News:

Do you need help?
Simutrans Wiki Manual can help you to play and extend Simutrans. In 9 languages.

Pak.Sweden....!

Started by Ves, June 01, 2013, 11:40:34 PM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

Ves

Quote from: JunnaThere's a simple reason that sleeper carriages would not be superior to regular carriages, particularly on heavy load services: they have a very low capacity due to their nature. Now, the game doesn't take into account time of day and variable loads, but over very long distances loading will generally be less - it will be interesting to see how with the new distance-based generation this will turn out - and overall sleepers would be preferable on very long few-stop services due to their low capacity and simultaneous high revenue.
.... and very high loadingtime! Excellent, this settles it! I will start painting sleepcars! There have been sleepcars/bed-cars since the 30-generation so this I will start to paint! It should actually be quite easy, its only the windows that really differ on some of the cars.

Iron-Bite, is it still in development?

Otherwise I really like the experimental version because of all the detail one can make that is not possible in the standard version (livery, upgrade, convoyspacing, comfort etcetc..) I think Experimental and Standard are equal as games and I think the experimental fork deserves some more attention. The biggest reason (I think) that there only is one truly experimental pakset, is because all other paksets were created before the experimental fork, but I might be wrong there.

I think actually it will be easier to form it with the experiemental engine in mind, and then adapt it to standard and take what is needed. Because the experiemental has more variables (for instance upgradepaths and liverys). In a Standard edition, there are no way to make a car appear in two ways, nor the possibility to upgrade or set a comfort, and this might affect the choices we make. Then its in my opinion better to make the desicions as detailed as possible, and then take what is needed to make a Standard pak.

In order to make such an idea with 'building' the factory more likely to be accepted by the simutrans community, I think some things are very very important:

* Building time over at least one year = reducing micromanagement, and making it possible to earn money.
* Time between spawning about one year = also reducing micromanagement.
* It should choose the nearest possible resources, and only a few (maybe 1-3 different types)
* If no resources are available (=empty map, or lack of some chains), it should 'cheat' and just have the currently not available resources in stock. If all stock are full by default, just place the industry!
* You code it ;)

Regarding the vehicle load, it would be an awesome feature, and implemented right, would not make it very difficult for players, rather enhance their experience. But I will ask James at some point :)

You definitely should have a look at experimental! ;)

Max-Max

On the Server, under document/color.png, you can find the SJ color scheme.
- My code doesn't have bugs. It develops random features...

Ves

Nice! You find information like carving in butter! :D

I have downloaded Vladkis signs and putted them into the game.
Here is a screenshoot:



Even though the signs are very nice (!!) I think they is maybe little too big? Vladki, is it very difficult for you to resize them, or if you find it too timeconsuming (boring :P ) is it ok if I or someone else are resizing them? What do you all think?

I have finally got to upload some of the sources to the Max-Max-server, so you can check them out there. Only uploaded the trains since I can not create folders so all the dats and pngs would just lie in the same folder and become a mess.....

Im starting to pinpoint out which sleepcars should be included. I think some newbuilt is allways needed, but they should be upgradeable too but not too many cars are needed! :)
thinking, thinking...........

Ves

Im sitting and looking for sleepcars. Well I found some and realized that some cars I already painted are capable of having beds.

But the question is:
Most of the 'real' old sleepcars have the littera BC-something, which means they are a combo of 1 and 2 or 1 and 3 class. I have deliberated avoided any car having two classes, since this makes it complicated to program since the simutranscars only can hold one cargo at a time. One way to do it is to make two cars out of it, one containing the graphic and one of the classes, one containing just space and the other class, much like the pak.britain-Ex boats are dooing right now. It is a workaround and I dont like it very much since it complicates the depot proces, making it very difficult to get an overview. I have actually already done this with some of the smal 'Spårvagnar' (no idea to phrase that word in english, but the smallest train in my set of trains :P ) where there are both people and mail holds.

What do you think?

Vladki

Quote from: Ves on March 01, 2014, 08:41:28 PM
Even though the signs are very nice (!!) I think they is maybe little too big? Vladki, is it very difficult for you to resize them, or if you find it too timeconsuming (boring :P ) is it ok if I or someone else are resizing them? What do you all think?

It is easy to reduce the height of the pole. However making the signs smaller would be equal to painting them again. I may do them if you agree on the final size. I can prepare one sign in different sizes as a sample. However I do not mind if somebode else does the modifications. I was planign to prepare blue highway sign and variants with yellow/red pole, and some signs for sharp turns, but I'll wait until the size is agreed.

Just a note to the sleepcars. They should have - low capacity (30-50), high loading time, and high comfort to allow for at least 8 hours ride. (somewhere around 200 and more). Anyway it might be quite tricky to get them balanced right.

The convertible Bc cars (at least the czech version) can have either 6 beds per compartment (and one compartment reserved for service), or 8 sitting places. You can either ignore it and just use it as 6 per compartment, or make two versions - Bc-sleeping and Bc-sitting, upgradable in depot at no cost.

Ves

Yeah i realize that there are gonna be much upgrading in the depot from the post I posted about that proposal :-P yes, I thought also that there could be two versions of those cars which are both daytime cars and verses, it was just the dilemma that half of the BC cars is 1class and the other half is 2class. Those require two different comfort ratings, but which is not possible yet. Some of those cars in the beginning of the century had the 'massive' capacity of 10 sleeping travelers. Of course this is on first class. That car is gonna have a high comfort rating...
Most of the coupe-cars at the time could be converted between bed and sitting

Ves

Sorry, my last post was maybe a bit unclear.. have to stop posting in the middle of the night! :P

I have looked even deeper in to which sleeping vehicles could be incorporated in the pak and im getting closer to a conclusion i think!
But we have to decide how to handle when 1'st and 2'nd and 3'rd class is involved in the same car, since one vehicle cannot hold two different cargos.

One solution is to make one car out of two vehicles (as I wrote in earlier post) one vehicle containing the graphics and loading first class and one without any graphics but loading second class. These two vehicles are forced together, never to be separated with constraints.
An other way is to imagine the entire car as a first class car, with the seat numbers it would have if it infact where first class (requiring some guesswork here....) and add upp to with overcrowding to the cars factual seatlimit. This should drag down the comfort rating when the car is getting overcrowded and (hopefully) make it level in comfort with a second class or third class car. What do you think?

Some of these cars are also not dedicated sleepcas, but cars with coupés where the sofa could be made into a bed, requiring as vladki suggest, two different vehicles to represent one.


Repainted the 40-generationcars with the colorscheme from Max-Max.

Maybe little too much in the red corner...? :)

Junna

I believe the X6 EMU's for suburban traffic were of a similar red (as well as white) livery, and I believe also that the 100-seater suburban loco-hauled stock had a similar red livery, so it might be that they had such a bright red colour.

Max-Max

I have seen this read color on some early Litt G. SJ seemed to have quite bright colors in the early 1900.

I made a concatenator pole in SketchUp from various notes, reports and half blueprints :P I'm not sure from what period this is, but I believe it is the first half of 1900.

- My code doesn't have bugs. It develops random features...

Ves

Some of the pictures of the real life cars from that generation, shows a 'very' red car, but the paint tend to get more brown/white during the ages.
We'll I'll leave it there for the moment. :)

Maxmax your details are wonderful! Have you tried to render it for the game? Just use my catenaries as a template.
What about the car you made earlier?

It's a shame now you made all those details, that the cannot be visible in the game, but maybe you are hoping for the simutrans 3D ;)

Max-Max

#115
Quote from: Ves on March 04, 2014, 01:34:43 PM
Maxmax your details are wonderful! Have you tried to render it for the game? Just use my catenaries as a template.
What about the car you made earlier?
When I saw all those drawings I felt the urge to do them in SketchUp. I still have to figure out how to best transfer them to Blender for some proper rendering.

Quote from: Ves on March 04, 2014, 01:34:43 PM
It's a shame now you made all those details, that the cannot be visible in the game, but maybe you are hoping for the simutrans 3D ;)
I put in as many details as possible so we can scale them up and down as we like and if there ever will be a Simutrans 3D or another 3D train simulator. As it looks now I'm thinking of writing a Blender Export filter to SketchUp or a SketchUp Import filter in Blender.

With all these details I start to think if we should go for a 192 or even a 256 pak instead? :P

***EDIT***
Seems like I managed to import from SketchUp without to much hussle. I just need to plan a bit when I design the objects.

How ever, I need some fundamental info.

1) Are we designing for Experimental or Standard?

2) Is this going to be a PAK 128, 192 or 256?

3) Standard has implemented the half slope patch, do we know if this will be implemented in Experimental at all and should be design the half slope version?
- My code doesn't have bugs. It develops random features...

Ves

As far as I'm concern, experimental will adopt the new half height and climate zones. Please correct me if I'm wrong!
I thought that we could make this for experimental, to get all these functions. If you can agree on that (and no one else is making some really good arguments against it) I think we should head for that!

I must admit that making a pak 256 would be awesome. But then my works would be lost. However I think about it, it might be possible to just double up the size on my paintings. Then they would look exactly as now, just twice the size and I would have to add all the details not yet visible....
Is it at all possible to compile a pak that size?
At least I don't think that the packing program I'm using can do it... How do you pak files from sources?

Junna

PAK256 looks terrible (as does 192 in my opinion) due to the inflexibility of the tiles making large things look cumbersome and rigid.The oddity of the excessive width of ways and tiles becomes even more obvious at such a large scale, too.

Max-Max

If things looks odd or not in 192 or 256 has nothing to do with tile size. No one is forcing us to use the whole tile as a road, we can still define and use proportions that looks right, just bigger end more detailed.

I think the truth is that a smaller pak size is really out of proportions, having different scales for vehicles and buildings. I think a 256 pak will be more close to some more realistic overall proportions. In the end we can decide these proportions ourself...
- My code doesn't have bugs. It develops random features...

Ves

I just downladed the 192 pak comic to try a bigger tilesize, and I got a little deja vu from when I first time tried pak 128 after playing pak 64. You could really zoom in to the details! Also I noted that there might come some sort of limit somewhere to this kind of pixel-game that Simutrans is. The graphic is getting inconsistent, eg shadows on other objects are missing, vehicle cornerings etc. Simply, it risks to looks like a picture moving around on the screen, instead of an actual train traveling on the rails. Because you expect to see even more details like the mechanics on the old steam engines moving, more rotations between the directions. The game cannot fool the human eye too much. Junna, isnt this what you mean?

I dont know where that limit would come though.

I tried, to rescale some cars, and i must say, that if its gonna be a bigger size than 128, then it must be 256, since 192 would be equal to scale it with 1.5, making the rescaling very very difficult.
Any way its a big big step to rescale everything, so there really should be something to gain and some very good solutions presented to make the graphic inconsistence not so present.

But for the moment, I must say my vote go for the 128 size. Maxmax, your wonderful models, you could still save for the sake of a simutrans 3D ;)

Any thoughts on this?

vorlon

Hey everyone, and sorry I've not been able to participate lately. I have still read every message that has been posted and the discussion has been very interesting. I have my high school final exams going on and they will keep me occupied for a couple of weeks more.

I just wanted to briefly comment the issue of different tilesizes. I totally agree with Max-Max with regards to the roads reserving the whole tile. The biggest problem I have with simutrans graphics in general is that cities tend to be more pavement than buildings due to the disproportionate scaling of roads, and narrower roads would open some very interesting possibilities with sidewalks, special road markings in intersection etc. But I understand that for pixel-artists bigger tiles mean a lot more work, and the current graphics being all 128-sized I think the current size is the way to go. I would have no problem with painting bigger graphics though, if that is what we end up with.

Ves

#121
Quote from: vorlon on March 06, 2014, 10:14:04 AM
Hey everyone, and sorry I've not been able to participate lately. I have still read every message that has been posted and the discussion has been very interesting. I have my high school final exams going on and they will keep me occupied for a couple of weeks more.

I just wanted to briefly comment the issue of different tilesizes. I totally agree with Max-Max with regards to the roads reserving the whole tile. The biggest problem I have with simutrans graphics in general is that cities tend to be more pavement than buildings due to the disproportionate scaling of roads, and narrower roads would open some very interesting possibilities with sidewalks, special road markings in intersection etc. But I understand that for pixel-artists bigger tiles mean a lot more work, and the current graphics being all 128-sized I think the current size is the way to go. I would have no problem with painting bigger graphics though, if that is what we end up with.

real life has to go first ;)

I personally would not mind to convert my objects to another tilesize or to paint new stuff for a bigger tilesize. I like painting!
My concern is what Junna pointed out, the limitations of this kind of game, which might get overexagerated when using a big tilesize.
Some of the limitations that Im refering to:

* No antialliasing / blending between objects and the ground.
* Only 8 views for vehicles. Makes the cornering a 'jump' between different views. If there was maybe 16 or 32 views, then a smooth curve could be made, but I wont be drawing 32 views of the same car. Ever! ;-)
* No 'climb'-textures for vehicles. The default south/north, east/west images are shown when vehicles are climbing a hill.
* The ways also have only 8 directions, no smooth transition in corners and the 3- or 4 directions rail texture would be difficult to get realistic since all rail vehicles can turn either way.

I think these issues are easier to forgive when using a smaller tilesize so you "can not zoom in on all the uglineses". Big tilesize invites for more details and more expectations of some of these things.

I think that the comic pak.192 is dealing with these issues as to not aim for a 'realistic look', but a comic look (NOO.....!? ;) ). A cartoonish look which does not invite the player to expect the world to be realistic in the way which for instance pak britain does.


Look at it like the wodden BRIO-trains: They have a very typical design and a relative small size. The design and convention of the briotrain is the Simutrans engine, and whatever you paint on the BRIO-car is our pak. As long as the wooden train-car has that small size, its easy with some paint to convince yourself that 'this is a RC-loco'. Your imagination fills in the gap.
If you instead made the brio train-car huge, 1 m wide, I assume you would have some problems accepting that 'this as a very good representation of a RC-loco' no matter how much details you have painted on it. It would still contain the big plastic wheels, the track would still be what it is, and it would still be a flatsided cube.
The best solution to this would be to go over to Märklin or some other model-train hobby to get all the juicy details, but as we are not able to just switch the mechanics of Simutrans, we have to go with the BRIO-train.

(this is not meant to be any offence to the Simutras engine or its designers, Im just exagerating a little to describe my point ;) )


So, the solution to this 'equation', I think will be:
The more realistic, "serious", that we want the pak to be, the smaller it has to be.
The more comic, cartoonish, "obviously not realistic", we want the pak, the more we can afford to go up in tilesize, making the player not expect a perfect world.


Does this sound reasonable? :)
What do you think?

And finally, which pak should we aim for?
Im voting for 128 ;)

Junna

Quote from: Max-Max on March 05, 2014, 05:02:25 PM
If things looks odd or not in 192 or 256 has nothing to do with tile size.

It does. Ponder this: at a bend, because the game only has 45 degree turns, parts of a vehicle will always overlap with another tile. How much increases with the length of the vehicle, but the appearance of this will be further exacerbated by the tile dimension; the larger the tile, the more jagged the look of the movement will become. Consequently, in pak.192 it looks a lot worse than in pak.96, and pak64 generally looks best (as regards the annoying nature of the overlap). The vehicle proportions in this pak are already quite large, so would look absolutely ridiculous if it was 256.

Max-Max

Well, my point was that just because it is bigger than 128 doesn't mean it automatically looks bad. I'm all in for 128 if that is what the majority wants ;)

I have been experimenting with Blender to generate images for 128 and one of the biggest problems are when details are so thin that they all are antialiased. To remove the blueish background border around each object (due to anti aliasing) I'm masking out all anti aliased pixels, meaning small details will completely disappear.

I'm experimenting in the Compositor to somehow "save" these "weak" pixels.

But beside that I have managed to import from SketchUp and have it scaled correctly to 24m/tile :)
- My code doesn't have bugs. It develops random features...

Sarlock

I have spent an enormous time thinking about varying pak sizes and the challenges of trying to create a visually appealing final product in Simutrans.

I came to the rapid conclusion that bigger is not better, for two reasons:
1) The player spends a lot of his/her time zoomed out, so the detail is largely lost.
2) When zoomed out, you not only lose the fine detail that you worked so hard to create but you are also losing quality due to the algorithm used to zoom the artwork out.  A pak128 object will not look as nice as a similarly designed pak64 object when the pak128 object is zoomed out to 2:1. (the pak192.comic can escape this because of its bold outline and solid colouring style, it looks very nice with a large image size)
3) Larger pak sizes require exponentially more work to create the art and take up more memory

I have done a lot of pak128 objects with Blender, several published items and volumes of unpublished work (either unfinished or test objects).  If you put the PNG created by Blender through GIMP's Threshold Alpha remover, you can quickly strip the "haze" that you get from the alpha blending on your finished object and create a really nice end product.

Detail does look better, in my view, but only add detail items that will show up in the end render at the size you are aiming for.

I present an aerial photograph:



This is taken in a similar angle to our isometric view.  Look at the detail!  And yet, this barely pak64 size.

Take a close look at the building on the right side:



This building is 64 pixels by 59 pixels, including the sidewalks!

A few things I note:

1) Consistent lighting.  Very important.  If you want your pakset to look good your lighting levels have to be similar.
2) Very few shadows.  The few shadows that are there play little part in making the image believable.
3) Despite the zoomed in appearance being a complete mess of pixels, it looks amazing when zoomed out.  Even zoomed to 50% and 25%, the detail is still incredible.


The human eye has an amazing ability to pattern match and decide what these objects are even though the actual pixel representation is somewhat random in appearance.

Look at a few cars in the parking lot:



You can barely even tell what this is.  Yet, when you see it in the image, you can see that it's cars in a parking lot immediately.  Why?  Because your eyes/brain can instantly infer from the surrounding objects what that must be (primarily from the painted lines)... and because it decides it's a parking lot, the cars look exactly like what your brain has decided they are... even though they are just blobs of black and grey pixels.  In the zoomed picture it's hard to determine what you're looking at...yet in the regular zoom picture you know instantly that you're looking at 5 cars side by side.

I am really starting to edge towards a pak64 or pak96 size as being optimal for Simutrans, if the graphics are done right.  One limitation we do have, however, is that we lack an alpha channel blending between objects, so edges will be a bit more sharp and won't blend smoothly together.  This is especially limiting with trees and other foliage.
Current projects: Pak128 Trees, blender graphics

Max-Max

That truly looks amazing and it is a good analyse. However, highres monitors are being more and more used. I tried to play Pak64 on my 2560x1440 monitor and find it far to little. Zooming in only makes it worse and even the Font text is a little bit to small. The tool buttons and dialog elements are also in the smallest (hence why I tried to make the theme update to Simutrans Standard).

For a highres monitor, such as mine, 128 is about the minimum size, not smaller. I also propose that we make the tool buttons larger than the standard.

To remove the anti aliasing in Gimp might be an alternative, as long it can be automated from a command file.
- My code doesn't have bugs. It develops random features...

Ves

Very interesting reasearch and thankyou very much for sharing it! :)

Quote1) The player spends a lot of his/her time zoomed out, so the detail is largely lost.
Im actually playing mostly zoomed in, so I can see what Im doing ;) But you are right, that most players maybe play zoomed out most of the time.


Quote1) Consistent lighting.  Very important.  If you want your pakset to look good your lighting levels have to be similar.
This is very true! Im waiting for how the renderings from Max-Max will look like, and adapt my work towards it. We where talking earlier about having a more special northern lighting, due to our contries are more northern. Maxmax would do some research on this I think?
Quote2) Very few shadows.  The few shadows that are there play little part in making the image believable.
What do you mean? I think it looks like there are very much shadows on the picture you present..? Maybe even little too much, because you barely can see what is on the side turning away from the sun, but this is maybe not what you mean? Anyway, this should also be coordinated between all objects together with the ligting.

Very interesting with the car-parking"blob". It really is the brain making it a car-parking!
So, if you were about to start up a new pakset, you would use 64 or 96 tilesize?

Sarlock

I'd probably use 64, if only because then you can borrow a lot of the components from other paksets to get your set up and playable fairly quickly.  Then you can upgrade those pieces one by one as time allows.  If you have the time and people to help, pak96 might be a nicer compromise.  96 is also very nicely divisible in to 2, 3, 4, 6 and 8, which is a useful benefit... though if rendering the graphics rather than hand drawing them it's not as important (depends on whether you're going for a nice hand-drawn, clean look or a purely rendered photorealistic style).

pak128 uses a sun angle of 60 degrees which roughly reflects the sun position at summer solstice roughly around northern Germany.  The downside of choosing a steeper sun angle is that the shadows become longer and while they look good on the tile itself it's hard to project shadows beyond that tile that look half decent (especially when zoomed out a bit)... especially with trees (pak128 uses the every second pixel is black method which looks alright but goes all black when zoomed out).  pak128.Britain does away with shadows altogether (as on a very cloudy day, which seems apt for a British pakset :) ).

What I meant about the shadows in the aerial photograph I sent is that they don't play any significant factor in you needing the shadow information to determine what is on the picture.  The shadows are mostly lying on the road behind the buildings and aren't that important.  If you removed the shadows you would hardly notice they are gone.



Once you come up with a few renders you like, you should probably do like pak128.Britain does and establish a common lighting setup that can be used by everyone to maintain consistency.  The intensity of your shadows will be an important component on the lighting setup (whether to use a stronger environmental lighting and less bright sun or vice versa).
Current projects: Pak128 Trees, blender graphics

Vladki

Here is an attempt on smaller roadsigns. Actually I just used the saize used for pak64, which is actually quite exactly the right scale - height is cca half of the railway electrification poles, and with a bit less that railway gauge. For comparison I put also pakBritains one way sign (back-to back with mine) and bus stop with a bus (not double decker). Big sign is 10px wide, 32px high. Small sign is 5px wide, 16 px high.

I think about making something in between - 7px wide, 24px high

Max-Max

Those smaller signs looks just about perfect  :thumbsup:

I know you guys are waiting for my document and I'm kind'a working on it (not so much written yet), but I have written about tile size, shadows and the sun.

I have also done some major research about the train history in Sweden and collected a lot of drawings and blueprints (on the server).
I have also contacted The model railway society here in Gothenburg and the Train museum in Gävle (where a lot of the original blueprints are stored).

It seems like I have to pay some sort of fee to get copies of the blueprints, so before I go ahead, I think we should make a road map of what engines and wagons we intend to include. Then I can focus on collecting information on things we actually will include :)

I have borrowed "Svenska Ellok" from the library and I'm scanning some material, but everything isn't complete.
- My code doesn't have bugs. It develops random features...

Vladki

Quote from: Max-Max on March 09, 2014, 05:47:17 PM
Those smaller signs looks just about perfect  :thumbsup:

Just to make sure - do you mean the small red-yellow "no entry" sign? The others - blueish are bus stops from pakBritain

Max-Max

Quote from: Vladki on March 10, 2014, 07:30:39 AM
Just to make sure - do you mean the small red-yellow "no entry" sign? The others - blueish are bus stops from pakBritain
I mean the smallest "no entry" sign. Have you tried to make some more signs in that scale?
To me it looks like the pole is a little bit to high (compared to the bus on the road).
- My code doesn't have bugs. It develops random features...

Max-Max

I have been holding on to this "design document" far too long now. This is just the very beginning of the design document, but at least we can discuss what's in there now.

You can find it on the server "Pak128Nordic_DesignDocument.pdf"
- My code doesn't have bugs. It develops random features...

Ves

I tried to paint a shadow using the technique from the threes you described on a train car. It did not look good at all..... Made a screenshot, but it's on my computer back home at the moment. So until the simutrans engine is making up a better way to handle shadows, I think we should avoid them for the moment.

Good point about using 96 instead as 64 (or 128). I have not tried the 96 pak (there is only one?) but I think I will agree with Max that 128 is a good size, not making it too small on the screen. 

Ok, so you mean shadows should almost only be on the object itself, not cast shadows on whatever is behind?

@vladki, that small no entry sign looks very nice! Good work! :-) how are the experiments on the different sizes going?

@maxmax, you really dig like a machine! Impressive! :-)
Ironically, now I can not go in to the server and check your document, I'm not home.....

By the way, I have uploaded the vehicles on the server along with ways and crossings. There is a document in the rail-folder explaining how to operate the GIMP-file, since it might be a bit confusing with lots of layers.

greenling

Hello All
The New Photos looks very nice out.
On what do the devloper in moment work?
Can i on thuesday post some new ideas for those Pakset?
Opening hours 20:00 - 23:00
(In Night from friday on saturday and saturday on sunday it possibly that i be keep longer in Forum.)
I am The Assistant from Pakfilearcheologist!
Working on a big Problem!

Ves

Hello greenling! Thanks for the comments :-)

Currently, I'm in the process of painting sleep cars and widening some trains (x2000 for instance) and some other stuff...
Vladki is painting signs and maxmax is gathering material and rendering.

You are most welcome to post ideas to the pakset! :-)

By the way, maxmax, I had made an excell sheet with the vehicles I have been thinking to include. It contains only the locos (gods/pax) and all passenger. I have not included any freight cars in it. It also contain information like weight, load etcetera.  I will upload it when I get home. Then we can discuss and update the file.

Ves

#136
Here the screenshoot with the awfull ground-shadow:

The method I have used is every second pixel black. In some way, I think then it is difficult to continue with the idea that we should have a lower sun in our pak. Any suggestions how to solve this?
Notice also that I have painted the X2000 a bit 'thicker' in order to make it match the other vehicles and also added wheels to it. I Think it looks a bit better now :)

Nice document! A bit empty at the moment, but we will fill it once we get through the work. God to have some ambitions and references in the document, so one is not getting lost in just creating pak-stuff...

Somethings that could be added:

Timespan We could ambition to have a limitted timeline, at least to start with. My suggestion is from arround 1930 to present time (this is the timespan my vehicles range). Any suggestions?
Train compagnies (in Sweden, and you could write down the compagnies in Denmark,Finland and Norway as well if you want)
Namingconvention: Big pak = many files = A MESS! We need to name stuff properly in order to not get lost. My suggestion: Countrycode_Name_Decade. This yealds for instance the pakfile: vehicle.Se_Du_1950.pak, or way_Se_50_road.pak. One has to know the name (real name) of the object in order to find it, but with this it will be easier to navigate the day pak.nordic includes more than Sweden. The reason to include the year, is that cars gets upgraded, so its called the same, but just another decade after. Also this helps one identifying the objects.


I have my excell sheet with the trains I think should be included. Im currently working on it so it will be able to be readable by someone else than me :P
writing,writing,writing...............

Vladki

Well, I didn't have much time to play around with the signs. Those tiny signs are just signs from pak64, with only minor modifications - color and position. I have a complete set for pak64: and . However thay are so tiny, that it is almost impossible to paint anything on them. And I'm afriad, they'll be almost invisible when zoomed out. I personally play pak128 mostly zoomed out so that it looks like pak64. I zoom in, and unhide trees and buildings only for amusement. However here is an attempt on signs 6 and 8 px wide - antialiased and not. Antialiased look quite good, bud the question is - what backgroud should be used for antialiasing? In the game it can be anything - road, grass, some building...

Ves

#138
Sorry I have been away for a time.. have a lot of other work at the moment..  :)

hmm that is not good if you can not see the sign. How is the little bigger sign looking in the game?
I think the signs themself should be antialized, but not the edges of the signs. A sharp border to a smooth (and good looking  8) ) sign.

I have tried to acces the server, is it down at the moment? I have a vehiclelist I have worked on and Im sharing it on my dropboxfolder under 'documents'. its definitely not an absolute list, just a incomplete list of vehicles I could think of that would benefit the game. At the moment there are no freight cars (no true freightcars at least), no diesel/steam-engines and there are almost only SJ stuf in it.
The list is a bit messy as I have tried to fill it with as much information as possible. You can sort the information by date to get a timeline (default is by littera). The information in the file is unfortunately not 100% reliable (especially some tractive effort entries) other infos as weight and capacity and KW should be mostly accurate.

Please think about what vehicles more to include! :)

The sleepcars / cars with sleepingabilities I think should be included are: (cars with "W" in the name have slightly more comfort)

ACo4
Ao5
BC2
BC4
WLABo4
WL1
WL11
WL2
WL4
WL6

The "BC2" is an old 1940-generation car. In the 70'ties SJ had many different cars from that 1940 generation rebuilt into "BC2", and they are still using them today. There are currently 5 different cars included that could possibly be converted to BC2, meaning that I have to paint all these 5 cars in all blue and black variations. 5 cars * (3 blue variations + 1 black variation) = 20 new cars to paint!  :o

Also, I have put up some fresher compiled vehicles on the dropbox.

Ves

I have now painted the 30- and 40-generation cars with sleepcars. There are some very minor differensies in some of the variation like the roof has been made a little taller (in order to make room for more beds)
Not all variations are visible on the screenshoots, but all the paintjobs should be visible.
I also adjusted the color of the red and brown. Is it too dark?

Anyway here are the picture:



Are you guys working on something and if, what are you doing? :)

Im planning to make the 60- and 80- generation sleepcars next....