Monday, Dec 20, 2004
Integrating new features to visual plasma.
Visual Plasma v0.4
snapshot:
New feature
Video matting
Editing video volume/alpha
Adding titles
Editing title duration/alpha
Preview clip/result
Results
- snapshot:
- demo video:
result.avi
Visual Plasma v0.4
snapshot:
New feature
Results
- snapshot:
- demo video:
result.avi
Saturday, Dec 18, 2004
Developing title feature.
Visual Plasma
Interface
To add a title, user can select a animation effect from
the title tab (currently only still text was implemented)
User can edit (text, position, scale) the title in the edit
panel just like MS Powerpiont
User can modify the alpha (transparentcy) and duration by
draging sliders.
snapshot:
Implementation
Title object data structure
- :
- :
Draw a editable text in the edit panel
- :
- :
Bindings for resize and move is similar to video clip's
- resize:
- move:
Database update is similar as before
Problem
During input text, the text can not automaticly warp to newline
when the text reach the bound of box.
Since scale tast is perform by zinc command, the text became ugly
after scaling.
Visual Plasma
Interface
snapshot:
Implementation
- :
- :
- :
- :
- resize:
- move:
Problem
Thursday, Dec 16, 2004
Developing crop feature
Visual Plasma
Crop Window
When user select a video in the edit panel, it will also appear in the crop window. Usr can draw a box on the preview picture, and the part outside the box will be crop off. Then the edit panel will display the new croped video clip but the picture in crop window will remain the same (full) so that user can draw new crop box
snapshot:
Implementation
Initail Zinc
-
-
Update dabtabase
-
- it will call
Cropping image
-
- Source code:
Problem
Crop method is a very slow, it will cost around 1 second to
produce one image file.
The reason should be it was written in Tcl.I have try to use
Dalí library to perform this task, however, I haven't find the
right way. The following code is my attempt, it create a mask and
copy bytes to the output.
- source code:
crop.tcl
- result:
Very fast, less than 0.1 second to produce same result, it works fine if the crop point is {0 0 X X}, but works anomally when left-top corner point is not (0, 0). For example:
Visual Plasma
Crop Window
When user select a video in the edit panel, it will also appear in the crop window. Usr can draw a box on the preview picture, and the part outside the box will be crop off. Then the edit panel will display the new croped video clip but the picture in crop window will remain the same (full) so that user can draw new crop box
snapshot:
Implementation
-
prop_panel_init will create a black zinc.-
Button-1, B1-Motion bindings allow to draw box on zinc-
db_update_crop will update database after user
draw a box- it will call
edit_panel_crop to produce a
cropped image and update edit panel.-
edit_panel_crop will read the PPM file byte by byte
and only write the pixel byte which is inside crop box to the output
file.- Source code:
set f [open $filename r]
set fout [open "tmp_crop_$img_id.ppm" w]
set c ""
set hdr ""
while { $c != "\n" } {
set c [read $f 1]
set hdr "$hdr$c"
}
set hdr_list [split $hdr]
set 3w [expr 3 * [lindex $hdr_list 1]]
set h [lindex $hdr_list 2]
set 3x1 [expr 3 * $x1]
set 3x2 [expr 3 * $x2]
set w2 [expr $x2 - $x1]
set h2 [expr $y2 - $y1]
fconfigure $f -translation binary -buffersize 65536
fconfigure $fout -translation binary -buffersize 65536
puts -nonewline $fout "P6 $w2 $h2 255\n"
for {set i 0} {$i < $h} {incr i} {
for {set j 0} {$j < $3w} {incr j} {
if {$i >= $y1 && $i < $y2 && $j >= $3x1 && $j < $3x2} {
puts -nonewline $fout [read $f 1]
} else {
seek $f 1 current
}
}
}
close $f
close $fout
Problem
- source code:
crop.tcl- result:
Very fast, less than 0.1 second to produce same result, it works fine if the crop point is {0 0 X X}, but works anomally when left-top corner point is not (0, 0). For example:
Tuesday, Dec 14, 2004
TkZinc: Zinc Is Not Canvas!
TkZinc is a powerful widget for Tcl/Tk, Perl/Tk and Python/Tkinter. It is very similar to Tk canvas in that they support structured graphics. Like the canvas, TkZinc implements items used to display graphical entities. Those items can be manipulated and bindings can be associated with them to implement interaction behaviors. But unlike the canvas, TkZinc can structure the items in a hierarchy, has support for scaling and rotation, clipping can be set for sub-trees of the item hierarchy, supports muti-contour curves.
snapshot:
Feature
- arc, curve, group, icon, map, rectangle, reticle, tabular, text, window ...
- support more options:
aplpha, linestyle, linepattern, fillpattern- rotate, scale, skew, transform ...
- color gradient
- antialiasing
- transparency
need the text input support (zincText.tcl), it apply to those sensitive text item with tag "text". example:
package require Tkzinc
package require zincText
zinc .z
pack .z
# Activate text input support from zincText
zn_TextBings .z
.z add add text 1 -position {50 230} -text \
{this text is editable} -sensitive 1 -tags text
Monday, Dec 13, 2004
Embedding new version of plamsa to visual plasma.
Visual Plasma v0.3
Change log
When user click a image in canvas, it will be put into
the forground.
-
The cursor will change when user enter a image or box
-
User can put image in any position
- a new varible called layer number is stored in clip object. When user move or resize an image, its layer number will update to be the
- when making movie, a new algorithm will parse the position database, and generate plasma command.
step 1: sort database as before;
step 2: find width and height of result video, and create the background layer with a black JPG image;
step 3: sort the database with the layer number;
step 4: for each clip in the database, create a new video obejct , and modify its scale and position:
Problems
The new encoder seems slower than before, and it cannot
stop
Since the old packing algorithm has been discard, need
a new method for auto-alignment
Visual Plasma v0.3
Change log
-
$canvas raise $img_id-
$canvas configure -cursor top_left_corner- a new varible called layer number is stored in clip object. When user move or resize an image, its layer number will update to be the
max_layer_number + 1.- when making movie, a new algorithm will parse the position database, and generate plasma command.
step 1: sort database as before;
step 2: find width and height of result video, and create the background layer with a black JPG image;
step 3: sort the database with the layer number;
step 4: for each clip in the database, create a new video obejct , and modify its scale and position:
set i 0
foreach item $list {
media .m.id$i [lindex $item 5]
mpack .m.id$i
set item_w [expr [lindex $item 3] - [lindex $item 1]]
set item_h [expr [lindex $item 4] - [lindex $item 2]]
set item_x [expr [lindex $item 1] - [lindex $list 0 1]]
set item_y [expr [lindex $item 2] - [lindex $list 0 2]]
.m.id$i scale $item_w $item_h
.m.id$i pos $item_x $item_y
incr i
}
- result:Problems
Friday, Dec 10, 2004
Vee Jay Online
VJOnline is digital audio/video stream mixer. User can multiplex a few A/V stream into one and broadcast over internet via RTSP(Real Time Streaming Protocol). It supports serveral transition effects when user switch the main channel and picture-in-picture effect. User can also add in text maquees and other text effects such as blinking.
Feature
Transition effect
during playing, user can switch channel with transition effect. -
-
Picture-in-Picture effect
Text rendering
- top/bottom maquee effects are implemented using keyframe pos motion tween.
- blink effect use alpha motion tween.
Source
- source code:
vj.tcl
- demo video:
vj.avi
VJOnline is digital audio/video stream mixer. User can multiplex a few A/V stream into one and broadcast over internet via RTSP(Real Time Streaming Protocol). It supports serveral transition effects when user switch the main channel and picture-in-picture effect. User can also add in text maquees and other text effects such as blinking.
Feature
during playing, user can switch channel with transition effect. -
swich_channel will mute the old channel,
create and pack the new channel-
fade_in, pop_in, slide_in and curtain_in is procdure
used to make spectial effects. They are implemented using keyframe
alpha/scale/crop motion tween. remove_old_channel will
remove all the sub layers.
add_layer will add a new layer(176X120) to the old
channel, user can specify the anchor point.
- top/bottom maquee effects are implemented using keyframe pos motion tween.
- blink effect use alpha motion tween.
Source
- source code:
vj.tcl- demo video:
vj.aviThursday, Dec 09, 2004
Zhu Tao release new version of Plasma!
Plasma Beta v3.0
This version is released as a dynamic loadable module (tclplasma.so) for Tcl 8.4.8. instead of a executable file as before. Some syntax changed, and some new functions were added. Zhu Tao also implemante a simple Tk application called VeeJay Online, which can make some interesting effects such as trasition, PiP, and title.
Change Log
New syntax
- create a video or audio object:
- output A/V:
- for pack(
- event, what is key?:
- keyframes can be used to make animation:
- apply the effect?:
- new option for media:
New features
- support format Mpeg2,4, AVI, WMV, MP3, JPG
- achive PiP using layer concept, user can specify the position of foreground layer and one image can over other images.
example:
- titles can be added as a layer, user can give width and height of one charactor and modify the position. Here is a example:
- A/V streams motion tween (alpha, size and position tweening, volume fading). Example:
Problems
How to concatenate videos
when packing a child object to parent, if parent is a container, it will be packed side-by-side, else, if the parent is a video, the child will be packed as a new layer witch produce picture-in-picture effect.
How to continue encoding
when one layer reach its end, the encoder will stop its work, which make the duration to be the shorter than expected.
How to terminate encoding process
after finish encoding, the program get stuck, do not konw how to terminate in other way than Ctrl+C :(
How to use event
the syntax has changed, waiting for Zhu Tao's documentation.
Plasma Beta v3.0
This version is released as a dynamic loadable module (tclplasma.so) for Tcl 8.4.8. instead of a executable file as before. Some syntax changed, and some new functions were added. Zhu Tao also implemante a simple Tk application called VeeJay Online, which can make some interesting effects such as trasition, PiP, and title.
Change Log
- create a video or audio object:
media $object_name $file_name(path | url | dev)- output A/V:
mplay/mout $object_anme $destionations //dev=playback- for pack(
mpack), when parent is not empty,
-side whole is no need.- event, what is key?:
mevent $object_name [-timer $num_of_frames] [-key $key]- keyframes can be used to make animation:
mkeyframe $object_name [-motiontween] $frame_number- apply the effect?:
mset $object_name- new option for media:
mute, grayscale, volume, pos $x $y, alpha $a,
crop $x1 $y1 $x2 $y2, script{$s}, text "" $x &y- support format Mpeg2,4, AVI, WMV, MP3, JPG
- achive PiP using layer concept, user can specify the position of foreground layer and one image can over other images.
example:
- titles can be added as a layer, user can give width and height of one charactor and modify the position. Here is a example:
media .m ./susie.mpv media .m.t text "Hello World!" 40 30 mpack .m.t .m.t pos 0 0 mout .m test.aviNote that user need to set environment variable PLASMA_FONT_FILE to a ttf font file
- A/V streams motion tween (alpha, size and position tweening, volume fading). Example:
media .m susie.mpv media .m.t1 cactus.mpv mpack .m.t1 .m.t1 scale 120 90 .m.t1 pos 10 10 .m.t1 alpha 0 mkeyframe .m.t1.kf -motiontween 50 .m.t1.kf alpha 100 mset .m.t1.kf mout .m test.mpg
Problems
when packing a child object to parent, if parent is a container, it will be packed side-by-side, else, if the parent is a video, the child will be packed as a new layer witch produce picture-in-picture effect.
when one layer reach its end, the encoder will stop its work, which make the duration to be the shorter than expected.
after finish encoding, the program get stuck, do not konw how to terminate in other way than Ctrl+C :(
the syntax has changed, waiting for Zhu Tao's documentation.
Wednesday, Dec 08, 2004
Convert interm progress report to tex format.
LaTeX - A document preparation system
LaTeX is a high-quality typesetting system, with features designed for the production of technical and scientific documentation. LaTeX is the de facto standard for the communication and publication of scientific documents. Here is a good tutorial website Getting to grips with Latex
Commands
Document Structure
-
-
-
-
-
-
-
Importing Images
- support JPEG and PNG formats
- need to put
- options:
- usage:
Floats, Figures and Captions
- need to put
- float example:
- figure example:
Formatting
- Dalí :
- bold :
- footnote :
Interim Progress Report
LaTex source
- directory: www.comp.nus.edu.sg/~liubing/UROP/code/latex/
Compiled with socreport class
pdf(1.0M)
Problem
do not know how to add url in the bibliography.
LaTeX - A document preparation system
LaTeX is a high-quality typesetting system, with features designed for the production of technical and scientific documentation. LaTeX is the de facto standard for the communication and publication of scientific documents. Here is a good tutorial website Getting to grips with Latex
Commands
-
\part{part}-
\chapter{chapter}-
\section{section}-
\subsection{subsection}-
\subsubsection{subsubsection}-
\paragraph{paragraph}-
\subparagraph{subparagraph}- support JPEG and PNG formats
- need to put
\usepackage{graphicx} in the header- options:
[width=x, height=x, scale=x, trim l b r t, clip]- usage:
\includegraphics[options]{x.jpg}- need to put
\usepackage{float} in the header- float example:
\usepackage{float}
\floatstyle{ruled}
\newfloat{program}{thp}{lop}
\floatname{program}{Program}
\begin{document}
\begin{program}
\begin{verbatim}
class HelloWorld {
public static void main(String[] args) {
System.out.println("Hello World!");
}
}
\end{verbatim}
\caption{The Hello World! program in Java.}
\end{program}
- cross-refrence:
\ref{xxx} and \label{xxx}- figure example:
\begin{figure}
\begin{center}
\reflectbox{\includegraphics{tucan.eps}}
\end{center}
\caption{A picture of the same tucan looking the other way!}
\label{toucan}
\end{figure}
- Dalí :
Dal\'{i}- bold :
\textbf{xxx}- footnote :
\footnote{xxx}Interim Progress Report
- directory: www.comp.nus.edu.sg/~liubing/UROP/code/latex/
pdf(1.0M)do not know how to add url in the bibliography.
url="www.apple.com" can not be handled by socreport.bst,
so use the misc instead:
@misc{imovie,
author = "Apple",
title = "i{M}ovie, http://www.apple.com",
year = 2004
};
However, it makes the reference page looks ugly :(Tuesday, Dec 07, 2004
Meeting Notes
Need to add new functions
- add titles: new version of plasma has this feature, ask zhutao how to use and CVS the new one
- overlay: should check Tk menu again about how to put image on the foreground in canvas
- powerpoint is a good example, check it out
Temporal editing
- timeline is hard to use, storyboard can be considerd
- should also think about other visalazation to allow easy-to-use temporal editing
Final report
- start to write report now
- should write in LaTex format
- LaTex class can be download here
- add titles: new version of plasma has this feature, ask zhutao how to use and CVS the new one
- overlay: should check Tk menu again about how to put image on the foreground in canvas
- powerpoint is a good example, check it out
- timeline is hard to use, storyboard can be considerd
- should also think about other visalazation to allow easy-to-use temporal editing
- start to write report now
- should write in LaTex format
- LaTex class can be download here