Moderator: Graphics Moderators
It's very nice that you ask, and coincidentally I have actually restarted work on BRIX lately after quite a while.
The current state is basically changing every day, I am trying to do so many things, so I am trying to settle down and realize what things do I want to do first.
The plan at this moment is to make the current BRIX look better instead of adding anything new. So I am working on a version 0.0.2 which does literally nothing new.
As I mentioned in some posts around Christmas 2016, I am really focusing on hand painted postproduction in Photoshop. Over the last 8 months I have been painting a little bit here and there and eventually I kind of have many things done. Not everything, but a lot.
Now I just need to put it into the pipeline which spits out the final sprites. I have succeeded in doing that for rails, trees and rail signals. I still need to do quite a bit of painting but it's going quite nicely.
In a bit more detail:
- After looking at BRIX for a while since the first release, I started to Really hate the super flat and kind of empty style, so I am adding more details, even if "not real details" ... therefore all terrains have a very basic and temporary noise drawn over them, defining more colours and stuff. Long term plans here are multiple, simpler of them is to create land textures on my own and place them there instead of the rather generic noise, and probably eventually I will completely rework the terrain tiles as I think I am starting to discover the true visual identity of BRIX.
- Rail signals are something that many people either love or hate. I believe the system in the design is really good as it lets you do all the things - distinguish all of rotation, signal type, and current state (red/green) at a glacnce. However, similarly to the terrain, the colours and execution of them is just plain, bland and boring - it has no detail for you to look at. So again I am painting over them to make the picture more rich, have better contrast, and seem more "full". Rail signals have their pipeline working but still need to paint most of them, which I might even finish tomorrow or something like that. Eventually I think I will strongly tweak the signal design so that they aren't such a massive triangular monolith, signals should be more lightweight. But the system of the lights will definitely stay, and the signal type identifiers (yellow, white, red pieces at the bottom) will definitely not change too much. System good, looks bad.
- Trees are getting colours now! Finally. It was just too much hassle to do it in the 3D scene so I am colouring them as a part of photoshop postproduction. However I will remove many of the current designs and replace them with other trees, or even things like stumps, tall grass patches, little rocks and things like that.
- Rails kind of work fine but need some finishing touches, the biggest differences are visible on tunnels which were massively lacking contrast, mainly black ends of the spectrum, but the highlights were also kind of bland.
- The huge spritesheet of bridges needs painting (very carefully to keep it working as many things must fit into each other), I have already started with the tubulars, but I have like 8 sprites out of 50 or something like that. That will take a while, I might even skip bridges for 0.0.2
- Roads might get complete rework, I don't feel too strong about the current design.....
And here I am getting to something very important. Roads were something that I made as the first thing in this project, composed them out of cubes, and called the project BRIX basically for that reason. Spontaneously and randomly. However, as I work on the project very slowly, it is very hard to keep visual consistency mainly because I don't really know what do I want to create. In the beginning I had some vision based around playing with saturation and colour codes, using the ultra clean style to save time on texturing, keeping clean 3D workflow without these, and so on. Over time the style has developed. First I discovered that white terrain won't work too well with recognizing snow and desert, so I started making grass green. Then I started adding noise to the grass to actually make it look like grass - or at least grass resembling of the old TTD graphics from 1998. And now I am removing the bricky roads? I think you can easily see how the original concepts are quite quickly disappearing from BRIX, and this is something that makes me think a lot about what should BRIX actually be. I won't shine a light on that now, but let's just say I have learned many things from all the experiments, I have seen what is totally out of question and which approaches seem to be nice, and I am slowly putting together a very specific vision. Let's see if I ever manage to complete that.
Regardless, I am hoping to have 0.0.2 at most sometime in September, let's see how that goes...
Adding 2 pictures worth ... 1000 words?
- signals-00.png (645.47 KiB) Viewed 5149 times
- (3.26 MiB) Not downloaded yet
The green grass is much easier on the eyes, I like that change!
The geometric cube-like shapes for trees look nice too, friendly and pleasant. You also have a more conventional xmas-tree in the 2nd shot, not sure how these combine. Can't you keep the geometric shape, where holes appear in the flat surfaces, eventually reducing to frames along the light-green edges, and then breaking down completely?
The black of the signals looks good, I hope the path signals also have a nicer colour.
One thing that hit me when looking at the bigger picture was the tunnel entrances at the left. I expected to see some form of connection between the edge of the tile end the bottom of the tunnel wall. Right now it more looks like the rail-bed and tracks were lifted up to enter the tunnel. I am guessing however you haven't done the tunnels yet for the new looks.
I'd consider it quite an improvement over 0.0.1
I am considering to make the basic block signals shorter but I fear it would become messy. Maybe I will just re-design them altogether sometime later to make them more lightweight.
Next up is finishing some final touches to rail tunnels, figuring out what to do with roads, and trying to make trees somewhat nicer, and that should be it for v 0.0.2
Plans for 0.0.3 is at the moment mainly trains. I have started them and it's a big shame they are just sitting on my disk so far. There is still a ton of work to be done on them, but well, everything is a lot of work.
What do you say?
The people who follow this thread since the beginning (which is quite a long while by now) might remember that originally I had some concept in mind. More specifically, I mean the idea that gameplay-unimportant things would be less saturated and important things would be rich in colour.
The first post in this thread in fact demonstrates basically that ... download/file.php?id=186186
Over time I realized this is quite boring as it has literally nothing to look at apart from the overall shape of the terrain, so I started giving the land a very slight amount of colour while adding a lot of it to water... download/file.php?id=189106
It is good to note that at that point I had very rough first versions of rails which I completely remade afterwards, finally giving them the colour they needed so I could see how it's shaping up.
When I got to making snow, the white grass was obviously a problem - the snow would have to be something super specific, blue-ish, but I didn't like any of the things I tested (and I tested a bunch) so I decided to make grass green.
This was an important point because that's where I dramatically changed one of the key points of the original colour concept...
However after that i went fix various issues and improve/redesign some of the bridges, I started train models (didn't finish them though), I made the foundations with frosch's help, I wrote a script which helps me convert 32bpp into 8bpp (that isn't in it's final version yet either, doesn't have full functionality yet), and so on. Basically I was doing things which didn't make me do anything with the terrain for the time being.
Then I bought a new wacom tablet and started to be really obsessed with hand-painted postproduction, and realized that this really is the way to go. Painting over the 3D renders by hand is something that the computer simply can't do and it gives the graphics some sort of resemblance to the the original pixel art from 1994/1998 by Simon Foster. Of course the aim of BRIX isn't to just copy that, but being able to get a similar feeling like when we played this game 20 years ago would be interesting.
Even though my plan was that this hand painted process would not take me that much time, I reduced my activity on this project overall, and I didn't really do anything except painting here and there, quite occasionally in fact.
Then I released 0.0.1 without actually using the postproduction, I started to feel that the amount of tasks I am attempting to make is just insane, and with my current schedule this isn't going anywhere, while the current stage at the time wasn't the best I could image, I wanted people to be able to play with it and see for themselves.
Since then I have been doing some more hand painted postproduction here and there and eventually I came to have touched almost everything I had in spritesheets so far.
When I re-opened my 3D scenes not too long ago, I realized that I seriously hate working with 3DS MAX at this point, I probably just got too used to the comforts of Blender. So I started having ideas again like migrating all of the 3D files into Blender. I migrated the landscape, but the rails are quite a bit more complicated and it would take a long time, not even talking about bridges, that's probably even worse.
So I realized that if I want to release 0.0.2 anytime reasonably soon, I should avoid this. That's why I went even more into hand-painting and mainly - playing with colours of basically everything.
As you can see, trees are finally coloured (although trees are the thing that needs the most work right now, I will get to that later), all of the terrain has some noise.
When playing with BRIX 0.0.1 I really started to hate the ultra-clean terrain. At first it might be interesting, but over time it just becomes empty, boring, and I even find it makes my eyes hurt more. So I added some noise, then I tweaked colours, and I basically got to this... download/file.php?id=199116
However when I look at it all together, I don't really like the colours overall - train signals have much less extreme colours now (as they should), and they are kind of not too visible, everything is quite saturated and it just becomes a mess.
At the same time I was doing a lot of sketching and thinking and I might rework the terrain completely, and starting to give BRIX a specific style (it takes a while to realize what I want to create, I have really needed to search for the identity BRIX should have).
Redoing things is again not very productive in terms of releasing 0.0.2 soon, so I will leave that for later. What I am seriously considering though, is a return to desaturated ground. I mainly started experimenting with this because my After Effects pipeline is quite flexible, allowing me to change the colours of all terrain quite easily.
Here is what I am talking about:
However, I understand that there isn't that many people who will like BRIX just because it's not realistic, and that's fine, the aim of NUTS wasn't maximum amount of players either. But of course it is still nice to see people who actually use the thing. For that reason, at least having the ground green feels more "natural" to most of the remaining people. Still though, I really want to do the desaturated version because I believe that overall it's a more solid plan and it will eventually be much better graphically because the levels of saturation will have some clear sense.
So the current plan is that every new version of BRIX I release will have a different grfID (unless it's just a bugfix release) and you will be able to pick what you want. Even better, since BRIX has parameters which allow you to disable every feature individually, you will be able to use multiple versions of BRIX if you want to change just terrain for example.
I intend to release version 0.1.0 with green grass, and immediately after also 0.2.0 with gray grass.
- (3.87 MiB) Not downloaded yet
I recognize the process, scientists I work with go through the same thing (and working with them, I do it too). They look for that one solution that has all the good things they want, and they jump from experiment to experiment gaining new insights each time until they hit the exactly right combination of elements that does the trick.
The really fun part here (and I hope you will reach it too), is that once you found the solution and you look back at your search, you wonder how on earth you managed to fail in seeing that correct solution for such a long time. In fact, it is considered a tell-tale sign that the solution is really good
Glad you're still working on it, may it give you many new insights in how colour and shape work, and I do really hope you find that ultimate solution!
Today was ... interesting.
I started with a glorious plan of reworking trees, or at least starting it. So I did. I opened the old 3DS MAX scene, exported it and replicated some part of the system in Blender. But then I had to leave for a few hours and I spent a lot of time waiting there, so I took my sketchpad and just did some thinking on paper.
I realized that I might want to completely change the style of the trees, but the longer I sketched, the more I was indecisive because each of the ideas had their advantages.
When I came back home I didn't feel like doing trees anymore simply because I wasn't sure about my decision, so I continued on fixing all kinds of problems with the painted postproduction on tracks.
The tracks are using a very complicated pipeline of taking the rendered sequence from 3D, doing -things-, and spitting out a giant spritesheet of both x4 and x1 zoom.
This pipeline ages since I was making RAWR. At that point it was also for outputting in 4 different modes - 1 for each climate which I don't need now as BRIX has just one climate for landscape.
Over time this thing has grown to absolute abomination in size and complexity, and especially now when I am hacking into it all kinds of things that I want from the Photoshop edits, sometimes things just go really bad. Because I am running quite often at "wtf is this" when working with this system, I decided to go check out what's under the hood.
The following picture shows the sewage I have managed to build up in the pipeline over the years...
https://dev.openttdcoop.org/attachments ... wchart.png
So I started thinking about reworking this. I simply have more experience and I believe I could create a much more robust and maintainable system.
However, building something like this isn't just done in an evening, especially considering that it should do everything that the previous version did. So I will probably rework it, but for later versions, not now. Then I just proceeded to hack my way through the chaos, had to do some compromises and leave out some photoshop layers which will be quite hard to notice that they aren't even there.
The outcome of all this is that I am getting close to finishing maglev tracks for version 0.0.2 so I went ahead and made a big screenshot with signals and things so you could see the current status in a bigger picture. (Literally, the following file has about 125MB)
https://dev.openttdcoop.org/attachments ... test-4.png
Obviously the trees at this stage are properly bad in compare to the temperate climate. I plan doing something about that for 0.0.2.
But other than that, I admit that in screenshots where the land is separate, the gray can get kind of boring. However when I start looking at busy networks like this, the clarity is just outstanding and I still get enough colour from all of the other things in the game.
As I said previously, there will be 0.0.2 as green version and 0.0.3 as gray version, both with different GRFID so you can really choose whatever you like. It's not like everyone builds giant networks.
And then there is one more thing. I loaded my wip version of BRIX as static and went to join our #openttdcoop Public Server, I was just curious to see how it works. I was positively surprised by realizing that the gray land blends excellently with 8bpp graphics. And that's something that is overall a BIG problem of 32bpp - people either want to replace everything with it, or nothing. Mostly because of the inconsistencies. Here I really feel like it's fine. (see attachment)
So, finish tracks, make trees work in all climates, make snow and desert work properly, give a little bit of love to roads, release 0.0.2 and 0.0.3 No more reworks and system redesigns, that will come later, now I really want to just give this to people to play with.
- (1.84 MiB) Not downloaded yet
In the last two weeks I have been silent here and now I will show you why!
Earlier on, I have mentioned that I have problems with 8bpp conversion and I need a special solution. I probably even mentioned that I am going to write a converter in python for myself. Soon after that I did, but a lot of the functionality was clunky and unfinished. Now I finally gathered energy to invest into finishing this converter script, I call RGB DESTROYER.
Basically it's a python script which allows you to automate 8bpp conversion of however many files you want, including many parameters for customizing how the conversion behaves, and with multithreading which is really useful for large files.
Even though I have the script finished, I won't flood you with a wall of text about it yet. The plan is that first I try to use it for BRIX in real scenario, see if there is anything that needs to be added right now, and then I will release the tool for anybody to use, because for anyone who is making 32bpp graphics this tool should help a lot.
As I said in previous post, the idea now is to work on trees. I am doing a lot of experimenting and I don't really know where it will end up.
Here I am posting a comparion of two pictures in both 32bpp and 8bpp. You can also notice that I am working on arctic trees. The plan is to have all trees for all climates in the next version. Another thing, the terrain is basically unrecognizable between 8bpp and 32bpp
I have finished all types, growth stages and snow caps for all arctic trees so what you see here is quite representative of what the upcoming trees could look like. Even better, the trees are made so that it's snow version is always exactly the same except the extra snow, so if you have bulldozed land on which you place a tree and the bulldozed land turns into snow so the trees on it convert, they just add the snow cap without changing the tree. You can of course also reach this case by using a NewGRF which adjusts the snow line over time so it's nice to have this consistent.
I plan to add a little bit more colour variety so that the trees are less similar, although it kind of works. I will mostly just make the early growth stages more lush and bright green colour, while the last stages will be darker and more browny.
Let me know what you think
Onwards to finishing temperate and starting tropic trees. It's an overwhelming amount of effort, an amount that I am really starting to go crazy over it, but slowly it seems to be coming together. Soon(tm)
- (6.3 MiB) Not downloaded yet
The main problem is, temperate trees are madness mainly because there is so many of them - 19 types with 7 growth stages each makes a total of 133 individual sprites, which means 133 individual 3D models I need to create.
To put this into some perspective, arctic trees have 16 types with 7 growth stages each, but the big difference is that it's basically 8 without snow and 8 with snow, where otherwise the tree is (and should be as said in previous post) the same.
Right now I am playing with colours and postproduction to see where I want them to end up, and I am making my mind what do I want to do with tropical trees.
On one hand I would really like to finish all this, on the other hand I really want to make a release as soon as possible, and tropic climate has another 20 types of trees, out of which 2 types are cacti for desert. I have some concept in mind so I will probably at least attempt to make them to make BRIX work in all climates (if I manage to complete tropic, toyland is just another 8 types of trees at which point it's just minor after all the other work)...
- (6.81 MiB) Not downloaded yet
Hi Bob, thank you for your comment, it's greatly appreciated I obviously did think about exactly this to save time, but I am trying to be in the mindset of "if I finish this, I don't need to revisit it again" ... let's see how much bravery and energy do I have but this is the plan.Bob_Mackenzie wrote:I'm always impressed with the sheer imagination and creativity that goes into your work, the trees look excellent. To speed things up you could recycle the same sprite for different tree slots, at least in the short term?
The time it takes isn't really that much of a concern, with the super long time of development already, a few weeks here or there don't really make a big difference. It's just about my mental sanity and not overdoing things.
recently I have been busy as hell with trees for tropic landscape. They are not entirely finished yet, but here is a preview of how they could roughly look like.
Many colours are likely to change and some tree types might disappear completely, but here goes.
The thing that I am questioning the most are the palm trees. On one hand they are kind of showing how overgrown the plants are, on the other hand everything else in the game already shows that. I might replace them with the grass patches instead. To be experimented with...
I feel like it's cute to also add an output from my rendering script. It seems like rendering all trees will take roughly over 5 hours. However I think most of the time is spent by "Synchronizing objects", not raytracing the image. I need to figure out how to cut that synchronizing time down. As you can see in the other picture, it's a giant scene with all the trees.
I might do something really weird like generating sub-scenes with a python script, and the sub-scenes would render things, and see how much the rendering time drops... An interesting test to do at least!
Looking forward to hear your opinions.
Users browsing this forum: No registered users and 8 guests