1. Could you explain why, when the factor in the mix color node is changing, it doesn't look how I would expect, which is that there are two different images being cross faded?
2. This would work if we mix the W value of a 4D noise texture, instead of the z-location on the mapping node, right?
3. I accidentally left out putting the linear interpolation in the graph editor for both mapping nodes, but it still seems to loop perfectly. Why do we need it? Is it just in case we want to change the duration of the timeline?
I can answer number three. Yes, it's just in case you want to shift things around or extend the duration of the animation, it'll just interpolate forward and you wont have to worry about tweaking that further.
In question one, I think that is exactly what is happening, the images are being cross faded but since it's just an amorphous blob the transition gets masked nicely. Kent inserted the handy graph to get an idea of what's happening:
Question two, I think the answer is yes, but I'm not too sure, that would have to be tested, but it sounds legit.
1. It's a good question because..I'm having trouble explaining exactly why. It makes more sense in the second loop project where we're mixing the noise patterns BEFORE they influence the voronoi's vector coordinates. It's like we're burying the transition in an undetectable place.
But the others, yeah, you'd think they would result like a simple cross fade. My thought aligns with Omar's. Since the mix inputs are so similar, perhaps the crossfade is less perceptible. Surely it would be a noticeable crossfade if the inputs were totally unrelated.
2. I believe so! I seem to remember 
spikeyxxx informing me about this at some point. Silly me just isn't used to the W coordinates. I'm an old dog slow to learn new tricks 👨🦳
3. Yes, as Omar affirmed, linearly EXTRAPOLATED curves are best for timing adjustment.
Hi Benjamin,
As to question 1, it is indeed the fact that two similar patterns are being cross-faded, not an image of a monkey with an image of a lion.
I have a side note about the Vector Mixing not working, but I'll put it in a seperate post under this lesson.
As to question 2, yes, you are absolutely right, a W going from 0 to 1, for instance, would even have a similar pattern changing speed as a Z-Location going from 0m to 1m. (Wouldn't be the same pattern though.)
As to question number 3, it will still loop perfectly, but the pattern changing speed would not be constant...you might not notice that, but Wayne Dixon certainly will!