Southern Railways "Schools" V class [WIP]

[Gumboots]

Hey Bomber.

I'll behave myself now, and go back to some of the other unfinished projects.

[bombardiere]

Oh dear. You are so quick. I did not even open RRT3 or 3D software whole weekend. One of my workmates was leaving, so we had an afterwork party. First too much fun. and then next morning not so much fun.

There's no way that green is Bulleid malachite. It's definitely an olive of some sort. If anything, the lettering may have yellowed a bit with age

However I am still convinced that the picture is for Bulleid Malachite. My arguments are the lettering and position of the loco number. Both are Bulleid style.

See the lettering on the tender. It is different lettering than Maunsell was using. That look like Bulleid lettering type.

Here is a comparison.

Typical Maunsell

https://fox-transfers.co.uk/transfers/s ... -lettering

And pre-war Bulleid

https://fox-transfers.co.uk/transfers/s ... -1936-1941

Also the loco number is on the cab. Maunsell had it on the tender. That is why I believe that the photo is Malachite Green engine. Even though the colour is so dark. So I think the dark green look is either a colour distortion or this is a re-coloured photo.

[Gumboots]

You may be right about the malachite, but I'm still not sure. I have seen plenty of shots of models of the Schools class that have tender lettering in yellow, while still being painted in some shade of Maunsell olive, so some people obviously think there is some basis for that combination. Anyway, it doesn't matter much. Anyone who particularly wants something will be able to grab the PSD and tweak it to suit their liking, or even just change the lettering directly on the DDS.

I'm currently thinking I should maybe try and clear the locomotive backlog a bit, by doing what I've done with the Schools: finish sorting the mesh out, get them running, and leave final UV mapping and detailed skinning for later. It may provide a bit of an incentive to do further work on them, and I could probably churn out a bunch of them to that basic stage pretty quickly. There are at least a dozen sitting around that are close to usable (Gresley A1/A3, Peppercorn A1, Gresley V2, Gresley P2, Black 5, DX Goods, Royal Hudson revamp, a couple of Garratts, etc, etc, etc).

[bombardiere]

I'm currently thinking I should maybe try and clear the locomotive backlog a bit, by doing what I've done with the Schools: finish sorting the mesh out, get them running, and leave final UV mapping and detailed skinning for later. It may provide a bit of an incentive to do further work on them, and I could probably churn out a bunch of them to that basic stage pretty quickly. There are at least a dozen sitting around that are close to usable (Gresley A1/A3, Peppercorn A1, Gresley V2, Gresley P2, Black 5, DX Goods, Royal Hudson revamp, a couple of Garratts, etc, etc, etc).

You get a plus from me for that one. I love to play with those even if are bit unfinished. I can happily play scenario even if that loco is bit raw and anyway in comparison to my models, even your raw version is a huge improvement.

Your work has inspired me to play again. It is fun to play old scenarios with new engines. Thank you.

You may be right about the malachite, but I'm still not sure. I have seen plenty of shots of models of the Schools class that have tender lettering in yellow, while still being painted in some shade of Maunsell olive, so some people obviously think there is some basis for that combination. Anyway, it doesn't matter much. Anyone who particularly wants something will be able to grab the PSD and tweak it to suit their liking, or even just change the lettering directly on the DDS.

Oh. All Southern passenger and loco lettering are Yellow. Both Maunsell and Bulleid. I am not sure about the exact shade and they may have different shades. Sorry, I was referring to letter font. That tender has Bulleid type of font. That is why I believe it being Malachite. It would very fit to caption "New light green livery." But of course it can be a test mix and match livery.

Anyway, you are right and it doesn't matter much. I hope you understand that I am discussing this in academic curiosity. If I want some special livery or combination, I can do that myself. You take your view on what you see is fitting and I will be happy with it.

[Gumboots]

Oh. All Southern passenger and loco lettering are Yellow. Both Maunsell and Bulleid.

I've seen models with white lettering too. But those are only with Maunsell olive, not with malachite. I don't know where the model builder got white lettering from, but I assume they had some basis for it. Maybe Maunsell experimented a bit too. I know that immediately after grouping the SR pax livery was the old LSWR sage green, and Maunsell didn't bring in his olive immediately. They could have tried all sorts of things I suppose.

And yeah, academic interest. It's kinda nice to have a couple of liveries that are accurate, if we can get them, just for a sense of history, but I'm not going to go nuts about it.

[Gumboots]

Ran some quick tests at the weekend with skinning this thing. Since I had the other skins roughed out I figured why not chuck them on the running beta and see how they look. Might as well.

The dgVoodoo graphics fix has made a big difference to what is going to be possible. Back when I first tested an exported model (first one ever for Blender to RT3) the darker green was so dark it was basically unusable. With dgVoodoo handling the game's graphics now, the same dark livery comes alive and looks like it should. It's now not just usable, but is one of the best.

This got me testing the other ones, and they have promise too. The dark purple-red St. Trin's was originally suffering from the same problem, so I lightened it up some. Looking at it now, I could go back to the darker version and still have it work well.

The other one which is now looking dramatically better is the LB&SCR Stroudley livery, which totally went to crap before dgVoodoo. Still needs some tweaking, but the boiler striping is now perfect and the gradients on the boiler seem pretty good too.

I seem to be going through one of my more motivated spells at the moment, so will see if I can get the running beta upgraded to a better standard. I'm coming to the conclusion that it's better to get things running even if they aren't perfect, as long as the mesh and UV mapping is set up so that tweaking the texture to a better standard later is possible. That way people can play with them and have some fun, and if anyone (like me) gets inspired to do a few improvements they can be done without it being a major drama from scratch.

[Gumboots]

Allllllllllllllllllllllllllllllllllll righty then.

I have had another crack at this mongrel of a thing. I was determined to knock it off and not be defeated. It's now at the stage where, for the first time ever, all of it fits on the texture, in a rational arrangement, at the correct scales and everything. This is even with me deciding to get fancy and have the correct counterweights on the driving and coupled wheels, which meant chewing up more space for wheels (coz they're different).

It has proper vector/alpha work in most places it needs it. Still needs a little bit of tweaking for UV packing, but that's just to make some details of skinning it a bit easier. Obviously the skins still need some detailing too. The bottom line is it will all work. Not only that, but it is going to come in at a lower poly count than the rough beta.

I'm kinda chuffed about it. Beats me why I didn't just do it this way to start with.

The main brainwave I had was to do non-critical components at 20px/RT3 unit, which is the same scale as the Pennsy H3 skin, and decent enough for a lot of things, but to do more critical components at 24px/RT3 unit to get better clarity on some of this beastie's horrendous fancy detailing. Which it has lots of. Complicated little critters, the Schools class. I seem to remember some beginner saying he was going to do one because they were nice and simple.

[Gumboots]

Made some more progress. UV mapping has been rationalised (more or less repacked the whole thing) and some extra skinning details have been added. Cab window is currently gruesome, but the lining is turning out quite well. I'm coming to the conclusion that it's almost not worth working in raster at all, and it might be better to do everything in vector format.

Anyway the whole thing is coming along nicely. I feel like I'm on the home stretch now.

[Gumboots]

Made a serious start on the tender skinning. It's looking a lot better already.

A bit more ratonalising has been done on the locomotive too. Saved a few more polys without making it look rougher. This thing is forcing me to think outside my usual box, which is good. I'm learning more all the time. I'll probably cut the poly count a little more by the time it's done. It's looking like it will end up around 100 lower than the existing beta, somewhere in the low 1800's. Anyway, things seem to be going smoothly now.

[Gumboots]

Just getting this thing finalised for mesh, and a slight problem occurred to me.

UK locos of the period didn't have headlights as such. The running lamps they used at night were only to indicate to trackside crew what sort of train was coming. They didn't provide any headlight beam for the driver, so if I do put a running lamp on the front of this thing it would only have a small glow light anyway.

The catch is that the running lamps weren't used between sunrise and sunset. That's when the head code (as it was called) was done by those white discs you see on old UK locos. These replaced the running lamps during daylight hours. This means that for RT3 purposes it can't be correct all the time. It has to have either the daylight head code (white discs) or the night head code (running lamps). It can't have both, so I have to choose between them. Well ok, it could have both, but then it would be wrong all the time instead of wrong some of the time, which doesn't seem to be an improvement.

At the moment I'm thinking it should have the white discs. There are more daylight hours than night hours, and the discs can be done with a much lower poly count, and all RT3 locos have lights at the front so using the disc head codes would be something different. If anyone has a different opinion, speak now or forever hold your codpiece.

[RulerofRails]

Good call. IMO, generally looks at day have higher preference than at night. I have my nights on the brightest possible setting. It's nice when doing a sandbox. But it also feels wrong to do some industrial scouting or a large track build in the middle of the night.

Hope all goes well for finishing this one up. It's looking better and better in the shots.

[Gumboots]

Y'know what I could do, just for the heck of it, is add a couple of files for small lights but without bothering with the mesh for running lamps. So it'd have the white discs all the time, but could have a couple of "lamps" up front at night too. If I use a small glow radius and set them forward a bit, they wouldn't illuminate the white discs but would still give the general effect. At least if you weren't looking too closely and weren't too persnickety. Can always try it with and without, since a couple of light files are a no-brainer and don't affect anything else.

Anyway yes, it's going well. I've done more repacking of course, since I thought of a couple more details I'd missed earlier and it might as well have. It's now sorted and is sitting on 1,798 tris. That's 121 less than the existing beta, but with better detailing. I've decided to call it done at that point, since trying to go further would be bonkers even for me, but it was a good exercise in figuring out what could be cut without making things look worse.

Only things left to do now are sort out enough eye candy to make the thing satisfying, then pack it up. I expect that will take another week, what with one thing and another, but the end is in sight.

[bombardiere]

You are tempting me. I have been bitten by Train Fever, but when you get this one ready, I need to fire up good old RailRoad Tycoon again.

Funny thing that you say about 1798 tris. I have checked Transport Fever locomotives and models and even the simple ones start with 5000 plus tris. And some modders do not even provide LODs. Times has changed a lot since RTT3 was introduced. And I am not talking about my greying hair.

[Gumboots]

I don't think RT3 would handle 5,000 tris on locos. It would be easy enough to test it and see what happened, but I think the frame rate hit would be too much. We were noticing some difference even with 2,000 tris, although it wasn't a huge difference. I do have a 4,000 tri model sitting in Blender. I could get that running as a test model, just for the heck of it. Although frankly I'm not sure there would be any point given that I can do just about any locomotive to a pretty good standard without going over 2,000 tris.

The real restriction in RT3 is the 1024x1024 limit on textures. Even if it allowed 2048x1024 it would allow heaps more scope for nifty things. I'd love to hack the .exe to allow that. Not only would it allow using fewer images overall, it would even allow cutting mesh without sacrificing visual quality, because the extra space would allow more scope for using alpha instead of mesh. So, weird as it may seem, allowing bigger textures could (providing you didn't go nuts) allow a performance advantage.

Anyway the Schools is now down to 1,780 tris. After saying I wouldn't take it further I went and took it further anyway. I sort of got on a roll and started seeing more opportunities to cut the tris, to the point where I then got into adding back in some extra detail for missing bits that weren't there before. For example, the firebox door on these things is a shallow cone, and although just using a flat plane with graphics worked well on the Pennsy H3 it didn't work so well for the Schools. The Pennsy has the large red numberplate (separate mesh) and a large headlight, which I think tend to draw attention away from the flat plane behind them. The Schools doesn't have those, so looks better with some extra mesh in the firebox door on the top LOD. Obviously it'll drop back to a flat plane by the third LOD.

Naturally all this did involve repacking most of the UV's again, but that's not as big a deal as it sounds. I'm now convinced that doing any loco from scratch is going to involve a complete repack at least 3 times, and you might as well accept it and just do it. Mesh really is close to finalised now. No, really. I mean it this time. Really. It's literally to the point where one or two tris might change, but apart from that it's just eye candy.

[Grandma Ruth]

I just got a "Schools" class in the Latvia map - I thought they were still under development? (I know we had a conversation about it because there is one IRL near where I live, which I intend to go see when and if we ever get released from this blinking lockdown!) Is this the finished article?

[Gumboots]

No the "finished" one is still sitting in Blender. I could pack it up pretty quickly. I just haven't bothered so far because a/ there's a bit more graphics detailing to do and b/ I haven't been playing any scenarios that could use it. But if you want to give it a go I'll sort out a PK4. It's a lot better than the beta in this thread.

[sbaros]

...the head code (as it was called) was done by those white discs you see on old UK locos. These replaced the running lamps during daylight hours. This means that for RT3 purposes it can't be correct all the time. It has to have either the daylight head code (white discs) or the night head code (running lamps). It can't have both

Having seen these "animations" on various RT3 models, could there be an animation where, say every 12 hours, the head code discs disappear and are replaced by running lamps?

[Gumboots]

Possibly, in theory, if anyone knew how the animation files worked in detail.

[sbaros]

Sadly no, I just ran some tests with locomotives. They won't load animations at all, only buildings can call them.

[Gumboots]

That doesn't surprise me. It's not a big deal anyway.