wiki:superBatchTransform

Version 1 (modified by bpt, 3 years ago) ( diff )

Creating page about super batch transforms

"Super Batch Transform" allows specifying transforms in a single file with a good degree of flexibility. Each row defines most of the parameters of a transform, including output of several frames per row. The columns defined in the file should be:

Subject (String filename without extension) Source (String filename without extension) Target (String filename without extension) Start value (floating point number) End value (floating point number) Output (String output file name) Steps (integer number) Shape (integer - if 1 include shape in the transform, otherwise don't) Colour (integer - if 1 include colour in the transform, otherwise don't) Texture (integer - if 1 include texture in the transform, otherwise don't) Mask (integer - if 1 mask the transform, otherwise don't) Sample (integer - if 1 include sampling in the transform, this is about using points along curves not just the anchor points in the template) Resize (floating point value - for overall scaling of the output, so 1 is no rescaling)

Note that some of the global settings, such as Average -> shape normalisation and Average->warping option are used and will have an impact on the output. Things can also get very slow with certain options (e.g. Thin plate spline warping with sampling of the points can get very slow!)

An example file with two rows is given below:

CFD-WF-006-002-N CFD-WF-006-002-N_MASK aria_grande_average_m 0 1 10 test 1 1 0 0 0 1 CFD-WF-006-002-N CFD-WF-006-002-N_MASK aria_grande_average_m 0 1 10 resize 1 1 1 1 0 0.5

Both rows apply the transform to the same sequence producing 10 steps from 0 to 1. The first row outputs files test0.jpg to test9.jpg and the second row outputs resize0.jpg to resize9.jpg and the corresponding .tem template files. The first row applies the transform with shape and colour, but no texture, whereas the second uses shape, colour and texture. The first row does not mask the transform, but the second does, and neither use point sampling. The first row is applied to the full size image, and the second row shrinks the image by 50% in width and height.

Note: See TracWiki for help on using the wiki.