Today, let's recreate the LOKI intro sequence. I firmly believe that the best way to master
fusion is to take reasonably complex projects, try to break them down and recreate it, so
that it takes you out of your comfort zone. It'll make you think of ways to solve problems
that you may not be facing in your day-to-day projects. Now, this is the original LOKI intro sequence. This sequence is exactly 400 frames in length. And if you scrub through from the beginning,
you'll notice that there is a font change that happens every 10th frame. And this font change occurs until frame number
320, at which point of time the master fonts, the logo fonts are displayed once. There is an explosion of light, as you can
see here. And then these master fonts stay for the remainder
of the animation, but there is a minor camera shake at every letter. So this is kind of the breakdown of what's
going on. And now let's try to reproduce it. The way I'm going to approach this is I'm
going to do two fusion compositions. One composition for where the font changes
every 10th frame. So that is from frame zero to frame number
320. And I'm going to use another fusion composition
between frame 320 to frame 400 for the last 80 frames. So my timeline is 24 frames per second. In the edit page, let's create a new fusion
composition and bring it in. And here, I'm only going to take it for 32
frames. It's going to be a relatively short composition
for 32 frames, and I'll explain why later. Let's bring in four text nodes, one for each
character. So this is going to be my L. This is going to be my O. This is going to be my K. And this is going
to be my I. I'm just going to use these as my base characters, the idea here being that
if you want to change the character at any time in the future, you can do it. We might not be doing it in this particular
animation, but it's a good practice to have. Next, let's create the Ls. Now, I told you that there are 32 font changes
because there are 320 frames, and there's a font change every 10 frames, so 32 font
changes. You could create 32 versions of the Ls and
32 versions of the Os, but I'm just going to do it six times, and I'm going to reuse
the same fonts across the animation. But if you feel like it, go and do it 32 times. This is my first node for the L. What I'm
going to do is I'm going to pin this. Go back to text five so that I can have the
pin down here. Right click, expression, take the whip, and
connect it. So I'm looking for text one style text. What I've just done is I've connected text
five to text one. So if I bring text one to the first monitor
and text five to the second monitor, you'll see that they both have L in it. If I go back to the text one and I change
it to something else, text five also changes. So that's what the whip does. So let's go back to L. So text five is now
connected to text one. I would also like to go here and then start
picking a random font. That looks nice. Now we need to duplicate these font changes
five more times because I want six fonts. So copy text five, and instead of doing a
paste, do a Control-Shift-V. This will give you an instance node. What an instance node does is that it'll retain
the values of text five. So in the future, if you want to make a change
to text five, the same changes. The areas highlighted in green here, those
values will automatically carry over to any instance node. So in instance text five, I want all values
of text five to carry over except for the font value. I want a different font. So to do that, right click where the font
is and choose the instance font group. Once you've done that, every parameter except
the font group is now tied to text five, but instance text five has its own font choice. So now choose instance text five and change
the font. And now simply redo exactly that four more
times. I've duplicated this four more times, and
I've created a different font for each of these nodes. So those are your six versions of your Ls. Now the next thing to do is we don't need
all six fonts to be present on the output at the same time, right? We need to choose between one of those fonts
every frame. Now, every time you see a node structure like
this, the first thought to combine these nodes would be to bring the merge node in. So something like, say like this. The issue with the merge node is that the
merge node is used to layer one node's output on top of another node's output. So if I bring merge one into monitor two,
you'll see that both the Ls are present at the same time, and that's not quite what we
want. We want either the L from text five or instance
text five to be present at any given time. So in order to do that, the node you would
use is not the merge node. The node you would rather use is called the
dissolve node. So let me bring in the dissolve node. Now what the dissolve node would do is depending
on the background foreground value here on the right, if the value is one, it'll pick
the foreground node. And if I were to move the slider all the way
to zero, you'll see that the other L is now displayed. So what that tells us is if the value of the
dissolve node is zero, then it picks the input coming from the background. And if the value is one, it picks the input
that's coming from the foreground, the green input here. Now this is kind of what we want, but the
challenge with the dissolve node that we have is that if there's an intermediate value,
you'll see both the inputs being displayed and they dissolve from one input to the other. Now this is quite not what we want. We want either a zero or a one. We do not want to handle intermediate values. So to solve that, what we can do is we can
use expressions. If this video has been helpful in any way,
please hit the like button and leave a comment and tell me how you enjoyed it. If it has not been helpful, please also tell
me how I can improve the next time. So right click on the background foreground
slider, choose expression and type this expression here. If mat.random is less than 0.5, then zero
else one. What this piece of code is doing is you're
asking resolve to generate a random number between zero and one. And if that random number is less than 0.5,
you're asking it to make that random number zero. And if the value is above 0.5, you're asking
it to make that value one, which means that at any point of time, the value is going to
be either zero or one. So if I scrub through, you'll see that the
value here becomes zero or one at any given time. So only one of the two letters is chosen at
any given point of time, which is exactly what we want. This dissolve node is able to choose between
text five and instance text five, any one of the inputs. We need to do the same thing for the remaining. So let's copy this dissolve node, do an instance
paste and connect these two, another instance paste and connect these two. And we need to keep going, another instance
paste and connect these two. The idea here is that at the end of it, we
just want one L to be chosen. Now with this structure, if I were to bring
this instance nodes output into monitor two, you'll see that as I scrub, depending on the
random values generated by these five dissolve nodes, a random L is chosen every frame. Now what we wanna do is we wanna duplicate
this. And now I want to change this to be my O's. So I'm gonna choose this and instead of text
one dot style text, I'm looking at text two because text two is my O. So I'm gonna change
this to text two dot style text. That makes it the O you can see here. I'm gonna bring this into monitor one. And I'm just gonna replace every instance
as well with an O. And now if I bring this instance dissolve into monitor two, you'll
see that a random O is generated every frame. Do the same thing for your Ks and your Is. Now once you've repeated it for your L's,
O's, Ks and Is, just go ahead and group them for easy management. So that's your, I'm gonna rename this. So that's your L group them using control
G and then rename them using F2. So that's your O cool. So these are your four characters that are
randomly generated every frame. We need to combine these four and I'll use
a multi merge node to combine them. Node one is now incorrectly connected to the
background. I actually need a background node as well. So let me bring in a background node. I'll connect the background node to the background
input, the yellow one and I'll connect the L to the white input. Let me bring in the original Loki animation
for reference. Let's put it in media one here, bring this
into media one. And I just want to get the background color
correct. So with the background selected, with the
color picker, let's just go here and choose a background color so that it's correct. Yep, I think that's okay. Cool, hopefully what I've done so far is clear. So I've gotten the multi merge to connect
to the background node as well as all the four. I'm gonna bring the multi merge into monitor
two. And you can see that what the multi merge
has now done is that it's placed all the four inputs in the middle. That's not quite what we want. We want them to be spread out because you
want to see the LOKI being displayed correctly. In order to do that, I'm gonna bring in a
transform node after the L and start positioning them correctly. So let's move the center, maybe there. Another one here. Move the center. Move the center. Cool, and you can see, you can already see
the effect coming through, right? There is a change in font every frame. If you go back to the edit page, and I'll
look at the original, you'll see that there's a glow around the edges of every text. So let's do that next. You can create a glow by adding a soft glow,
decrease the gain, increase the glow. So something like this, if you go back to
the edit, you'll notice that the edges of the font is not very sharp, it's a bit blurry. And we can achieve that by adding a blur node,
a Gaussian blur node for every text. So what I'll do is, okay, cool, so that looks
good. The next thing, if you go back to the original,
is you'll notice that there is some sort of a grungyness here going on. And the other thing is the font kind of seems
to shake its position every single time. There's a camera shake going on. So let's bring that in next. So after the soft glow, I'm gonna add a camera
shake node. And in the camera shake, it's a little too
much now. So let's bring the overall strength down somewhere
here, and the speed low as well, and the randomness low as well. Cool, and I'm just gonna copy this contour
camera shake and add it to every node here. I suppose the other thing that we're missing
right now is in the reference, even the glow kind of changes between frames. So you'll see that the L has a lower glow
than the O, but in the next frame, the L has higher glow. So let's maybe work on that as well. So go back to Fusion and choose the soft glow
node. And in the glow size, what we can do is we
can see that a number between say a 10 and a 30 seems all right. So maybe I can write an expression here and
say that 30 minus 10 into math dot random plus 10. This formula will not give me a value in every
frame between 30 and 10. So that's what that formula will do. So I'm just gonna copy this formula and add
it to the other soft glow nodes as well. I've downloaded a grain pack. So I'm gonna use one of these grain packs
here, maybe this one. I'm gonna add it to the background node. I get all my assets from Envato Elements as
a subscription service. It's of huge value to me because I can get
WG templates, stock overlays, stock videos, music, and sound effects with just one subscription. And for me, that's incredible value for money. And the link to Envato Elements is in the
description. And in the merge, let me reduce the blend
really low, something like that. And maybe also make the apply mode screen. So you have something that looks like this. Maybe the blend is a little too high. Let me reduce that even further. And the other one I want to do is also add
a grain node. And you can play with the grain size and the
grain softness. Now I'm also gonna bring another grunge pack
that I've downloaded. With the merge two selected, reduce the blend
and apply mode, play around with it, right? It's see what works for you. So maybe that kind of works for me. The other thing I've noticed is that the original
animation has a hint of a greenish color where the O's look. So look at it. It's a bit of a green color going on here. So I want to bring that green in. In Fusion, after the multi-merge, I'm going
to add a color corrector. And in the color corrector, I'm just gonna
push this towards the greens, just tip it. Not too much, just a hint, something like
this. The next thing you'll notice in the original
is that the animation kind of zooms in slowly. So I'm gonna do that next. So go back to Fusion. After this color corrector, let's add another
transform node. And with the size key, key framed all the
way from the start to the end, let's just increase the size. So you have something that looks like this. So this is going to work for our first 320
frames. We still have 80 frames at the end, which
is the hero frames. The difference between the hero frames and
the initial frames is that the font does not change the hero frames. You need the master low key fonts present
there. Now I've gone ahead and downloaded those fonts
and you can find the fonts in the description below. What we'll do is go back to the edit page,
duplicate the Fusion composition that we just created. But I'm gonna reduce this down to eight frames
only because I need 80 frames and then I'm gonna stretch those eight frames to 80 frames. So one, two, three, four, five, six, seven,
eight. That's it, that's all that I need. So I'm gonna cut this down and bring this
composition into Fusion. Now in this case, I also want to remove these
Ls and I want to only retain a single font. So let's just do something like this. So the fonts to use are for the L, you use
US Angel. For the O, you use Old English Text MT. For the K, you use ARB 85 Poster Script. And for the I, you use Cloister Black. What I also notice is that the camera shake
seems to be exactly the same for each of these and I don't like that. So go into each camera shake node and just
change the randomness of each of those so that they seem to move independent of every
other thing. Yeah, that's okay. So now what we have is we have two Fusion
compositions. One that is 38 frames in length and the other
that is eight frames in length. So let me create two exports. So I'm just gonna call this export one. Cool, now come back to the next one, do the
same thing, deliver, add to render queue, render. Bring an export one and I'm gonna retime this
to 10%. And if I look at it
and then bring an export two and then retime this as well to 10%. And if you play them together
now all that's left is to add the music. So let's drag in the sound. I had loads of fun doing this and here is
a Netflix animation tutorial that I did a couple of weeks ago, which I think you will
also enjoy. So go ahead and watch that next. Bye.
