Tranny NPM | npm.io

tranny

A little library for converting compound CSS transforms into their matrix equivalents.

usage

# using the default matrix
new Tranny

# or using a custom 4x4 matrix (meaning some transformations are already applied)
matrix = [
	[1, 2, 3, 4]
	[5, 6, 7, 8]
	[9, 0, 1, 2]
	[3, 4, 5, 6]
]
new Tranny matrix

example

tranny = new Tranny

tranny

# set some transforms
.rotate '90deg'
.translate3d 50, 100, 200

# get a matrix back
tranny.getMatrix()


#=> [
#		[9.870993963020204, 0.7, 0, 0],
#		[-0.5, 0, 0, 0],
#		[0, 0, 1, 0],
#		[443.54969815101026, 35, 200, 1]
#	]

# .. or as a CSS property
tranny.getMatrixCSS()

#=> "matrix3d(9.870993963020204, 0.7, 0, 0, -0.5, 0, 0, 0, 0, 0, 1, 0, 443.54969815101026, 35, 200, 1)"

# set a new base matrix (on the basis of which transforms are applied)
tranny.matrix [
	[3, 1, 4, 1]
	[5, 9, 2, 6]
	[5, 3, 5, 8]
	[9, 7, 9, 3]
]

supported transforms

perspective
rotate
rotateX
rotateY
rotateZ
rotate3d
scale
scaleX
scaleY
scaleZ
scale3d
skew
skewX
skewY
translate
translateX
translateY
translateZ
translate3d

why not use actual matricies?

If you're performing caculations dozens of times per second (in the case of animations. 60 times per second), performance is essential. Not only do we need to be able to perform calculations for complex compound 3D transforms at 60FPS, but we need to leave head room for any precursor computations, as well as compositing and painting (ideally, the last two should take place on the GPU). Performing the calculations ourselves (instead of leaving it to the CSS engine) gives us finer control over where (CPU vs. GPU, this thread vs. a WebWorker, sync vs. async) and when (precomputed, partially applied, realtime) computation occurs.