My aim within the industry is to become a narrative designer, a role that combines both gameplay and storytelling, as such I felt a fitting specialist project would be to create my own dialogue system as well as a story, characters and dialogue to go with it, which I felt worked well towards my aim, I was also planning on doing all the modelling for the assignment as well.
For the first part of my specialism I focused on just getting the framework for the dialogue system working, a challenge in itself given the sparse resources on the subject of dialogue based systems, the few that are out there are limited and outdated and inevitably lead to more problems down the road anyway.
However by the end of the first semester I had a solid, if not a little janky, working base to build upon going into my second part of the project, where my goal was to improve upon the system as well as write a story and dialogue with about seven characters to talk to.
Of course not everything goes to plan, in fact it rarely goes to plan, as such there were numerous changes throughout the project for a number of reasons, some good and some bad.
One of the first major aspects of my plan that I changed was myself doing all of the modeling for environmental assets and for all of the characters as well, I instead opted to use premade assets and to do none of the modelling myself, I changed this for a number of reasons but the most prominent would have to be that it was for the most part irrelevant to what I want to specialise in, nameley narrative and gameplay. I felt this would have distracted from the more important areas of my project and at the same time they didnāt really push me towards my own goals. This combined with me being somewhat burnt out when it comes to modeling for sci-fi environments ultimately led to me deciding against doing the modeling for the level. This would also free me up to work harder on both the dialogue system and actual dialogue for the characters, as well as giving me additional time to bug fix and refine the game rather than rushing out a project that does everything in a medicor fashion.
In the end I feel this was definitely the right call for the project as to try and do all the modelling would have only led to lesser results on all fronts
All environment art assets were acquired from the unreal marketplace, all character models and animations were acquired from Mixamo
So when planning out the story and characters for my game I had planned on their being more characters than ended up being, I envisioned about eight characters in the game that the player could speak to and gain information from, as the story is a murder investigation I was going to have four suspects and four information givers, however I found that too many characters would be too confusing to keep track of and two of the information givers were rather meaningless to the story. Also wary of the games length becoming a tad overlong I decided the best course of action would be to merge some of the characters together, this kept the story a little simpler and a little more concise while also making the four main suspects seem more intertwined a little bit.
The end result is five characters in total, one information giver/ tutorial and then the four suspects whom I merged most of the information giver dialogue into.
One thing that I decided to add as oppose to remove was objective markers that point the player towards the characters they can talk to as I felt it would be too easy for player to miss speaking to a character by walking past them and just not noticing them, the objective markers ensure they notice all the characters.
They objective marker works similarly to the game Dishonored, were when facing the marker it appears more prominent but when facing away from the marker it scrolls around the edge of the screen, it also tells the player the distance between them and the character they wish to speak to.
It was actually rather simple to create in the end as the instructions for how to make it were easily findable online. Ā
A feature I was originally planning on having a journal list or evidence list of sorts that would provide the player a way to see the info theyāve already gathered, I managed to implement this in a basic form but I ultimately removed this for a few reasons. It would make the investigation too easy and I wanted to strike a middle ground in terms of difficulty and I wouldnāt have enough time to properly balance it, the second reason I removed it was because I wanted players to pay attention to the dialogue and not just skip through only to read the journal list at the end, I felt this was the best way to encourage players paying attention. And finally the last reason I had it removed was because of me having lowered the number of npcs in the game which made the story and characters easier to remember, had I kept the original eight characters that I had planned then I would have probably kept this feature too but with the lowered amount it works fine without.
So obviously a major part of my project is the dialogue system so I felt it extremely important to get this right especially. Ā
I already had the framework from the first part of my project but their was still a lot to improve on, one of the first things I aimed to improve was the possible length of a conversation sentence, in the original dialogue system for the first part of specialism it was restricted to one sentence replies from the npcs which didnāt create for very fluid dialogue and made it extremely hard for me to get an adequate amount of information across, so I saw this as the first thing to improve, I blueprinted a system that let npcās have continued sentences which allowed me convey more information to the player and extend conversations naturally where they would occur which gave me a lot more freedom when it come to writing the dialogue.
For much of the rest of the project, the improvements I made to the dialogue system were to make it easier to use on my end, making it faster and more efficient for me implement my own dialogue and sorting the blueprints into reusable functions that allowed me to streamline the process, while it still wasnāt great or even that practical to work with, it was still a lot better than before the second semester.
Finally I improved the UI of the dialogue system, where it had previously been using a weapon wheel for its dialogue wheel, I changed it to better suit the six option limit I intended for the dialogue. I also added an overlay so that the text stood out against the background and option boxes with lines drawn to the wheel to make it clear which option the player is picking.
The dialogue was surprisingly difficult to write for, while I was no expert when it came to writing in general, Iād had a little experience, but writing branching dialogue for a dialogue system proved quite the challenge.
Once I had planned out the story and characters I moved onto the dialogue, in order to differentiate between the characters, i planned each characters dialogue a little differently to each other, both in the structure of the dialogue branches and the dialogue itself, after all the dialogue was planned and written I tested it out in game to see which kind dialogue structure worked and which didnāt, finding some dialogue to be too easy to fail too quickly I went back and rewrote certain characters dialogue with one of the dialogue structures that worked better to ensure that players didnāt feel cheated or dissatisfied with how easy it is to fail.
A number of challenges arose when writing the dialogue however, specifically in writing enough options that were of equal quality to each other, it wasnāt easy writing dialogue for options that i knew no one would see. but because of the possibility that someone might just by chance happen to click that option I needed to make it the same quality as the rest.
Further challenges arise in trying to write dialogue for a game feature, were dialogue needs to be a certain length because too much could cause it to go off screen or clip into another option and each npc reply is another set of blueprint nodes that I have to make and implement, its also an interactive story driven game so replies need to be long enough that players feel they got a worthy amount of story from picking an option but not too much otherwise it would become boring.
There is also the issue of having to write dialogue that makes sense regardless of what players picked previously, which makes it really hard to reference back to previous things stated during dialogue because it's possible that this specific player didn't pick that option in the past and so it wouldn't make sense to them now.
The dialogue system, mostly, while it is still very janky even after Iāve polished it some I feel it ultimately succeeds at what I was hoping to set out for.
From people who have tested the game i've heard they liked the dialogue specifically, though Iām overly critical of my own work it is nice to hear such positive feedback all the same, it is however still an area that I wish to improve on going forward in future though.
Iām glad I managed to get the character models and animations in as even though they don't specifically relate to narrative design they do help players identify and reaffirm the characters personalities for the player.
Also the dialogue system, while I am overall pleased with the result it doesnt change the fact that this was the most difficult thing to make and refine of the whole project, hampered by its initial limitations as the tutorial that I used as a starting point for it clearly did not intend for it to be used for something like this and that it's incredibly impractical for such a full scale project.
Another thing that didnāt turn out like I hoped was the level itself, while I created a much larger level with a wider array of assets, I was unfortunately unable to get it it package Ā or build due to an unknown lighting build error for which I couldn't find a fix.
