It is a
pre-processing only filter.
· smooth = %f [default 0.06]
Value for smoothing ]0.0 1.0[
· smoothlast = %f [default 0.06]
Value for smoothing last sample ]0.0, 1.0[
· algo = %d [default 1]
Algorithm to use (1 or 2). 1=uses a 1 value memory and coeffi‐
cients new=a*old+b*cur (with a+b=1). 2=uses several samples to
smooth the variations (standard weighted mean on past samples)
null - demo filter plugin; does nothing
null was written by Thomas Oestreich. The version documented here
is v0.2 (2003-09-04). This is a video and audio filter. It can be
used as a pre-processing or as a post-processing filter.
pp - Mplayers postprocess filters
pp was written by Michael Niedermayer et al, Gerhard Monzel. The
version documented here is v1.2.4 (2003-01-24). This is a video
filter. It can handle YUV mode only. It supports multiple
instances. It can be used as a pre-processing or as a post-process‐
ing filter.
· hb = %d:%d [default 64:40]
Horizontal deblocking filter
· vb = %d:%d [default 64:40]
Vertical deblocking filter
· h1 (bool)
Experimental h deblock filter 1
· v1 (bool)
Experimental v deblock filter 1
· dr (bool)
Deringing filter
· al (bool)
Automatic brightness / contrast
· f (bool)
Stretch luminance to (0..255)
· lb (bool)
Linear blend deinterlacer
· li (bool)
Linear interpolating deinterlace
· ci (bool)
Cubic interpolating deinterlacer
· md (bool)
Median deinterlacer
· de (bool)
Default preset (hb:a/vb:a/dr:a/al)
· fa (bool)
Fast preset (h1:a/v1:a/dr:a/al)
· tn = %d:%d:%d [default 64:128:256]
Temporal Noise Reducer (1<=2<=3)
· fq = %d [default 15]
Force quantizer
· pre (bool)
Run as a PRE filter
preview - xv/sdl/gtk preview plugin
preview was written by Thomas Oestreich. The version documented
here is v0.1.4 (2002-10-08). This is a video filter. It can handle
RGB and YUV mode. It is a post-processing only filter.
XXX: Write me
pv - xv only preview plugin
pv was written by Thomas Oestreich, Tilmann Bitterberg. The version
documented here is v0.2.3 (2004-06-01). This is a video filter. It
can handle YUV and YUV422 mode. It is a post-processing only fil‐
ter.
· cache = %d [default 15]
Number of raw frames to cache for seeking
· skip = %d [default 0]
display only every Nth frame
· fullscreen (bool)
Display in fullscreen mode
The filter listens to mouse and key strokes. If you click
into the preview window, the first time say near the upper
left corner and the second time near the lower right corner,
transcode will draw a rectangle and will print out the coor‐
dinates of this rectangle on stdout and the socket. See the
table below for available keys.
When you start transcode with the --socket option and the pv
filter with (for example) cache=20 you can talk to transcode
and the pv filter at runtime using the socket.
transcode -i file.avi -J pv=cache=30 --socket /tmp/sock
Available Commands
Key Socket* Effect
RET draw redraws the image, applying filters.
u undo goes to image before draw
SPACE pause pause the preview (and transcode)
UP fastfw in pause mode, step forward 5 frames
RIGHT slowfw in pause mode, step forward 1 frame
DOWN fastbw in pause mode, step back 5 frames
LEFT slowbw in pause mode, step back 1 frame
q display toggle display of frames
s slower slow down
f faster speed up
y toggle toggle displaying only every 5 frames
j grab save a JPEG
r rotate rotate AVI file after next keyframe
(*) all commands must be prefixed with "preview
".
resample - audio resampling filter plugin
resample was written by Thomas Oestreich. The version documented
here is v0.1.4 (2003-08-22). This is a audio filter. It is a pre-
processing only filter.
skip - skip all listed frames
skip was written by Thomas Oestreich. The version documented here
is v0.0.1 (2001-11-27). This is a video and audio filter. It is a
pre-processing only filter.
· fstart1-fend1 [ fstart2-fend2 [ .. ] ] = %s
apply filter [start-end] frames
slowmo - very cheap slow-motion effect
slowmo was written by Tilmann Bitterberg. The version documented
here is v0.3 (2003-29-11). This is a video filter. It can handle
RGB and YUV mode. It is a pre-processing only filter.
This filter produces a simple slow-motion effect by dupli‐
cating certain frames. I have seen this effect on TV and
despite its the simple algorithm it works quite well. The
filter has no options.
smartbob - Motion-adaptive deinterlacing for double-frame-rate output.
smartbob was written by Donald Graft, Tilmann Bitterberg. The ver‐
sion documented here is v1.1beta2 (2003-06-23). This is a video
filter. It can handle RGB and YUV mode. It is a post-processing
only filter.
· motionOnly = %d [default 0]
Show motion areas only
· shiftEven = %d [default 0]
Blend instead of interpolate in motion areas
· threshold = %d [default 12]
Motion Threshold
· denoise = %d [default 1]
Phase shift
This filter only makes sense when fed by -J doublefps. It
will take the field-frames which filter_doublefps produces
and generates full-sized motion adaptive deinterlaced output
at the double import framerate.
smartdeinter - VirtualDub's smart deinterlacer
smartdeinter was written by Donald Graft. The version documented
here is v2.7b (2003-02-01). This is a video filter. It can handle
RGB and YUV mode. It is a pre-processing only filter.
· motionOnly = %d [default 0]
Show motion areas only
· Blend = %d [default 0]
Blend instead of interpolate in motion areas
· threshold = %d [default 15]
Motion Threshold
· scenethreshold = %d [default 100]
Scene Change Threshold
· fieldShift = %d [default 0]
Phase shift
· inswap = %d [default 0]
Field swap before phase shift
· outswap = %d [default 0]
Field swap after phase shift
· noMotion = %d [default 0]
Disable motion processing
· highq = %d [default 0]
Motion map denoising for field-only
· diffmode = %d [default 0]
Motion Detection (0=frame, 1=field, 2=both)
· colordiff = %d [default 1]
Compare color channels instead of luma
· cubic = %d [default 0]
Use cubic for interpolation
This filter provides a smart, motion-based deinterlacing
capability. In static picture areas, interlacing artifacts
do not appear, so data from both fields is used to provide
full detail. In moving areas, deinterlacing is performed
smartyuv - Motion-adaptive deinterlacing
smartyuv was written by Tilmann Bitterberg. The version documented
here is 0.1.4 (2003-10-13). This is a video filter. It can handle
YUV mode only. It is a pre-processing only filter.
· motionOnly = %d [default 0]
Show motion areas only, blacking out static areas
· diffmode = %d [default 0]
Motion Detection (0=frame, 1=field, 2=both)
· threshold = %d [default 14]
Motion Threshold (luma)
· chromathres = %d [default 7]
Motion Threshold (chroma)
· scenethres = %d [default 31]
Threshold for detecting scenechanges
· highq = %d [default 1]
High-Quality processing (motion Map denoising)
· cubic = %d [default 1]
Do cubic interpolation
· Blend = %d [default 1]
Blend the frames for deinterlacing
· doChroma = %d [default 1]
Enable chroma processing (slower but more accurate)
· verbose = %d [default 0]
Verbose mode
This filter is basically a rewrite of the smartdeinter fil‐
ter by Donald Graft (without advanced processing options)
for YUV mode only. Its faster than using the smartdeinter in
YUV mode and is also tuned with its threshold settings for
YUV mode. The filter detects motion and static areas in an
image and only deinterlaces (either by blending or by cubic
interpolation) the moving areas. The result is an image with
high detail in static areas, no information is lost there.
The threshold settings should be sufficent for most users.
As a rule of thumb, I recommend setting the chroma threshold
to about the half of the luma threshold. If you want more
deinterlacing, lower the thresholds. The scene threshold can
be easily found by turning on verbose mode and the preview
filter. In verbose mode, the filter will print out, when it
detects a scene change. If scenechanges go by unnoticed,
lower the scene threshold. You can completly disable chroma
processing with the doChroma=0 option. Here is a sample
commandline
-J smartyuv=highq=1:diffmode=2:cubic=1:Blend=1:chromath‐
res=4:threshold=8:doChroma=1
smooth - (single-frame) smoothing plugin
smooth was written by Chad Page. The version documented here is
v0.2.3 (2003-03-27). This is a video filter. It can handle YUV mode
only. It is a pre-processing only filter. It supports multiple
instances.
· strength = %f [default 0.25]
Blending factor
· cdiff = %d [default 6]
Max difference in chroma values
· ldiff = %d [default 8]
Max difference in luma value
· range = %d [default 4]
Search Range
"single-frame" means it only works with the current frame,
it does not need the next or the previous frame for opera‐
tion. Usually smoothing is done by talking the data of pre‐
vious frames into account to see which parts of the picture
can be "safely" smoothed, this filter only needs one frame.
Usage -J subtitler="[no_objects] [subtitle_file=s] [color_depth=n]
[font_dir=s] [font=n] [font_factor=f [frame_offset=n] [debug]
[help]" f is float, h is hex, n is integer, s is string.
no_objects disables subtitles and other objects (off).
color_depth= 32 or 24 (overrides X auto) (32).
font= 0 or 1, 1 gives strange symbols... (0).
font_dir= place where font.desc is (~/.subtitles/font).
font_factor= .1 to 100 outline characters (10.75).
frame_offset= positive (text later) or negative (earlier)
integer (0).
subtitle_file= pathfilename.ppml location of ppml file
(~/.subtitles/demo.ppml).
debug prints debug messages (off).
help prints this list and exit.
testframe - generate stream of testframes
testframe was written by Thomas Oestreich. The version documented
here is v0.1.3 (2003-09-04). This is a video filter. It can handle
RGB and YUV mode. It is a pre-processing only filter.
· mode = %d [default 0]
Choose the test pattern (0-4 interlaced, 5 colorfull)
text - write text in the image
text was written by Tilmann Bitterberg. The version documented here
is v0.1.4 (2004-02-14). This is a video filter. It can handle RGB
and YUV mode. It is a post-processing only filter.
· range = %u-%u/%d [default 0-4294967295/1]
apply filter to [start-end]/step frames
· string = %s
text to display (no ':') [defaults to `date`]
· font = %s
full path to font file [defaults to arial.ttf]
· points = %d [default 25]
size of font (in points)
· dpi = %d [default 96]
resolution of font (in dpi)
· fade = %d [default 0]
fade in/out (0=off, 1=slow, 10=fast)
· antialias = %d [default 1]
Anti-Alias text (0=off 1=on)
· pos = %dx%d [default 0x0]
Position (0-width x 0-height)
· posdef = %d [default 0]
Position (0=None 1=TopL 2=TopR 3=BotL 4=BotR 5=Cent 6=BotCent)
· notransparent (bool)
disable transparency (enables block box)
see /docs/filter_text.txt
tomsmocomp - Tom's MoComp deinterlacing filter
tomsmocomp was written by Tom Barry et al.. The version documented
here is v0.1 (2004-07-31). This is a video filter. It can handle
YUV and YUV422 mode. It is a pre-processing only filter.
· TopFirst = %d [default 1]
Assume the top field should be displayed first
· SearchEffort = %d [default 15]
CPU time used to find moved pixels
· UseStrangeBob = %d [default 0]
?Unknown?
· CpuFlags = %x [default 5f]
Manual specification of CPU capabilities
unsharp - unsharp mask & gaussian blur
unsharp was written by Remi Guyomarch. The version documented here
is v1.0.1 (2003-10-27). This is a video filter. It can handle YUV
mode only. It is a post-processing only filter.
· amount = %f [default 0.0]
Luma and chroma (un)sharpness amount
· matrix = %dx%d [default 0x0]
Luma and chroma search matrix size
· luma = %f [default 0.0]
Luma (un)sharpness amount
· chroma = %f [default 0.0]
Chroma (un)sharpness amount
· luma_matrix = %dx%d [default 0x0]
Luma search matrix size
· chroma_matrix = %dx%d [default 0x0]
Chroma search matrix size
· pre = %d [default 0]
run as a pre filter
This filter blurs or sharpens an image depending on the sign
of "amount". You can either set amount for both luma and
chroma or you can set it individually (recommended). A posi‐
tive value for amount will sharpen the image, a negative
value will blur it. A sane range for amount is -1.5 to 1.5.
The matrix sizes must be odd and define the range/strength
of the effect. Sensible ranges are 3x3 to 7x7.
It sometimes makes sense to sharpen the sharpen the luma and
to blur the chroma. Sample string is:
luma=0.8:luma_matrix=7x5:chroma=-0.2:chroma_matrix=3x3
whitebalance - White Balance Filter - correct images with a broken
white balance
whitebalance was written by Guillaume Cottenceau. The version docu‐
mented here is v0.1 (2003-10-01). This is a video filter. It can
handle RGB and YUV mode. It is a pre-processing only filter.
· level = %d [default 40]
Level of blue-to-yellow white balance shifting (can be negative)
· limit = %s
Limit to specified ranges (+fnumber toggles on, -fnumber toggles
off)
This filter allows correcting movies with a broken white
balance, e.g. bluish movies.
xharpen - VirtualDub's XSharpen Filter
xharpen was written by Donald Graft, Tilmann Bitterberg. The ver‐
sion documented here is (1.0b2) (2003-02-12). This is a video fil‐
ter. It can handle RGB and YUV mode. It is a post-processing only
filter.
· strength = %d [default 200]
How much of the effect
· threshold = %d [default 255]
How close a pixel must be to the brightest or dimmest pixel to
be mapped
This filter performs a subtle but useful sharpening effect.
The result is a sharpening effect that not only avoids
amplifying noise, but also tends to reduce it. A welcome
side effect is that files processed with this filter tend to
compress to smaller files.
Strength 'strength' (0-255) [200]
When this value is 255, mapped pixels are not blended
with the original pixel values, so a full-strength effect is
obtained. As the value is reduced, each mapped pixel is
blended with more of the original pixel. At a value of 0,
the original pixels are passed through and there is no
sharpening effect.
Threshold 'threshold' (0-255) [255]
This value determines how close a pixel must be to the
brightest or dimmest pixel to be mapped. If a pixel is more
than threshold away from the brightest or dimmest pixel, it
is not mapped. Thus, as the threshold is reduced, pixels in
the mid range start to be spared.
yuvdenoise - mjpegs YUV denoiser
yuvdenoise was written by Stefan Fendt, Tilmann Bitterberg. The
version documented here is v0.2.1 (2003-11-26). This is a video
filter. It can handle YUV mode only. It can be used as a pre-pro‐
cessing or as a post-processing filter.
· radius = %d [default 8]
Search radius
· threshold = %d [default 5]
Denoiser threshold
· pp_threshold = %d [default 4]
Pass II threshold
· delay = %d [default 3]
Average 'n' frames for a time-lowpassed pixel
· postprocess = %d [default 1]
Filter internal postprocessing
· luma_contrast = %d [default 100]
Luminance contrast in percent
· chroma_contrast = %d [default 100]
Chrominance contrast in percent.
· sharpen = %d [default 125]
Sharpness in percent
· deinterlace = %d [default 0]
Force deinterlacing
· mode = %d [default 0]
[0]: Progressive [1]: Interlaced [2]: Fast
· scene_thres = %d%% [default 50]
Blocks where motion estimation should fail before scenechange
· block_thres = %d [default 1024]
Every SAD value greater than this will be considered bad
· do_reset = %d [default 2]
Reset the filter for `n' frames after a scene
· increment_cr = %d [default 2]
Increment Cr with constant
· increment_cb = %d [default 2]
Increment Cb with constant
· border = %dx%d-%dx%d [default 0x0-32x32]
Active image area
· pre = %d [default 0]
run this filter as a pre-processing filter
see /docs/filter_yuvdenoise.txt
yuvmedian - mjpegs YUV median filter
yuvmedian was written by Mike Bernson, Tilmann Bitterberg. The ver‐
sion documented here is v0.1.0 (2003-01-24). This is a video fil‐
ter. It can handle YUV mode only. It can be used as a pre-process‐
ing or as a post-processing filter.
· radius_luma = %d [default 2]
Radius for median (luma)
· radius_chroma = %d [default 2]
Radius for median (chroma)
· threshold_luma = %d [default 2]
Trigger threshold (luma)
· threshold_chroma = %d [default 2]
Trigger threshold (chroma)
· interlace = %d [default 0]
Treat input as interlaced
· pre = %d [default 1]
Run as a PRE filter
AUTHORS
transcode was written by Thomas Oestreich
<ostreich@theorie.physik.uni-goettingen.de> with contributions from
many others. See AUTHORS for details.
SEE ALSOtranscode(1), transcode_import(1), transcode_export(1), tcmodinfo(1),
1 modules transcode(filter)
