Stuka - music video creator for studio recording sessions
v1.0.9 by arnaud sintès & nicolas combe

[home] [using Stuka] [scripting 1/2] [scripting 2/2] [download]


scripting Stuka (1/2)
overlay( 8mm, 100 ) {
    capture( 17, 26, 692, 448 );
}

marker( 4, 1top_3bottom ) {
    display( "gh1" );
    display( "et1" );
    display( "et2" );
..

The main part of Stuka is a text file defined by the user to set up all the building operations.

Basically, the script contains different sections to describe:
  • the output video file information
  • a set of input video files information and adjustments
  • a set of group definition
  • an optional overlay description
  • a set of markers to set up the timeline and the display layout of videos

    First of all, you have to set up the output video file configuration, using the following command:
    output( "output.avi", width, height, interpolation_type, "master_track.wav" );

  • "output.avi" is the full path to the output AVI file
  • width / height are the output video file dimensions in pixels
  • interpolation_type [optional] is the name of the interpolation filter
  • "master_track.wav" [optional] is the full path to an input wave track to insert into the AVI

    Interpolation type determines the filter used to rescale input video frames.
    Values can be: "fast_bilinear" (default), "bilinear", "bicubic", "x", "point", "area", "bicublin", "gauss", "sinc", "lanczos" or "spline"

    It may be usefull to set it to "fast_bilinear" when working with previews, and to switch it to "bicubic" when dealing with final rendering, as long as "fast_bilinear" is faster but less accurate.

    On the same topic, don't forget that a 640x360 rendering will be faster than a 1280x720 one.

    The input wave track to insert may be a Wave file, PCM signed 16 bit, stereo at 44.1KHz.


    Then, you have to consider a bunch of video files you've made during the recording sessions of a track:
    input( iid, "input.mov", focus ) {}

  • iid is a unique ID relative to the particular input video file, e.g. "arnaud_front"
  • "input.mov" is the full path to the input video file (MOV, MPEG, AVI etc.)
  • focus [optional] determines a point of interest in the video image

    Focus type can be: "center" (default), "top", "left", "bottom", "right", "top_left", "top_right", "bottom_left" or "bottom_right"

    The focus is particularly important when you are using small viewport layouts (with five or six displays), because some cropping can occurs.
    The cropping area will always be computed to display the focused part of your video.


    All these videos are relative in time to the master track you have under your audio sequencer (ProTool, Cubase, Sonar or whatever you prefer), so it is really important to indicates how to anchors them in the timeline.

    To do that, I personally use Virtual Dub to extract the audio part of an input video file (open the video file and use the "File/Save WAV" menu), then I use Goldwave to search a point in the input video file WAV (a distinct beat), to note the point position in millisecond and then search the equivalent point in the master WAV track (it's easy to find it directly onto your sequencer) and note its position.

    By these operations, you are able to anchors a point in input video file relatives to master track, using:
    input( iid, "input.mov", focus, anchorMs, TLAnchorMs ) {}

  • anchorMs [optional] determines a point position in millisecond in the input video file
  • TLAnchorMs [optional] determines a point position in millisecond in the master track (timeline)

    by default, both anchors are set to 0 millisecond.


    As long as you got a lot of different input video sources, it may be important to adjust their images:
    input( iid, "input.mov" ) {
        color_balance( CR, MG, YB );
        brightness( b );
        contrast( c );
        saturation( s );
        noise_reduction( level );
        vertical_flip( flip );
    }

  • color_balance( CR, MG, YB ); [optional] modify the balance between Cyan/Red, Magenta/Green and Yellow/Blue, all values are set between -1.0 to 1.0 where 0.0 is the default color balance.
  • brightness( b ); [optional] determines the brightness ('b' equals -1.0 to 1.0 where 0.0 is the default)
  • contrast( c ); [optional] determines the contrast ('c' equals -1.0 to 1.0 where 0.0 is the default)
  • saturation( s ); [optional] determines the saturation ('s' equals 0.0 to 2.0 where 1.0 is the default)

  • noise_reduction( level ); [optional] apply a smart algorithm over the image to reduce
    the noise amount (temporal), 'level' is in percent (0: no NR (default), 100: full NR, try 6% for happiness !)

  • vertical_flip( flip ); [optional] indicates how the decoder must flip the input image
    'flip' values can be: "auto" (autodetection, by default), "true" or "false"

    The best way to adds a new input with correct image adjustment settings is to use the adjustment panel.

    To help with future markers creation and cycle layout, we can determine groups:
    group( gid ) { iid, iid, gid, ... }

  • gid is the unique ID of the group, e.g. "arnaud_all"
  • iid is the id of an input, previously defined with an "input() {}" statement

    Note than you can create a group including some input AND some previously defined groups

    When working with preview, it's really usefull to bound the range of markers to compute:
    range( begin_pos, end_pos );

  • begin_pos is the position of the first marker to constraint the build range
  • end_pos is the position of the last marker to constraint the build range


    On the same topic, it may be really usefull to display some debug information:
    debug();

    this command adds frame count, input marker position, signature and tempo, interpolated marker position / beat count, and viewport input iid and gid.