Assumptions about an <animation>'s <sampler> INP

I am wondering if it is ok to assume that the <input> with an INPUT semantic of an <animation>'s <sampler> is a time in seconds IE: the INPUT semantic is ALWAYS a time in seconds.

I ask because a while ago I noticed that DAZ Studio exports a mesh UV values with “U” and “V” as parameter names, even though the spec states they should be named “S” and “T”, I posted on their forums about the “bug”, but I was told parameter names are meaningless and semantics is all that matters, thus, the TEXCOORD semantic meant the parameters where in fact UV’s.

In the case of an animation sampler the “INPUT” semantic is rather vague (unlike “TEXCOORD”) as it could be given as a TIME or as a FRAME, there is nothing other than the parameter name inside the <source>'s <technique_common>'s <accessor> that differentiates one from the other, I have only seen “TIME” so far, but I am wondering if “FRAME” is a possibility or if there are other possibilities out there.

So any Ideas?


There is no notion of ‘frame’ in COLLADA.
The specification says all times are in seconds.

Alright, so as far as version 1.4.1 goes the INPUT semantic on that context is always TIME (in seconds) right?

Just to be 100% clear :smiley:


The specification does not limit INPUT to be TIME. But I have not seen other semantics implemented by exporters at this time.
TIME is always in seconds.

Alright, thanks :smiley: