Firstly, allow me to say a massive thank you to the team, Cybertronix. This trailer was amazing, not because I worked on it, but the sheer effort and work put into everyones piece, even from a less optimistic members, and for the smallest team out of the three, I think we pull it off really well, and with that I want to thank each member individual.
Micheal Prior,
Joey Brown,
Stuart Wade
and Curtis Stoyles.
And above all, a thank you to Mel Brown and Chris Dewey, for whom without there nagging, tutoring and constant help and advice, not to mention pulling the showcase off as well, the end product probably wouldn't be here, or, nearly as fantastic as it end up, and that goes for all teams, when I say that, the end products where amazing.
This was our final product, compiled by Joey. One thing our trailer was complimented most on, was our use of lighting. I will say now that some of us, where not that strong in modelling, or had rather simple areas to replicate. With that, we knew we had to bring our scene alive more, and give it that "UMPH" we needed to get it looking 100%.
I used various lighting techniques in my environment, including direction, volumetric and even a couple of omni lights, applied with shadow mapping. One thing I will note, here that was key, was render time. Render was important in order to get our particular scene ready for the deadline, because we drove ourselves right up to the frontlines with our intial design, we ended up with very little time left to render our environments out. One problem with mine was, it was taking very long to render each frame. (About 10 minutes per frame, multiply that with 800 and your looking at 8000 minutes to render the entire thing..) It was also not very possible to leave machines rendering over an alloted time, because we are in a college, our machines are likely to be required for teaching purposes, and we have no render farm, so its not very practical to render over night, instead we used a different technique, but first, I want to go other a problem that I solved to cut down my render time.
SKYLIGHTS! Thats right, there a bane! Well, they apply ray tracing to creating ambience light around the entire scene, unfortunatly this kind of effect generates a lot of overhead, and takes time to render each frame, it wasn't until I removed the skylight to help relight the entire scene (it had gotten far too bright) that I realised the overhead that Skylights cause. So remember, Skylights = Bad. Least when your limited in development time :)
So now I have expressed my dislike of skylights and there treachery, I would like to specify how we overcame some of the rendering issues. Now, recall how games and movies, use frames. Each frame depicts a moment in scene, and games/movies move around 25 - 30 Frames a second, roughly. Well, thats how we did it. We didn't render it out as a full AVI movie, that would be tedious for the sake that AVI is actually larger in file size, but also, if our rendering machine crashed (Which they did.... alot) then we lost the entire file and had to restart... ugh.
So we rendered each frame out in bitmaps. That meant any crashes, problems, we could just start from the last bitmap it got to. Perfect, no more rendering from the start, which meant we kept a linear progress level.
Overall however, we managed to get it pulled through, Joey himself had to re-render his environment, due to a lack of frames given, it resulted a rather ugly and jumpy framerate that didn't remain consistent. Which brings me to the reviews. One thing i've alway said from the beginning was that I wanted to keep a trailer that was consistent through out. I specified it so much so, my team practically made sure each environment led off from one another.
Looking at previous teams, and also this years, I was able to see how they combatted consistency issues. Because my team was relatively small, and we were pretty much in the same idealism, we was able to generate a very consistent idea, that help with generating the trailer. Thanks mostly to the back story, that Joey typed up, we have a very linear path we could follow.. "From Joeys Environment, it follow down the stairwell, and into Priors".
While, it wasn't 100% consistent, give or take, we still used the same texture gallery (Photo's we took from around Saltaire) and we still managed to keep a similar pace in framerate consistency, which I loved alot.
So, one trailer I immediatly realised how they combatted differences in environments, was last years work called T.R.A.I.N.
The use of a train to explore each environment was a great idea, and also, the storyline chosen, allowed them to work with any nature of environmets, this created a sense of consistency, and an idea that I believe Wolfpack used, the idea of time and portal to other lands, which generated a level of consistency and exploration.
One thing to note, moving from the topic of consistency to comparison of quality, is that Wolfpacks trailer had a lot of good 3D work, one thing we did notice, however, was how easy it was to create a detailed interior rather than an exterior. None of the exterior environments where as highly detailed as the interior ones, from like Joseph Lee and Joshua Williams, we believe it to be down to how much detail can you really put in an expanded environment.
While I admit, if I was to do an interior it was unlikely to be as good as those in Wolfpack, I still stand by that producing an interesting exterior is much harder, with so much space, everything feels... empty. Suppose, the use of everything is too compact in an interior environment could be used against me there. :D
Back to my trailer, each environment was created in correlation with a level. The exterior was the introduction, the plaza level would be in gameplay, for searching and findin the entrance, hence the torchlight, the tomb would be simiulated with traps, and made to feel rather like the level on Amnesia, stuarts was going to be a way for the player to be introduced, rather like a tutorial level, and of course, the final level, the wine, would likely occur a boss fight.
Intially, we was going to add in various elements, a stone samurai was actually in development by me, however time scale, and my lack of ability in modelling, forced that idea to be scrapped pretty quickly.
But so you can laugh, where is some development shots I took of the stone samurai, armour breast plate I was working on developing.
As you can see, it wasn't coming along too bad, now I look at it, it seems to be better than I intially realised. Usually, I like to compare my models against professional examples, but thats obvious to what would be considered better and what I would want to aim for, I mean, I know I wanted this to look like a Samurai, instead, I am going to compare this humanoid model, to a humanoid model developed by one of the first year students, who happens to be pretty talented with 3DS Max, and compare my experience with the software, to his.
For all purposes the model above is actually under the intellectual property of Zero Matrix Interactive, and use of the model without permission will breach copyright. I'm serious.. :P
Credit for the model, goes to "Tom Pearson".
However, you can clearly see the amount detail in this. One key factor to realise here is that the armour plate I developed has around 1400 Polygons. Including smoothing, in some areas. This model, has grabbed around 2800 polygons, and its finished. So, my unfinished samurai would have likely had four times the amount of polygons that this model did.
One other key fact is well.. its finished. But, its arms, and hands are incredibly well detailed. I remember staring at awe in Ivo's attempt at creating hands last year, but now I look at Tom's work, no offense, but it blows that out of the water, and not just the hands, but the entire body posture. I do have some development images of the model, and i'll display a few for visual comparison to my development images, but I cannot display any updates, other work or textures used on the model, for sakes of IP.
Well, I was sort of going in the right direction... sort of. The face is my favourite one here, it has NURB SubDivision applied, to increase the smoothness and depth of the model, however, we have removed that because of the increase in polygon count it creates. However, the head is something I struggled with when developing the stone samurai.
My previous posts, depict my development of my own environment, but there hasn't been a direct link between the two. So here I will show and explain what I did in certain areas.
This time, i'm comparing to the real thing, because, when I was developing the actual building, I wasn't aiming for any particular style, I wanted to create a photorealistic environment.
Immediatly you can see the massive difference in the two images, if you can't, then your just being nice, and deserve a cookie.. I decided pretty early on that I wasn't going to replicate the frontal area exactly.
One reason for this was development time, of, what we lacked of it. We didn't have long to work on developing the environment, and as I said previously, exteriors are much harder to develop, due to the large size and open spaces. Intially, I had benches and bins about, but they looked unsightly, and decided it would have been best just to remove them them.
Knowing I had a huge grass field, and knowing that I had to do something, I decided to play on the renovation scenario we had planned, and its easily seen in Joey's environment, by making pillars and collasped environment present. This was great it allowed me to keep the look of the building, but present a essence of my own design into, and above all, cut out development time. I also applied some fog and volumetrics for that added depth, and the church and hills in the background were added for increase effect and to simulate a larger environment than it actually was.
I am asked alot, when I show people the actually environments, why I didn't develop the surrounding areas of Saltaire, well this was simple. Firstly, I wanted the Salts building to feel rather empty and alone, in comparison with the rest of the town, I achieved this with a large open grass field, and fog, which made it feel spookier than it actually was.
Another reason is I only wanted the Salts Building, why waste time building Saltaire, when i'll only show a small part of it. This was vital for both production time (I keep drilling that in don't I) and also render time.
I've discussed alot about trailer asthetics but I haven't talked much about management. I'll leave that for another post, as for now, i'll end this one. Hope you found this interesting, and keep a look out for productions posts, where i'll detail areas of my environment and steps I had to overcome, you may even learn something. I know I did.
No comments:
Post a Comment