DEVELOPMENT

18

2,093

Atonal Dreams Weekly Update 18 - Cleanup, Script Writing Next?

4 years ago - Edited 4 years ago2,672 words

I've cleaned up the intro cutscene a (little) bit by adding or changing some structural stuff, plus I've planned the story as a series of dialogue scenes. Next I should write the script in its entirety, though what about the demo I keep teasing?!?

I've not done as much as I'd like this week because I've been ill. I don't think it's covid! Seems to be more of a digestive system thing, though it's been lingering for longer than I'd expect from such a thing so hopefully it's not some kind of cancer or something... though that wouldn't be surprising considering all the other wonderful boons life has gifted me with. I'd see a doctor, but lockdown makes that difficult and risky, so hopefully it's something that'll go away on its own rather than getting worse.

I've been meaning to get around to replying to the comments on the Undertale post, since people had a lot to say about that, but this combined with depression and more specific negative feelings surrounding that game have got in the way, annoyingly. I've not even started a second playthrough yet.

I tried writing out a personal post about some of that and some other things, like three times, but kept scrapping it. Annoying!

I have been playing the latest Pokemon Sword DLC (Crown Tundra) though, which released a couple of days ago. I might write about that when I'm done, since I planned to write about Pokemon Sword in general but never got around to it.



I did do a bit on Atonal Dreams, at least! First, I made some tweaks to the scene I showed in ∞ last week's post ∞. Here's how it looks now:



...So not THAT much different, really, and it's still certainly not finished. Most of the changes I made were behind-the-scenes stuff that aren't visually obvious.

I fixed some of the ugly issues with Pierce summoning a monster from last time, though the particle effect is still wrong and he should play a sound effect but doesn't. I've been noticing that the musical sound effects for spells clash horribly with Pierce's battle music in the background and I can imagine people will complain about that... but eh, we'll see. I'll just leave it for now.

The dialogue's not really changed, but I'll get to that!

There's also an issue where Collie says "I am taking a turn!" (a test of turn-triggered interjections, temporary dialogue), where her dialogue box is placed incorrectly; I need to fix that.

I also ended the battle by having Savitr use a spell to one-hit-KO Pierce, which obviously wouldn't be any good for actual gameplay! The battle's meant to serve as a tutorial, so I'd probably lock out most of the skills and have Savitr take a defending/teacher role while explaining things to Collie. That still needs planning, though I've been focused on other stuff.



Someone mentioned in the previous post's comments that the framing shots of the characters felt amateurish, which is fair enough since I am an amateur and cinematography is a whole art in itself. How shots are framed is something filmmakers think a whole lot about, but which I've never properly looked into before, or even really paid attention to. Seems what I'm doing here is called a ∞ Shot Reverse Shot ∞, where the camera alternates between characters. Sometimes all characters except the one speaking are out of shot, other times you see from over another character's shoulder. I experimented with that:



Looks a bit stupid though, I feel, because Pierce is too far away. You could say it's maybe fitting if you're seeing what Savitr and Collie are seeing - and they're seeing him from far away - but as the very first shot, does it work? The way battles work, each side has eight 'slots' which characters or monsters can occupy, with fixed coordinates. During this scene, the characters are locked into slots on either side of the battlefield, and I couldn't position them closer together without recoding a lot and losing the smooth transition to battle mode.

I could also use an establishing shot of all the characters from the side:



This is a screenshot from the conversation editor I made for myself, where I can edit everything fairly quickly and in a way that's almost like a game in itself for me! (I've considered making a 'game' that allows the 'player' to make their own scenes using a similar system, though I'll talk about that some other time.) The lighting's not set up like it is in the actual gameplay though, so this all looks quite dulled.

I feel a side-on shot this would lack clarity, as it'd give a sensation of not knowing where to look, plus I feel the characters aren't particularly appealing to look at from this kind of angle. It's like how in the layout of the 'normal' conversations...


(EMOTIONS.)

...the characters aren't exactly facing each other side-on.



Wouldn't that look terrible? Even though it's more likely to be how people would position themselves if actually talking to one another. Cartoon characters - especially simple ones like in many modern western cartoons - are also almost always at a 3/4 angle because for whatever reason our minds register that as most appealing. To give the appearance of looking at each other, their eyes have to look to the side.



I find that psychologically interesting!

I'll probably be more aware of how films position their camera shots and characters in future, though for now I think what I have for this Pierce scene might be acceptable? Maybe?? I'm not sure whether opening on Pierce's not-exactly-sexy face is the best start really, but... maybe it's at least striking??



I composed some battle music for Pierce! That took time, and I'm quite proud of the result, since it's something I've been wanting to do for ages. I've posted the full version on ∞ my Patreon ∞.

It's made up of two separate parts: there's a short looping anticipatory intro, which would be unique to this scene. His actual battle theme begins at the end of the initial dialogue as the actual battle begins.

This is achieved by having the music's intro fade to silence for a line, then the battle theme begins abruptly on the next. Originally I planned to a smoother transition where the music would loop one section, then progress to a different one seamlessly once a certain variable had been changed; I think some other games do this, like maybe Octopath Traveler with some boss battles? FFXV did it with some intense battles against the eidolons (or whatever they were called in that game), too. But there are a couple of reasons why I chose not to.

One is because I bought an asset to handle seamless looping of music tracks. This is a feature I've wanted in my games for literally years, but I could never figure out how best to do it myself; there was usually some audible click during the division, especially if I wanted there to be a separate intro section. I remember this being an irritation in MARDEK. I don't normally buy assets for Unity, but this one was only $15 and solved the problem perfectly so I didn't have to fiddle around with it myself for ages. Unfortunately though it only allowed for an intro section, not for multiple sections that could transition between one another. That'd require some deep reverse-engineering or using midi or something.

I also felt that the fade to silence and sudden start of the battle theme felt more impactful anyway, so it doesn't feel like I went with an inferior option due to limitations!



Something I mentioned last week was that I wanted Pierce to emotively shake his 'sharp flute' around once he'd drawn it, but it wasn't possible with the animation setup I had previously. I put some work into making that possible, which took some time. The only immediate result was that brief animation in the scene (though I feel it is an improvement over how it was before), though I've completely redesigned the setup to allow for greater potential in the future.

Here's how the animation state machine - or 'AnimatorController' - looked in Unity before:



Each state block contains either a single animation or a series of animations that can be blended between based on a property (eg "Movement" contains a walk and a run which are blended based on a speed property), and the AnimatorController as a whole has properties which can be altered, and which each state can use to determine whether to transition to others or not.

For example, the posing system previously used two states: "poseA" and "poseB". When the pose was set to change, the animation for the state that wasn't the current one would be set programmatically, and a boolean (true or false) variable, poseAB, would be set to its opposite. If the current state was poseA and poseAB was changed to true, it'd change to poseB, and if it was poseB and poseAB was false, it'd change to poseA.

I hope you were paying attention to that because there'll be a quiz later.

When I started using Unity, I had far more complicated state machines where every single pose had its own state, and I had to keep track of dozens of different transitions. So this did make things easier. It had limitations too, though. The battle pose was separate to the normal emote poses, and each of the possible battle animations had its own state, all of which could transition to and from that basic battle pose. For emote poses to work with the battle pose, I'd need to add transitions between them and poseA AND poseB and... well, the whole thing would be a mess.

There were also a bunch of properties and animations for now-removed features cluttering the thing up, since I'd been working with this all year. I'd been wanting to clean it up for ages!

So I did:



Much better! Now all emote animations AND the battle idle are stored in a single 'idle' state. It still transitions between A and B states, set programmatically, but this time it's a Blend Tree and poseAB is a float value...

...and it's all quite technical and not worth going into too much, but the point is that it took time, it's simpler now, and it's more versatile. Good!!



The biggest thing that I did this week was another pass over the planning for the game's plot structure.

Previously, I talked about how I'd decided that the game had three parts: overworld, underworld, and drealm sections. I'd decided that the overworld could be divided up into a few small islands, since they were easy to make and you could satisfyingly 100% complete each one, though I'd need to work out a way to travel across the seas between them.

I've had that at the back of my mind for a while, and I've been wondering how many new assets it'd require. I'd need to plan that sea-traversal method in a way that's harmonious with the lore but restrictive enough for gameplay reasons. Should it have its own music? Should each of the islands have different music? Different tilesets? I'd been thinking I've finished making most of the gameplay mechanics now... except for all that, and it was a heavy thought.

I think I might revise it, though. Make it simpler, kind of like what I had originally. The overworld could just be one biggish linear island, with no travelling between separate islands at all. No need for new assets or mechanics or anything!

As I said, I went over the plot again. I've got a good idea of the general structure of the story, but I haven't written it out as a script yet. Just a sort of synopsis, similar to what I had in the planning document I released alongside the Steam version of MARDEK. Going over it again helped refamiliarise myself with ideas I hadn't touched in weeks, but I was also able to plan it out as a series of specific conversations, listing what each of these conversations would cover.

I've got 33 conversations planned, though perhaps with some tweaking I could streamline that a bit.

When I made other games, like MARDEK, I wrote the dialogue scenes as I went along. I'd make a dungeon's tileset, monsters, music, dialogue scenes, and I'd have all of them finished before moving onto the next.

It might make a whole lot more sense though to spend a week or two writing a rough first draft of the entire script. This should allow me to add foreshadowing to the right places and make sure everything flows well and characterisation is consistent.

If I had the full script planned out, then I can imagine the rest falling into place fairly smoothly!

One of the biggest issues I'm facing at the moment is uncertainty regarding how to approach this intro cutscene; what the characters should specifically say at this point in the game. If I had the big picture planned in detail, I wouldn't be blocked by so much uncertainty.

Something people seemed to like about Undertale is that it allowed for so much choice that every playthrough was different. Some scenes had so many branches they'd require complex flowcharts to illustrate, and there were a whole bunch of dialogue scenes available for players curious to examine every nook and cranny multiple times.

I'd like to have some elements of choice in this game, but I definitely don't want to go down that route! Apart from being completely impractical in terms of time and energy costs - I don't want this to take as long to make as Undertale did - I personally don't even find that kind of structure interesting. I'd much rather have a fully satisfying experience my first time rather than being burdened with the thought that I had to replay it to experience it 'properly'. I'm already ugh-ing about the thought of having to replay Undertale. UGH.

So I'll be trying to write something coherent and hopefully satisfying, rather than a series of disconnected 'things that happened' like I felt Undertale was.



It seems sensible, then, to work on writing the script next. But that presents an issue: What exactly do I do about this demo or alpha test or whatever that I've been talking about for ages?

The plan was for it to be the starting island, including the starting cutscene, but it makes sense to write much or all of the script before the starting cutscene so as to have the best chance of giving a first impression that's reflective of what's to come.

I'll probably remake the map too (not much of a loss; the one I have currently only took a few hours).

Maybe what I could do is use what I currently have and add some temporary dialogue - like the ∞ Belief demo ∞ - and add that to a Steam page for some alpha testers, probably my patrons...

...but I feel that the plot is the main focus of the game, and that wouldn't be getting at the actual plot at all, so I don't know.

What do you think??



Here's the quiz you were absolute waiting for and didn't forget I mentioned:

"If poseA is travelling at 3 hectares per gigabyte and poseB has ten apples, what are the primary narrative themes of Beethoven's famous piece, Ravel's Bolero?"

I want your answers on my desk by the time I've finished writing this post. Why haven't you handed them in yet. Why. I'm in tears. There's no justice in this world. Buy Sindrel Song.