I'm making open world Playstation 1 graphics.
interpolator
From another thread, I promised to show something.
I've been inspired to make something loosely based off PS1's limitations. I want to do it to better understand real world game limitations and to create something I've dreamed of for a while.
I did some modeling and texturing but half way through, I realized something was wrong.
Even though I've been careful with how I added my geometry, I've realized what I'm doing is too much for my objective. Since I want to have a lot of assets, I'm going to have to reduce my polycount by almost a half for everything.
This means remaking everything I was working on so far.
I wanted to post this thread as a learning experience. I'm starting to understand how easy (or hard) it must be to create assets, only to have to scrap them or heavily modify them to meet certain game industry goals.
I've been inspired to make something loosely based off PS1's limitations. I want to do it to better understand real world game limitations and to create something I've dreamed of for a while.
I did some modeling and texturing but half way through, I realized something was wrong.
Even though I've been careful with how I added my geometry, I've realized what I'm doing is too much for my objective. Since I want to have a lot of assets, I'm going to have to reduce my polycount by almost a half for everything.
This means remaking everything I was working on so far.
I wanted to post this thread as a learning experience. I'm starting to understand how easy (or hard) it must be to create assets, only to have to scrap them or heavily modify them to meet certain game industry goals.
Replies
For reference, I would suggest to check out the Lowpoly thread. There are many ps1 spec assets there.
I mean, I would love to play this on an actual Playstation (I have one!), but at the same time, I realized not everything is going to be accurate.
For example, I was reading about how the PS1 lacks a z-buffer, which made sorting through objects extremely painful, and that the CPU wasn't powerful enough to handle large polygons, so something like my skyscraper might have already killed it.
The actual art I'm making though will be very faithful though. I plan on keeping the entire geometry count somewhere between 3000 - 6000 triangles, textures will be smaller than 256 x 256 with tons of indexing to save on memory. I will also try and limit the use of lights, etc.
Nobody does PS1 games and graphics nowadays, even smartphones are more powerful, seems like you're wasting your time. If you want to do some nice low-poly stuff, do it , but for mobile or something. Here you're going to put something in your portfolio that is completely irrelevant for today's standards.
Well if you just do it for fun nobody will try to stop you...
I agree with this you would not learn much that is fantastically relevant to today and starting a new project or finishing an old one is best for the learning side.
For the fun side do whatever you want really no one is really going to stop you there.
However, I kept going into next year. Rather than drop the project, I believed I rather keep going, because I really liked doing it.
I don't think this is a waste. For example, ever since doing this assignment, it has taught me how I should mange my polygons carefully.
Now you could say the same for mobile, but I was only inspired to start this because of Playstation.
Any ideas where you want to take it aesthetically? IE hand-painted, photo-based, the Sims vs GTA III, etc. Your texture work is going to play a huge role in how well everything turns out. If I was you I'd probably start uving and texturing some of those assets if you haven't already to prevent the feeling of having an overwhelming pile of models to texture at once. Make some sweet tilable textures laid out smartly so you can reuse them multiple times.
Taking a few of those and seeing them through to completion could also be good for portfolio purposes.
All of that said, best of luck!
You're right about the textures. I have devised a "master lighting" system that's going to take advantage of it, heavily.
I'm going to showcase this environment under multiple lighting conditions viewed from every angle. I'm hoping to impress people with what will only use 1MB of memory, for everything!
I feel like you are letting yourself get caught up in extremely tight technical details and making things difficult for yourself for no reason. Making a cool 500 triangle car? Cool! Feeling like you have to half tri count to theoretically match system specs on a 20 year old console that you aren't actually going to run this on? Not cool!
Opinions aside, I found this cool thread that had a tidbit of info about cars from Gran Turismo, which you may find interesting/useful:
http://www.polycount.com/forum/showthread.php?t=133462
I would hate to see you spin your tires and not get anywhere because you are overly focused on the technical things. Hopefully that doesn't happen. Good luck!
Say hello to the Ford Transit and Bugatti Veyron!
To be honest I don't get the alpha plane wheels and especially the blocks behind them. It does seem like those blocks at that size would be visible as blocks when the plane is masked out.
I'm hoping when I get to texturing them, and I post more angles, it wont be noticeable. However, if the textures don't fix it, I will go back and retouch the meshes.
All 3 cars right now are under 300 triangles which was the spec of PS1 games. Only the Bugatti went slightly above that, just because I think it needed the most detail IMO.
I really don't think the PS1 era used that many smoothing groups (split vertices).
But all in all, you're seriously just wasting your time, even if it is just for fun. Use your time to improve your real portfolio. That's what will get you a job.
I hope you don't take offense at what we are saying. We are trying to say this in the kindest way possible. We're sincerely just trying to help.
But if I'm getting your guy's blessing to officially stop, then I will.
Honestly, wtf? Either I am getting this whole community and industry completely wrong and there is a polycount-has-to-like-you rule that I never heard about or you have some odd understanding of all this here. Or you are taking us all for fools - in this case: nicely done, you caught me :P
Go and do something smart with your time and do not just try to appease people on the internet. They are giving you smart advice and I am quite frankly certain that there is still a lot you have to learn to become a good 3D artist so you should listen to them and not try to proof something (especially something that most of the people you are adressing deem senseless anyway).
From what I've seen so far quality (of work and knowledge) gets you respect around here. So if you put up something good that's done correctly people will rather listen to you than if you try to fullfil your own set up little quest that has nothing to do with the reality that you will face in a job.
You don't need to argue against every crit, or everyone who thinks you should do this, that, or the other thing. You can listen, or not! It's up to you! Just keep creating.
Though I don't think this will look good on portfolio as a PS1 stuff. But if you turn it into lowpoly for mobile games instead, it would still be relevant to nowadays industry, and would look better to show off.
This project reminds me of the first Gran Turismo game. The dev team on that project really pushed themselves to the limit. Would love to see that kind of dedication here.
You're embarking on an unusual project but it will work out if you put your heart into it.
If you want to keep it around the ps1 era of fidelity in restrictions then by all means that's your choice. I just think it might help you improve a bit more by selecting one of your favorite car models you've been making and really focus some time on polishing the UVs and texturing so it's something you can really be proud of and show how you carried it from start-to-finish. Nailing that process down on one model will help you decide on the process and overall style for all of the others, besides the current set poly-limit.
It's also sometimes easier to give more-in depth critiques on one or two models rather than an entire games worth, so people can give you better suggestions as to what areas you may need to focus more on.
Just my 2 cents, feel free to take or nay whatever parts you will
Here's my progress of today (I'm exhausted).
All of Sunday I was unwrapping the first car (Crown Victoria). I've also begun some very early texturing. It's very hard when I don't have photos of the real thing to work with, so I'm trying to recreate it through photoshop and the internet.
The Crown Victoria is finally textured. It took me 5 days to create.
Also, the lighting is baked in. I plan on showcasing a different time of day (morning) but only when I finish the night level first.
Here are the two textures used.
The emissive map is colored but for no reason. You have the exact same colored lights in your albedo too, so the emissive can be gray scale. Or you could even select those pixels with an IF because their color are really different (a lot brighter) from the rest of the car so they are easy to find.
It would be more shader work and more flexible then but less textures if the "lighting" would be just a parameter in the shader, so you can change it dynamically with adjusting a vector parameter through matinee based on a daytime variable or something.
The UV of the side of the car is really stretched.
Also, to me the 256 texture resolution looks big, 128 could still display the details, and it woulf fit better to the specs in my opiniom. There are many many things in an open word game, so you should be really careful with the texture resolutions.
About the 5 days:
I made this in 6 hours from start to finish so the realistic creation time is a lot more shorter. We would have 5 days for a highpoly car at job lol.
(Inspired by Delko's cars)
After these all, don't get me wrong, I like the idea, I would just like to help you approach this better.
You're right though to be careful, as only the car right now is at that resolution. Everything else has been 64x64 or 128x128 with lots of indexing.
4k and 8k is possible nowadays but you really rarely use it because its very expensive so that its possible doesn't mean its optimal.
My another thought is that you should really do some texture atlasing.
Anyways, making "ps1 graphics" in UE4 doesn't really make sense, it will still have huge system requirements, because the code is full of with "unnecessary" lines and its not optimized for old devices so you couldn't run the end result on a ps1 even if you do everything correctly compared to the ps1 specs. Maybe it would be better to try an older engine. DarkBasic 1 for example would give good performance and with that, the result would be a lot closer to "ps1 specs".It can still do everything that you need. There are many old engines and engines that has just the old features, it would worth to check out.
If you want to make it optimal for ps1 then I would suggest this. If not, and you just want ps1 look , then stay with ue4.
For those who were interested in seeing a modern day PS1 game, do not give up! I still love and cherish this project and want to complete it some day!
However, I am aware this is something that is not immediately going to get me a job in the game industry. Because of that, I have to heavily rethink where my priorities lie, as I must balance work done at school and at home. This project is now being pushed back to being a "personal art piece" while I finish my real portfolio for AAA and Mobile video games.
Before I go on hiatus with this, I will share what was being worked on. I hope to revisit this again maybe next month or April.
Gas Station:
Church
City Hall
Bye for now!!!!!!!!!!!
I'll still continue with my knife but I'll also revisit this project from time to time.
Personally, I'm really interested to see where you're going to be taking this and I find the screenshots we've seen so far has a charm to them.
However, for your good, I would agree with others that it would be best to focus on something more relevant to today's industry. This can then be a side-project to do in your free time
I still want to work in the game industry, I just hope people understand that this project isn't trying to take away from it.
I'm never going to submit this thread to Ubisoft, Naughty Dog, Activison etc requesting employment (unless they do contact me and are interested in it). I just want to work on this whenever it makes me happy.
You're doing an awesome job. I just wanted to chime in about how gay I am for super lowpoly graphics. Keep it up dude!