The International Simutrans Forum

PakSets and Customization => Pak128.Britain => Topic started by: The Hood on January 21, 2014, 10:17:42 PM

Title: pak128.Britain scale problems
Post by: The Hood on January 21, 2014, 10:17:42 PM
Great work with the new trams, but I must say I'm concerned about the new scaling. See this:

(http://imagizer.imageshack.us/v2/800x600q90/845/1x2w.jpg) (https://imageshack.com/i/nh1x2wj)

Is it just me or does the tram not look way too small for the road and houses nearby (e.g. a double decker tram appears to be only as tall as a single floor in a building? I'm really not sure why you've had to do such a drastic rescaling anyway as I applied the 30m to one tile length scaling when I first drew them. I've investigated your blends but can't quite work out why - the length seems about right on your new blends (but e.g. the Blackpool Balloon is a bit short I think). One thing which is probably not helping is width and height. While I'm sure yours are correctly scaled compared to the length, most pak128.Britain models actually don't have a 1:1 aspect ratio as you might expect. They are actually slightly wider and taller than they should be compared to their length, partially to avoid visual oddities like the one in the picture above - hence the blends I originally drew. I know this isn't strictly accurate but given the scaling problems with simutrans generally (e.g. different length scales, the fact that in real life two rail tracks are much closer than the visual gap in simutrans between tiles), that's not a problem. More of a concern is the general graphical "feel" - and to me these newly rescaled models "feel" too small - I'm lacking the visual detail there used to be and the just look too small compared to what's around. After all this is a transport game so the main features should be the vehicles themselves - if anything they should be larger than their surroundings so as to stand out.

Personally I'd recommend keeping to the scale I originally adopted, which was what Kieron had used on the rail vehicles too (so in fact all rail, road and trams should already be in scale, although I admit a few earlier road and rail vehicles are possibly too large compared to others). I'd be interested what Kieron has to say on this though.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 21, 2014, 11:42:07 PM
Hmm - the trouble with using something that is "slightly wider and taller than they should be" is that it is very hard to work out just how "slightly" wider and taller that they should be, which results in terrible inconsistencies and things looking awful, especially when compared to each other. With the larger ships, Kieron's logarithmic scaling system works so that we can work out exactly what all the dimensions of the ships should be using a mathematical formula, and everything is consistent. The trams actually look so much more real when they are narrow like real trams. It would be very odd indeed, I think, for the trams to be a different scale to the trains, especially when they share the same track (and I have checked: the width between the rails is exactly correct for the scale that I have used).

I do not know how the buildings are scaled, but, if they are out of scale for the vehicles, this is the problem rather than anything else, I think. The road width seems correct in proportion to the tram, however: see this Hong Kong example, for instance:

(http://people.exeter.ac.uk/nkjdatta/photos/abroad/hongkong/tram.jpg)

or this rather lovely photograph from Dundee:

(http://bygone.dundeecity.gov.uk/sites/default/files/imagecache/wysiwyg_imageupload_lightbox_preset/wysiwyg_imageupload/9/trams.png)

If one adjusts for the fact that the latter photograph shows double track trams, what we see in the image above does not look too far off.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 21, 2014, 11:54:53 PM
Well, this goes back to it being impossible to have all scales in simutrans consistent. Road vehicles and trams were roughly the same scale as buildings previously. Buildings are to the same scale as roads, with trains and railways to a smaller scale. If you try and make railways and roads the same scale then either trains will end up very long with lots of clipping errors, or road tiles will appear unnaturally wide. Really roads should be two tiles across (which would solve double track trams at the same time), but that's another story, and a whole load of extra things to code.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 22, 2014, 12:02:59 AM
Quote from: kierongreen on January 21, 2014, 11:54:53 PM
Well, this goes back to it being impossible to have all scales in simutrans consistent. Road vehicles and trams were roughly the same scale as buildings previously. Buildings are to the same scale as roads, with trains and railways to a smaller scale. If you try and make railways and roads the same scale then either trains will end up very long with lots of clipping errors, or road tiles will appear unnaturally wide. Really roads should be two tiles across (which would solve double track trams at the same time), but that's another story, and a whole load of extra things to code.

I thought that you had said (http://forum.simutrans.com/index.php?topic=12764.msg126581#msg126581) that a consistent scale between rail and road vehicles is preferred?
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 22, 2014, 01:01:05 AM
The Hoods post made me remember why they weren't to begin with...
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 22, 2014, 01:16:00 AM
Hmm, that is after they have all been done.

However, and in any event, looking at the photographs, the scale seems correct. A single deck on a tram is much lower than a storey in a building, and the tram in the photograph appears to be, at most, barely higher than the first floors of the buildings opposite (and those are not buildings with high floors, although there are three storeys). The width of the tram tracks, which is on the same scale as the trams as I have scaled them, is broadly consistent with the width of the road, as seen in the photographs, assuming the road to be a reasonably but not exceptionally wide road. Given that the height, length and width of the trams are scaled to the same scale as the width of the tracks, the scale must also, it therefore follows, be in keeping with the scale of the road. If the scale of the road is in scale with the buildings, it would also be in scale with those.

It may well be easy to mistake objects such as vehicles for being larger relative to roads and buildings than they appear in the perspective used by Simutrans because our perspective of them at ground level (when we are much closer to them) emphasises their size relative to surrounding, larger objects. When viewed from above, as in the two photographs in my previous post, the scales seem more consistent.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 22, 2014, 08:00:51 AM
James, I know the rationale behind the scaling and I'm not doubting it's all been calculated correctly and carefully. However I think at times it's best to go with what looks good in the game. This of course is more subjective, and you may well have a different view to me on this, but for me the trams (at your correct scale) look too small. As I said, it's a transport game and to have all of the vehicles as little dots moving around somehow misses the point of the game to me. I'd advocate just using a bit of visual judgment to get scales looking visually "right" rather than trying to do everything by formula. The rough scale that I found worked was 2 blender grid squares wide and 3 tall for most "standard" vehicles - double deck trams worked out more like 4 tall. Length scaled as to the 30m per tile.

There is also the issue of building scale - this has never really been enforced properly and I think there are a few buildings which are clearly to a larger scale than others that could do with scaling down - Archon's green and white pub in the shot above is one but I know there's one or two of my houses that also could do with scaling down. Again, given the constraints of what fits on a tile in Simutrans judgment should be used.

As kieron says the best solution would be to eventually go for double tile roads but that would introduce clipping problems (length of vehicle relative to tile woudl double) and require a huge amount of code change and graphics recoding.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 22, 2014, 10:53:19 AM
When I first started with graphics for the pakset, I did not consider precise scaling, and just tried to make what looked about right. After a while, I realised that a number of the graphics - mostly rail vehicles at the time - were obviously out of scale compared with one another and looked very poor, and this problem became more acute as I added more and more vehicles, and it was necessary for the visual distinction between each of them (especially railway carriages, where earlier and later versions of similar carriages often differed only in length) became finer. I then started working out what the correct scale was and using that, and found that things looked far better in relation to each other when I did. I then found that a number of existing vehicles that were already in the pakset or that I had scaled based on those that were (Midland Railway locomotives in particular) were obviously overscaled, and, when rescaled, again looked much better in comparison to other vehicles. Now, there are still a good number of earlier steam locomotives obviously to the wrong scale which look silly when compared to other locomotives or carriages, or even the width of the track.

What looks better is, of course, subjective, but I have always found that things being correctly scaled look better to me (at least, when seen alongside other things of the same scale: individually, of course, bigger is usually better, and when one is accustomed to vehicles being a particular size in the game, one gets the short-term impression of them looking worse when made smaller), and I strongly suspect that the overall look of the pakset will appear more natural to people (and therefore better) the more consistent and accurate that the proportions of things are in comparison to each other. When judging the appearance of things, it is important, I think, to try to set aside one's memory of the larger (and therefore more detailed) versions of the images by comparison and focus on what the images look like in proportion to other things in the game as they are now scaled.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 23, 2014, 06:24:53 PM
OK, we really need to knock this one on the head. I've done some more playing around with scales. Firstly, a class 37. Two images below - one in its current form and one applying a strict 30m to one tile and 1:1 aspect ratio. Although the length was correct, the old image was higher and wider than a 1:1 aspect ratio would imply. I was already aware of this, and personally I favour the old approach as it fills the width of a tile better and gives better detail. However not so much of a problem so far.

(http://imagizer.imageshack.us/v2/800x600q90/89/bzah.png) (https://imageshack.com/i/2hbzahp)(http://imagizer.imageshack.us/v2/800x600q90/543/qr1t.png) (https://imageshack.com/i/f3qr1tp)

The next pair of images really highlights the problems we have at present. The first house is adapted from the 1880 terrace houses and is consistent in scale with that image. The second version is correct to 30m to one tile (actually it's a little large but I wanted to fit a whole number of semi pairs into a tile) - this model is accurately scaled as it's an idealised version of my own house so I know the real dimensions of it pretty well.

(http://imagizer.imageshack.us/v2/800x600q90/32/85c9.png) (https://imageshack.com/i/0w85c9p)(http://imagizer.imageshack.us/v2/800x600q90/197/1vr7.png) (https://imageshack.com/i/5h1vr7p)

As you can see the existing buildings are about 3x too large compared to what they "should" be if we keep to a strict scale with vehicles. One option of course is to enforce the 30m to one tile more rigorously and rescale all the buildings down to what we have in the second image. However I think this is not a good idea because (a) everything would be tiny and that seems to be against the idea of 128px tiles and (b) you'd still have major scale issues. For example a single carriageway road plus pavement is rarely more than 10m wide, so that would be 3x too wide on that scale (presumably where the 3x too large buildings come from if they are drawn to look good compared to the road!). You also still have the problems with the distance between double railway tracks being 30m (!) and ships and planes being on a reduced scale anyway.

In other words inconsistency of scale is going to be a part of the game. We just have to decide on what sort of inconsistency we accept and what we don't.

My views on this:
- the "general" scales should remain roughly as they are at present (this is what people are used to and have given so much positive feedback on, and it also means not so much redrawing)
- some things are still unacceptably out of scale however. In particular I'd look into scaling (some, or possibly even many) buildings down a bit.
- Accurately scaled vehicles at 1:1 aspect ratio and 30m to a tile are generally too small for good visual game experience, especially with small vehicles early in the game.
- Given the fact that widths are exaggerated compared to lengths by default (given the width of tracks/roads is much more than would normally be the case in real life because of the need for elongated objects on square tiles) I suggest we keep the exaggerated width and height that has been a feature of most road/rail vehicles in the set from the beginning. Lengths should still be at 30m per tile for road/rail vehicles, or if anything, scaled UP not down.
- We should have a standard width/height ratio to avoid irregularities. The class 37 was about 1.25-1.3 times larger in both these directions. I'd go with this for now and apply it e.g. to the new trams.

Discuss :)
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 23, 2014, 08:23:54 PM
Ahh, yes, a knotty problem indeed. I shall have to think about this more rigorously when I have a little more time; I should also be interested in Kieron's and others' views.

However, a few provisional points for now to be thinking about. Firstly, I actually rather like the second class 37. It looks more real and less like a toy; it gives, in other words, a more convincing illusion of actually being a class 37, and gives the player a more powerful feeling of actually managing a transport network; the suspension of disblief is more effective.

Likewise, I do like the buildings in the right hand side picture. It reminds me of some of the wonderful discussions that we all had a while back in this (http://forum.simutrans.com/index.php?topic=12424.msg122834#msg122834) thread, where we decided that, ideally, it would be good if houses were about twice as wide as the roads (that would still be splendid, incidentally).

The issue with the width of the roads and railways is a somewhat intractable one and it is most unfortunate that it is not easier to fix this in the code. This should not, however, mean, in my opinion, that vehicles should not have a consistent scale compared to other vehicles, although it might have to mean that vehicles have a different scale from buildings. Overall, I think, if things are going to be out of scale from each other by necessity, it is better for a whole class of things (such as buildings) all to be in accurate scale with each other but out of scale with a whole class of other things (such as vehicles), all of which are also in scale with each other. This is because inconsistencies of scale between different objects of the same type are much more noticeable than inconsistencies in scale between objects of different types.

Edit: It strikes me that there is a potentially significant and rather important advantage to be gained from having smaller buildings. Because buildings are limited to one tile square, it is difficult to have anything like enough range of size between the smallest and largest city buildings. This has economic and game-play as well as aesthetic problems. If the smaller city buildings could be grouped together as has been drawn above, this would greatly improve the situation, and allow much larger buildings on a single tile drawn to a smaller scale.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 24, 2014, 11:28:11 AM
I agree that vehicles should have a consistent scale, as should buildings, but it doesn't necessarily work to make bith groups to the same scale. As it stands, all the vehicles myself and kieron drew for road and rail were to a consistent scale (with the possible exception of a few steam trucks which I never quite thought looked right, but I had no data for them). In theory so too for the buildings but this is harder to get right as you have to fit a whole building into a single tile and that can be deceptive. What I'm advocating is that we stick to the original scalings that Kieron started and I followed with buildings, trains and trams. To do differently would (a) involve more work than I'm willing to put in to changing perfectly good graphics and (b) significantly change the look and feel of the pak, which is something most players already like about the pak (if I'm not mistaken). Without a thorough review of how the game works (e.g. multiple tile city buildings, one-direction roads per tile, clipping improvements, etc so that the current "scale" could be re-used on a pak64 tile base) I'm very reluctant to change too much. We have a game engine with certain limitations and we should concentrate on what works best within these limits.
Title: Re: pak128.Britain scale problems
Post by: prissi on January 24, 2014, 02:46:21 PM
Personally, and since this game is about transport, I found all pak128.britain vehicles too small. And clipping errors (like for larger trains) are long gone since Dwachs patch (as long as vehcile length stay shorter than 16).

There is even a patch around for multitile houses, which has the limitation of 2x2 houses only upgradeable by a 2x2 house. THis might not even be an issue, if most houses are actually 2x2.
Title: Re: pak128.Britain scale problems
Post by: VOLVO on January 24, 2014, 08:45:04 PM
I do think the vehicle should be enlarged,  but if the buses are going to be enlarged by that scale, it won't fit on the road.
The trains will be just about right though.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 24, 2014, 08:52:24 PM
Yes - I'd tend to agree with that view prissi. If anything I'd be in favour of increasing the scale of vehicles rather than decreasing them. pak128.Britain vehicles could all be scaled up by a factor of 1.25 and still fit on one tile (ie length<16). I've spent some time comparing paksets and from what I can tell pak128 vehicles tend to be slightly wider and taller but a lot shorter (many of them are still exactly half a tile long like in the old days). pak128.german has much larger vehicles than pak128.Britain, although looking at them I suspect they are not 1:1 ratios. Nevertheless they do look good. I've made a trial with your newer vehicles jamespetts and from what I can see exaggerating their widths and heights by a factor of 1.28 seems to give more visually pleasing results.

Good to know about the houses though prissi - is that making its way into trunk any time soon?
Title: Re: pak128.Britain scale problems
Post by: prissi on January 24, 2014, 09:21:52 PM
There are so many things to upgrade paksets: Maybe not before the next release.
Title: Re: pak128.Britain scale problems
Post by: Junna on January 24, 2014, 09:48:46 PM
Quote from: The Hood on January 24, 2014, 08:52:24 PM
Yes - I'd tend to agree with that view prissi. If anything I'd be in favour of increasing the scale of vehicles rather than decreasing them. pak128.Britain vehicles could all be scaled up by a factor of 1.25 and still fit on one tile (ie length<16). I've spent some time comparing paksets and from what I can tell pak128 vehicles tend to be slightly wider and taller but a lot shorter (many of them are still exactly half a tile long like in the old days). pak128.german has much larger vehicles than pak128.Britain, although looking at them I suspect they are not 1:1 ratios. Nevertheless they do look good. I've made a trial with your newer vehicles jamespetts and from what I can see exaggerating their widths and heights by a factor of 1.28 seems to give more visually pleasing results.

Looks horrendous on bends with longer vehicles. Already at the limit of what is bearable with the current lengths, I'd say.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 24, 2014, 10:13:28 PM
The Hood's post with comparison images shows that a picture is indeed worth a thousand words.

prissi is right in saying that pak128.Britain already has relatively small vehicles compared to other paksets. Making them even smaller isn't the way forward.

For what it's worth my original scaling was intended to be broadly similar to pak128 - originally pak128.Britain was an addon so I intended the vehicles and buildings could be interchangeable. When length of vehicles was no longer limited to 8 the lengths were rescaled but not the height and widths.

I agree with broadly keeping things the way they are. Although I support taking advantage of larger buildings if these are ever implemented in trunk. I'd say that 1x2 buildings should be used for most houses - some dense tenements could remain 1x1 while large blocks of flats could be 2x2. It would be good if there'd be a way of upgrading from larger (e.g. 2x2) buildings to smaller (e.g. 1x1) buildings though.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 24, 2014, 10:55:00 PM
It would be useful to know if anyone has any suggestions for vehicles or buildings which appear obviously out of scale with other objects. I suspect we'll eventually have to do some selective rescaling, even if we do choose a set of scales that is similar to what most objects are currently designed for.
Title: Re: pak128.Britain scale problems
Post by: ӔO on January 24, 2014, 11:19:15 PM
pak160.britain?
If I remember correctly, 160 gave a better scale when coupled with 20m^2 per tile.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 25, 2014, 09:53:35 AM
Problems with that are obviously (a) muchos redrawing and (b) exacerbating the distance between tracks for double track.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 25, 2014, 11:16:17 AM
To be clear on what is being suggested: is the idea that all vehicles (save for large ships, which have their own particular scaling as discussed elsewhere) are scaled consistently with each other, but are 1.25 times higher and taller relative to their length than in reality? If that is so, the trams will still be significantly smaller than they were previously, as the older generation of trams (but not the second generation, from Metrlink onwards) were considerably too long as well as to wide/high, and I suspect that the same applies to road vehicles, too. Early steam locomotives are also out of scale compared to later steam locomotives (with the exception of the early LBSCR locomotives that I added last autumn).

As to the buildings, what scale relative to vehicles are these now? The scale of buildings to vehicles is obviously a rather different question than the scale of vehicles to other vehicles. As to multi-tile buildings - this would be splendid, but 2x1 is likely to be more useful than 2x2 (unless, I suppose, the 2x2 buildings comprise multiple actual dwellings). I should still be interested to see what the smaller buildings look like in the game, however; as written above, there is a great deal to be said for allowing a greater ratio between the smallest and largest buildings by putting multiple smaller buildings on the same tile. Also, the class 37 looks like a miniature railway locomotive next to the larger sized buildings, which, to my mind, is a little silly.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 25, 2014, 11:27:04 AM
James, you seem to have access to much more data on the lengths of early vehicles than I could ever find. Could you please let me have the dimensions for as many of the locomotives and trams that you have drawn (or for the vehicles that you say are out of scale)? That would allow me to come up with a better standard scale for road and rail vehicles. Even then I think some of your rescaled vehicles are too small - e.g. your rescaled Jenny Lind was too short (I have found the locomotive only was 7.8m long and checked using your 15m ruler - this is somewhere between the length of your version and my original version). Incidentally the Rocket appears to be the correct length at 30m per tile.

In terms of rescaling road vehicles, I'd be wary of making them much smaller than they are now - the road would dwarf them!

As for boats I suggest we leave them as they are - after all they are all scaled consistently with each other now.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 25, 2014, 11:53:08 AM
The data on lengths of early steam locomotives are somewhat patchy; the lengths for many of the locomotives can be found in "The British Steam Railway Locomotive from 1825 to 1925" by E. L. Ahrons, but data on earlier locomotives is less detailed. Often, I have to use the wheelbase data instead of length, which means scaling the wheelbase accurately and judging visually whether the length is correct based on the proportion of the wheelbase in the Blender graphic compared with the photograph from which I am making it. The lengths of railway carriages seem easier to find, often because length was a defining feature of those carriages compared to later carriages. I have quite a number of books on railway carriages, both generally and of a few specific companies. In many cases, especially the later locomotives, carriages and multiple units, the length is readily available on Wikipedia, although I imagine that you have found that already. For locomotives and carriages for which I cannot find data, I scale them to match approximately the size of contemporary vehicles.

For the trams, the steam trams and their trailers are based on information from some specialist books on the subject that I have acquired recently. The horses are scaled according to information from Wikipedia as to the normal size of draught horses. For electric trams, I found good information both in a book that I had bought about London trams and also on the internet: see here (http://forum.simutrans.com/index.php?topic=6521.msg129252#msg129252) for a summary of some of that information.

As for actually providing you with the dimensions of some of the earlier vehicles, aside from the trams where I recorded the data in advance, I did not keep an independent record of the length or where I found it, applying the scale to the graphics as I drew them directly from the source. (Some of the graphics that I produced initially were not strictly scaled at all, and may need re-doing). The rail graphics have generally been scaled strictly to length with the width and height being either unadjusted from the original graphics from which I adapted the vehicle, or scaled/adjusted by visual appearance only. It would take a long time to reconstruct all of those data; perhaps I could give you a sample?

As to road vehicles, I am not entirely clear on what the scale of them is supposed to be. They really should be the same scale as the trams, or else they would look very out of place alongside each other; and the trams need to be the same scale as the trains because they use the same tracks and stations. That means that the road vehicles need to have the same scale as everything else (the ships, you will recall, are also the same scale as the trains, save that ships longer than 15m are scaled to the square root of their additional length above 15m to keep them all within a maximum tile dimension of 255x255). I think that Giuseppe has a separate scale for aircraft, but this is less problematic, since aircraft do not interact as directly with other vehicles as road, rail and tram vehicles do. For narrow gauge vehicles, I scaled everything based on lengths relative to the Double Fairlie that you had drawn, assuming that to be the correct size.

Edit: The lengths of many of the railway carriages are apparent in the .dat files.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 25, 2014, 12:40:25 PM
A sample of a few lengths of locomotives for different time periods would be handy.

I have to say though, the more I think about this the more I'm convinced that an absolute scale that we stick to rigidly is not a good idea. I've always drawn the graphics with the idea that they should be visually pleasing (which requires a certain level of detail which should be enough to distinguish that vehicle from other similar ones if you look closely enough), but also somehow plausible in their surroundings in the game. In this regard the human eye/brain is quite helpful as we aren't very good at distinguishing different scales for objects much bigger than us, and we also judge vertical scale differently (and usually worse) to horizontal scales. That allows for substantial leeway in putting two different lengthscales adjacent in the game without looking too odd.

For me, and I suspect most other players and artists out there, the game is as much about looking visually good as it is about being strictly accurate. I'm certainly not up for rescaling everything, especially in such a way as to lose most of the detail that has been put into these models. Much better to make new vehicles consistent with what we have (and indeed decided on deliberately in the past) rather than change mid-stream. You might well say that the trains look like "model" trains compared to some of the buildings - if there are specific ones which make this effect particularly bad I might rescale a few - but there are plenty of other things about simutrans that to me are more "silly" than that, things which stick out as being unrealistic far more than the juxtaposition of scales, e.g.: only 8 directions of travel allowed, only 4 of which can have stations; regular tiles of the same size forcing grid-like layouts on cities and only specific multiples of length for buildings and stations; every turn is 45 degrees; even with the half heights every change of height is still steep, etc etc etc. These won't go away by "consistent" scaling, so I remain utterly unconvinced of the benefits of accurate scaling. The scales we use should only be a guide so as to prevent blatantly over/undersized vehicles from ruining the graphical appearance.

I'm compiling a list of objects I think could do with rescaling because they appear too large or small for the other objects. Please suggest any yourself. But most of the rescaled trams are on that list for me - they just look too small and I miss the detail they used to have - this is less enjoyable as a gaming experience for me therefore.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 25, 2014, 01:17:16 PM
As I wrote in an earlier post, I started with the same principle as you describe above, but as I added more things, I saw more obvious and problematic anomalies in scale, especially with different vehicles compared to each other (and refer to my earlier post on the details of these problems: see also here (http://forum.simutrans.com/index.php?topic=12764.0) for a further discussion and some visual examples). These obvious visual anomalies have exactly the opposite effect of being pleasing: they are visually jarring and make the pakset appear crude and ill-executed. Accurate visual scale is not an end in itself; rather, it is an efficient means of preventing some things from looking out of place compared to other things and ensuring that the pakset looks good. I cannot think of any other way of achieving that end effectively. I do not think that it makes a difference that the Simutrans engine provides limitations to the level of graphical realism possible: the aim should be to make it look as good as possible within those limits, rather than for those limits to be a reason to give up on making it look good. Also, if there are not standardised scales, it becomes very hard indeed to know how big to produce any given model, which was another problem that I had before I worked on scale standardisation: I produced graphics for locomotives and railway carriages, then, when I came to produce another one that had to be larger than it, I found that it ended up being larger than another locomotive or carriage already in the pakset than which it should have been smaller, and then one has to go back and re-scale something, duplicating work, and it is difficult to work out which one should be re-scaled, not least because one re-scaling will often then require a whole lot of other vehicles to be rescaled, and so forth.

As to the trams, to what scale did you intend to re-scale them; did you intend to multiply their widths and heights by 1.25, or do something else? Did you intend to do that to all trams (in which case modern trams would be considerably larger than they were before), or just earlier trams? If so, where is the cutoff for modern/earlier trams, given the continuum at Blackpool (and the need for modern Blackpool trams to run alongside first generation Blackpool trams; it is noteworthy that the second generation of Blackpool trams did not exist when the original set of trams was drawn)?

On trains looking miniature compared to the buildings: I was referring to the images in this thread of the larger house and the Class 37; the smaller houses look much better in comparison to the locomotive than the large house.

As to what needs to be rescaled: if we retain a standardised scale for rail and tram vehicles based on the 15m rule and being 1.25x as tall and wide as they are long (relative to scale), then the only re-scaling necessary would be on the widths and heights of trams and the re-scaling necessary in any event on older steam locomotives and other miscellaneous out of scale rail vehicles. (I still think that a 1:1 scale looks considerably better, and not only because that is how ships are scaled, but also because the vehicles actually look like the vehicles that they represent rather better, but that is another matter).

The buildings generally as they stand are far too big for the vehicles. This can either be tolerated or fixed by making the buildings smaller. There are great advantages in making the buildings smaller, especially the ability to have a much greater ratio between the smallest and largest buildings, which will have substantial economic benefits for the game. (I have also noticed that the buildings are not always consistently scaled to each other, and larger buildings appear to be on a smaller scale than smaller buildings). Nonetheless, if it is thought that re-scaling is too much work, the current buildings look reasonably good in their own right. What I do not think makes sense, however, is to scale some vehicles, such as trams and 'buses, to match the buildings, and other vehicles, such as trains and ships, to a completely different scale, because it looks far worse when one vehicle is out of scale compared to another vehicle than it does when vehicles as a class are out of scale with buildings as a class (and see above on the necessity of 'buses interacting with trams and trams interacting with trains).

As to the locomotives, I shall look into compiling a few samples when I have some time.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 25, 2014, 02:40:49 PM
Further to what the Hood has said, I would say that trains should look good next to trams, and trams look good next to buses. That doesn't mean the scale has to be identical - road vehicles pass buildings more frequently so should probably more in scale with those, whereas trains cannot be at this scale.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 25, 2014, 07:02:40 PM
Quote from: kierongreen on January 25, 2014, 02:40:49 PM
Further to what the Hood has said, I would say that trains should look good next to trams, and trams look good next to buses.

I can't really understand how this can be achieved without using a consistent scale.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 25, 2014, 09:40:38 PM
We accept that there are different scales going on, that vehicles that are likely to be near each other should have similar but not necessarily identical scales. If trams are 1.25x scale of trains, and buses are 1.25x scale of trams for example it's unlikely that the differences compared to all being at the same scale will be particularly objectionable.

Ultimately having buildings being able to take up several tiles, as well as vehicles able to be several tiles long, possibly combined with each road lane being a tile, no limit of image size, and more realistic curves would possibly enable a consistent scale across all transport types and buildings. That is a long way off - if indeed it is desirable at all.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 25, 2014, 10:02:12 PM
Quote from: kierongreen on January 25, 2014, 09:40:38 PM
We accept that there are different scales going on, that vehicles that are likely to be near each other should have similar but not necessarily identical scales. If trams are 1.25x scale of trains, and buses are 1.25x scale of trams for example it's unlikely that the differences compared to all being at the same scale will be particularly objectionable.

To my mind, having different scales for different sorts of vehicles is sub-optimal as described above; but at least having different vehicles of the same type all to the same scale as each other (and specific known ratios as you propose between such vehicles, especially when those ratios are fairly small) is better than having different vehicles of the same type drawn to different scales. I should be interested to know what this would actually look like (and it might be necessary to try a few different scales).

QuoteUltimately having buildings being able to take up several tiles, as well as vehicles able to be several tiles long, possibly combined with each road lane being a tile, no limit of image size, and more realistic curves would possibly enable a consistent scale across all transport types and buildings. That is a long way off - if indeed it is desirable at all.

The things that you describe appear to be in descending order of desirability; the curves thing would be especially difficult and not likely to be worthwhile in the foreseeable future, but having city buildings able to take up multiple tiles would definitely be worthwhile and should in principle be achievable, although it would probably represent rather a lot of work which may prevent it from being feasible for rather a while, depending on the availability of the time of coders interested in this project.

As indicated above, if buildings were re-scaled (and, in the case of smaller buildings, multiple buildings drawn on a single tile as The Hood demonstrated), I think that it is perfectly possible and desirable to have consistent scales as things are now, although this would be even better with the things that you mention, especially those near the top of the list.

However, if we are not to have consistent scales, it is better that the scales be consistently inconsistent and that the degree of inconsistency be small, as you propose.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 26, 2014, 01:13:09 AM
I should say that the reason for suggesting those scales is to reduce the difference in size between different vehicles (much as the formula for ships does). Indeed, an alternative would be to use a similar formula as for ships. The overall aim is that players should get an impression of size, but without small vehicles being difficult to see, or large vehicles taking up too large an area.

Regarding buildings being a lot smaller - this has drawbacks as The Hood points out above and I would not support it.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 01:41:48 AM
From a practical perspective, the system for scaling ships would be more awkward to use, as this requires extra calculation for each vehicle; a fixed scale, I think, is preferable where possible (i.e., where not constrained by maximum size).
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 26, 2014, 11:51:23 AM
I too had been thinking about a non-linear scale, but for the smallest road/rail vehicles rather than the largest. I'd need some more reliable data on the dimensions of some early, short rail vehicles however. I'm sure we can all use a calculator easily enough to work out whatever scale factor to use if we do use a non-linear scale.

I'm also fairly certain that width and height exaggeration are a good idea so as to lessen the mismatch of scales between vehicles and buildings. Almost all road and rail vehicles I have tested which I have good data for have exaggerations of between 1.25 and 1.3 in these dimensions. I propose we set that down as a standard for all road and rail vehicles. I might try redrawing some of the new trams in this way and see how they look then.
Title: Re: pak128.Britain scale problems
Post by: Vladki on January 26, 2014, 01:05:41 PM
Hi, I would like to add my two cents to this discussion.

Firstly I must say that I like the visual appearance of pak128.Britain as is.
Secondly I understand that, once you notice the scale inconsistencies, you see them everywhere and they start to bother you more and more. A similar thing to bad typography and kerning. So hush, don't tell anybody, most players won't notice a thing.

Thirdly, I did some research and simple math:
Two lane roads (at least here in Czech republic) are designed to be roughly 6 m (local) to 12 m (main) wide. Highways are even wider. Main roads are usually so wide that three cars can fit comfortably. Railroad tracks are built 4 m apart (axis distance), tram track 3 m. So this all gives us 3-4 m wide lane for either road, tram or rail vehicles. Sidewalks should be at least 1,5 m wide. So lets have two sample simutrans tiles.

1 - narrow roads: 2x 3m for cars, 1x 3m for tram, 2x 1.5m for sidewalk = 12m tile width. If we rescale everything to 12m per tile, only standard buses will fit on one tile, modern rail cars (mk3, mk4) would occupy two tiles. Railway gauge (rounded to 1.5 m) would be 8 px. Hmm, that would be lot of detail, but a lot of work... Bad idea...

2 - wide roads: 2x 4m for cars, 1x 4m for tram, 2x 2m for sidewalk = 16m tile width. That is still not enough for articulated buses to fit on one tile. However railway gauge would be 6 px - and that is indeed the gauge used now. And 20/16 is 1.25 - the deformation ratio you claim was used on most road/rail/tram vehicles. So - everything is fine :)

Just leave the tile length to be 20 m and tile width to be 16 m. So that widths have to be exaggerated by 1.25. And you can leave most stuff as is, and just fix those that are completely out of scale.

Scale of houses is another thing. Some small terraced houses look indeed big, but the three houses on the image above are too tiny to my taste. Two on one tile would be fine. And some bigger and higher buildings would deserve occupying 2x1 or 2x2 tiles. However even real buildings have so much variety that it is hard to tell which one is out of scale. I think the only thing that could be used as scale guide is the tram or trolleybus catenary height. I don't know how high it is e.g. in Blackpool for double-deck trams, but for trains it should be at 5.5 m above the rails. And one floor is cca 3 m high (in modern houses), or slightly more on older houses. However the first (ground) floor can be raised or have a shop with higher ceiling so that it will be as high as tram catenary. So I would suggest to agree on consistent floor heights and scale the houses according to it. However, if vehicle heights are exaggerated by 1.25, houses should follow the same ratio.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 01:40:14 PM
I really do not think that a non-linear scale for rail/road/tram vehicles is a good idea: not only would it involve re-scaling nearly every single vehicle of that type in the pakset (and for rail vehicles, there are vast numbers) as the existing vehicles are not scaled in this way, but it would also make small vehicles look very out of place next to the larger vehicles. It would also cause real difficulties with vehicles, such as railway carriages, whose only real distinction from other nearly contemporary railway vehicles is the length (such as the 48ft/54ft LBSCR carriages); the size difference between these carriages will be inadequate if a non-linear scale is used.

There is also the issue not hitherto mentioned that the all the standard gauge railway tracks are scaled in their width to the same scale as the rail vehicles' length (I have not checked the narrow gauge tracks). If rail vehicles are to have a different scale on the x and z axes than on the y axis, should the rails not also have the same scale on the x axis? This would have the advantage of reducing the apparent spacing between the tracks for double track configurations. I should also note that, as things stand, most rail vehicles' wheels are far to far apart for the rails.

As to Vladki's point about noticing poor scaling: it's not an easy thing to keep from people - I certainly notice it, and there is no reason why any discerning user of the pakset might not come to notice it and think less well of the pakset thereafter because of it. It is best to try to prevent this from being a problem in the first place than hope that people do not notice.

As for measurements of early rail vehicles, here is some information on the trams from the "snippet" thread:

Quote
P. 79 gives some interesting information on some early (1875) designs of tram engine, these destined for Paris (whose steam tramway was not a success and which reverted to horse operation before electrification in 1896), but designed and built by English company Merryweather & Sons Ltd.. They were vertical boilered engines, although not to the Wilkinson patent, weighing 2t with dimensions of 1.6m long, 2.01m wide and 3.35m high; it had 5" x 9" cylinders with a 90psi boiler. The heating surface and firegrate areas are not given. This was the only locomotive built to these dimensions, the second locomotive, also sent to Paris, had 6" x 9" cylinders and weighed 4t. This latter type of engine became the Merryweather Type 1, and an example belonging to the Wharncliffe National Rifle Association, which used it in an annual makeshift tramway on Wimbledon Common for many years (p. 80).

Quote
...it is reported that the (standard gauge) Drypool & Marfleet Steam Tramway Co. paid between £500 and £600 apiece for some Thomas Green tramway engines in May 1889 (a caption on a picture on p. 41 gives a figure of £604 per engine). These were compound engines of two cylinders in an 0-4-0 arrangement with a 5ft wheelbase. Boiler details and cylinder dimensions are not given, but the overall length was given as 11ft 6in, a width of 6ft and a total height of 10ft 4in (not including chimney) (p. 39).

(These are probably equivalent to the Kitson no. 2 engines in the current pakset).

Some information on the size of electric trams by way of example:

QuoteIn 1905, the LCC ordered a new class E, and built 300 of them (page 41). These were of a bogie type and fully enclosed from new (apart from the driver's position). These were also 33ft 6in long, but only 15ft 9in high. The bodies were apparently of "sturdier construction" than previous cars (implying higher weight), but actual weight is not given. Most class Es were fitted with (again, I assume two) 42hp motors.

Classes F and G were single deck cars seating 36, which ran through the Kingsway tunnel before it was enlarged (p. 41). They were 33ft 6in long, ran on bogies, and were 11ft high. Power and weight are not given.

QuoteEdit: Another unverified source, this 'blog comment gives dimensions for the Feltham tram as 40′ 10″ long and 7′ 3″ wide.

Some examples of the dimensions of some early steam locomotives (all page references from "The British Steam Railway Locomotive from 1825 to 1925" by E. L. Ahrons"; where wheelbases are given in two parts separated by a "+" sign, this designates the separate distance between each pair of wheels from front to back):

* based on a scale drawing on p. 36 and my (somewhat crude) measurements of it, a Hawthorn type 2-2-2 engine of about 1837-1841 was just under 17ft long (not including buffers or tender);
* the first "long boiler" engines, which appeared in 1841 on the York and North Midland Railway, had a wheelbase of 6ft 4 1/2in + 4ft 4 1/2in in a 2-2-2 configuration and a boiler barrel 11ft 3in long (this was a 5ft gauge railway);
* a Crampton type engine built for the Eastern Counties Railway in 1848 had a wheelbase of 15ft 3in (2-2-2) (p. 72);
* the LNWR "Cornwall" as originally built in 1847 had a wheelbase of of 12ft 11in, and, as rebuilt in 1858, of 14ft 10in (p. 74);
* the LBSCR "Jenny Lind" type had a wheelbase of 7ft + 6ft 6in and driving wheels of 6ft diameter and carrying wheels of 4ft (p. 77);
* the Metropolitan "A" class of 1864 had a wheelbase of its leading truck alone of 4ft, 8ft 10in of its driving wheels and a total wheelbase of 20ft 9in (p. 155);
* a 2-4-0 LNWR express engine of 1866 had a total wheelbase of 15ft 8in (p. 169);
* the MR 156 class of 1867 had a total wheelbase of 16ft 6in (p. 170);
* the GNR Stirling 8ft Single 4-2-2 of 1870 had a total wheelbase of 22ft 11in (p. 185);
* the Midland 0-4-4 tank engine of 1875 had a total wheelbase of 21ft 9in  (p. 201);
* the LBSCR "Terrier" 0-6-0 had a wheelbase of 12ft (p. 202), although note the overhang at the rear; and
* the LBSCR C class 0-6-0 goods engines of 1873 had a wheelbase of 7ft 9in + 7ft 6in with 5ft dia. wheels (p. 203).

I hope that this will suffice for now; as indicated earlier, wheelbases are easier to come by than lengths for older locomotives, but this can be a satisfactorily accurate means of measuring length if dimensions are kept proportional.
Title: Re: pak128.Britain scale problems
Post by: Vladki on January 26, 2014, 02:29:10 PM
Ah, I somehow missed in the previous discussion that the scale was supposed to be 30m per tile... I thought it was 20m per tile. Anyway if we would take the standard railway gauge as the base for scaling - and I would like to keep that consistent with other pak128.* - we have 6px = 1.435 m or for simplicity 1.5 m which gives us 4px=1m, and tile length 64px = 16 m. So if any rescaling I would suggest scale UP, not down. The tram in the first post is narrower than the tracks and too tiny to play with.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 02:31:35 PM
The tram in the first post is scaled consistently with all of the rail vehicles in the pakset, and the tracks, save for those rail vehicles which are at the wrong scale. If the tram were any larger, it would not be in proportion to the tracks.
Title: Re: pak128.Britain scale problems
Post by: Vladki on January 26, 2014, 03:03:33 PM
If it is scaled according to the track, then how do you explain that I can see one rail and the lines on the road that should be hidden beneath the tram. Is it so badly aligned? Counting the pixels - it is 6 px wide and 24 px long, that makes it 1,5 m wide and 6 m long. Is that really the size of blackpool balloon (or whatewer it is on the screenshot?)
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 03:12:37 PM
I imagine that the visibility of the rail and the road is indeed due to an alignment issue: that is, as far as I could find, the size of the Blackpool Millennium (a rebuilt Balloon as featured in the opening post).
Title: Re: pak128.Britain scale problems
Post by: Vladki on January 26, 2014, 04:11:35 PM
Oh, come on, - is that beast really only 6m long? How would you fit 100 people inside? I tried to find some specs, but I couldnt. However judging form the many photos on web I would guess that it is about the same length as common buses - i.e. 12 m, and slightly wider than the track gauge - at least 2 m. That would make it 8 px wide and 48 px long. Two times longer than the picture in the first post.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 26, 2014, 04:47:38 PM
Some dodgy maths going on here... A balloon car is roughly 12m long, it's certainly longer than 6m! But don't forget at pak128 scale 64px is the width of a tile (because of the isometric projection) so 64px = 32m and the tram = 12m long - all correct there. Also the tram tracks are not the realigned ones that James has recently drawn although the trams are aligned for the newer tram tracks.

Thanks for the data James. I've done a few quick checks on the early locos and yes they are "too large" by the 30m per tile accurate scaling. However the graphics produced (e.g. LMR Patentee as a similar early 2-2-2 to the one you had data for) were tiny. Guestimating a similar scale for the Rocket and other really early locos produced similarly microscopic collections of pixels that were just about distinguishable as steam locos (with some imagination). The big question is, what is more important? Accurate scaling or graphics that are worth looking at and actually enjoyable to play with? In an ideal world, both of course. To my mind we can only have both if we SCALE UP most of the graphics, e.g. to 24m or even 20m per tile. That would mean a major redrawing project (and that's just for the vehicles before considering what changes might need to be made for tracks - I'm not sure you've had the "pleasure" of that James!). Given it's realistically practical to only have either a consistent linear scale or graphics that are large enough to see and use enjoyably, it's a no-brainer for me as a graphics artist given the myriad other ways in which Simutrans suspends reality.

Some images just to give a flavour of what I mean...


LNER A3 at current scale:
(http://imagizer.imageshack.us/v2/800x600q90/69/uone.png) (https://imageshack.com/i/1xuonep)


LNER at 24m per tile (NB I only extended the length as it already has exaggerated width and height)
(http://imagizer.imageshack.us/v2/800x600q90/845/g81x.png) (https://imageshack.com/i/nhg81xp)


LMR Patentee 2-2-2 as at present
(http://imagizer.imageshack.us/v2/800x600q90/834/hn1c.png) (https://imageshack.com/i/n6hn1cp)


LMR Patentee at accurate 30m per tile scale
(http://imagizer.imageshack.us/v2/800x600q90/202/7v3l.png)


LMR Patentee at 30m tile length with x1.28 width and height.
(http://imagizer.imageshack.us/v2/800x600q90/198/4ajx.png) (https://imageshack.com/i/5i4ajxp)
Title: Re: pak128.Britain scale problems
Post by: Vladki on January 26, 2014, 05:16:11 PM
Quote from: The Hood on January 26, 2014, 04:47:38 PM
Some dodgy maths going on here... A balloon car is roughly 12m long, it's certainly longer than 6m! But don't forget at pak128 scale 64px is the width of a tile (because of the isometric projection) so 64px = 32m and the tram = 12m long - all correct there.
OK, if the scale is supposed to be 64px = 32m, then the standard rail gauge would be only 3 px. That is less than pak64.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 26, 2014, 05:27:32 PM
Argh this is doing my head in now! This is precisely why I think we should just use our own visual judgement and have some *guideline* scales. Most vehicles are exaggerated at least on their widths to allow for the fact that tracks would be inordinately narrow otherwise. Earlier vehicles should probably be slightly more generously sized so you can actually see something. That's exactly how most graphics are already and that's the reason for it. As long as the general principle holds that for any given point on the timeline larger vehicles are generally larger than smaller vehicles, there really isn't too much of a problem.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 05:35:10 PM
Ahh, the real question is whether it is more important for the vehicles to look good individually or together with the pakset as a whole. Undoubtedly, when looked at individually, the bigger the better; but the whole thing looks awful when one realises that things are drastically out of scale (and note that I have scaled all the early railway carriages to match their accurate length, which then makes them much too small for the overscaled early locomotives). The lower image of the Patentee isn't too small as far as I can discern (indeed, even the second to last one is perfectly serviceable).

Upscaling everything to 24m/tile would be an insane amount of work, and would require redrawing every single vehicle and rail type way, bridge and tunnel in the pakset. The least work to get a consistent scale would be to re-scale all out of scale vehicles to 30m/tile length and 1.25x that in width and height, as most vehicles are at that scale already. The trams would need to be rescaled in width and height only (which would be relatively easy, as the ratios would be the same for all), and the earlier rail vehicles would need to be rescaled correctly.

QuoteAs long as the general principle holds that for any given point on the timeline larger vehicles are generally larger than smaller vehicles, there really isn't too much of a problem.

The trouble with this using anything other than a consistent scale (which is not the same as a linear scale; consider the ships - although applying this to other types of vehicles would require vast amounts of work) is that this can make adding new intermediate vehicles impracticable because (for example) they have to be bigger than an earlier vehicle but to make it any bigger than the earlier vehicle would make it bigger than or the same size as a later vehicle which is supposed to be bigger than it and with when the two can coexist at the same time. That is precisely the experience that I have had many times over when adding rail vehicles (especially early steam locomotives and carriages) that prompted me to be more strict about scaling in the first place.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 26, 2014, 06:10:18 PM
Quote from: jamespetts on January 26, 2014, 05:35:10 PM
Upscaling everything to 24m/tile would be an insane amount of work, and would require redrawing every single vehicle and rail type way, bridge and tunnel in the pakset. The least work to get a consistent scale would be to re-scale all out of scale vehicles to 30m/tile length and 1.25x that in width and height, as most vehicles are at that scale already. The trams would need to be rescaled in width and height only (which would be relatively easy, as the ratios would be the same for all), and the earlier rail vehicles would need to be rescaled correctly.

That's the conclusion that seems best to me. Trams and some of your newest locos would need this width treatment. Some older locos would need scaling down (e.g. Patentee), but the vast majority are probably fine. I also think a few houses could do with scaling down too but that can wait.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 06:29:01 PM
The newer rail (as opposed to tram) locomotives I had not sought to scale accurately for width, but based the width on existing types ultimately from Standard. I did not measure the heights but attempted to judge them visually. Do you think that these need increasing?

Edit: Incidentally, are the road vehicles also not rather larger than 30m/tile?
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 26, 2014, 09:35:53 PM
Some of the newer trains still appear too small - can't remember which just now. I'll investigate another time. Which road vehicles are bothering you? They *should* be the same scale as the trams as that's what I derived them from graphically.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 09:42:55 PM
Ahh, yes, that's the trouble: they are the same scale as the original trams, which are a different scale to the trains (except for the second generation trams, which were the same as the trains). The trams that I drew were to the same scale as the trains, save that they were scaled 1:1 on height and width rather than 1:1.25. Even if the trams are redrawn to be 1.25x wider and higher than they are now, the first generation trams will be smaller than before and therefore smaller than the road vehicles. Which of the newer trains appear to be too small, incidentally?
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 26, 2014, 09:54:01 PM
I still don't follow. The trams were done at 30m per tile (for the ones I had data for), albeit some of the earlier ones I couldn't find data for and made them too large (assuming your scales are correct, which I am because your research was more thorough than mine). Bizarrely however I notice your new Croydon tram is >1 tile even though it is the defining scale for the pak - it should by definition fit exactly into 1 tile. In that case some of your trams are *longer* than the originals!

LBSCR train wise it's particularly the early tanks - on closer inspection of the blends it's the height not width that makes the difference. They just seem to squat and exaggerating the height seems to improve things nicely.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 26, 2014, 10:11:30 PM
Hmm - that is rather odd. All of my scaling is based on my 15m rule, which I used for all the ships and boats, which is in turn based on the length of the Mk. I carriages in the pakset. When scaling the trams, etc., based on the historical data that I found in the thread to which I linked, I found that the second generation trams were almost all the correct length (but I reduced the width and height as I had adopted a 1:1 ratio rather than a 1:1.25 ratio), but the first generation trams I found to be far too big on all dimensions, and rescaled them accordingly (it was not possible to find data for all trams, but, where data were not available, I scaled them to match the size of similar contemporary trams for which data were available). They will now need to be increased in height and width but not length, which will mean that they will still be smaller than they were before.

One of the first generation trams - I forget which now - had to be increased slightly in length to match the accurate scale as measured by my 15m ruler. That might have been the Croydon tram, but I am not sure.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 27, 2014, 08:12:10 AM
OK, but a 2-car 2nd generation tram (e.g. Croydon Flexity Swift or Metrolink T86) should fit exactly into a 1-tile station. This will need tweaking once again.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 27, 2014, 09:33:20 PM
I thought that they were all supposed to be the same scale...?
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 27, 2014, 09:46:10 PM
Yes, but what I'm saying is that these trams DEFINE the scale for the set, so if your rescaled versions don't fit on the tile then something is slightly wrong with your scaler. The Mk1 carriage is scaled from the tram scale. Also bear in mind that the scale in simutrans is discrete (length=1,2,3,4,...) so you effectively have to round your scaling to 1/16ths of a tile (1.875m) or have undue gaps or overlaps between vehicles. Perhaps that is what caused the difference and the length=9 for your tram cars is an artefact of such rounding errors?
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 27, 2014, 09:52:17 PM
Hmm - I was not aware until this discussion that the Flexity Swift trams were intended to define the scale: I have scaled everything, including my re-scaled version of these trams, according to the BR Mk. I carriages (without any rounding).
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 27, 2014, 09:55:56 PM
That was kierongreen's intention from the outset so that this type of 2nd generation tram would still only require a single tile. Some 2nd generation trams are longer (e.g. NET/Sheffield Supertram) so won't fit, but they were drawn later.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 27, 2014, 11:25:19 PM
Indeed - I didn't see the point in minimum length trams not fitting in a single tile. Having trams slightly over this limit will make them disproportionately more expensive to run and difficult to build infrastructure for.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 27, 2014, 11:33:16 PM
But aren't the actual stops on Croydon Tramlink at least twice as long as a normal 'bus stop in any event?

(http://www.crosenstiel.webspace.virginmedia.com/trains/tramlink/r0300330.jpg)
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 27, 2014, 11:53:01 PM
But buses themselves are far shorter than a single tile. You can see this effect in simutrans where traffic jams have large gaps... Single units of short modern trams (including DLR) should only need one tile stations, of course these are quite often doubled up.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 28, 2014, 12:09:21 AM
Would that not have the effect of meaning that trams and 'buses need similar stop infrastructure in Simutrans, whereas in reality trams need greater stop infrastructure?
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 28, 2014, 07:12:43 PM
Short trams don't really need much more than a bus stop. Once you start having units coupled together, or longer trams that's a different matter. Consider that original DLR stock (28m) and a Croydon CR4000 (30m) are only a few metres longer than a Class 153 (23m) for example. There's also consistency here - it shouldn't matter what type of short tram you use it should only use 1 tile otherwise people will get caught out upgrading between seemingly similar models. If you're coupling more than one unit together (like the DLR does now) then you'll need more than one tile. Similarly longer trams, for example Edinburgh Trams (43m - when these are added...) should naturally need more than one tile for a stop. The new Croydon Variobahn trams (again, when added?) while only slightly longer (32m) than the original CR4000's (30m) should probably need more than one tile - the reason being that these are articulated units with multiple segments. With both these and Edinburgh Trams players should have the option to construct units with fewer segments that fit in one tile, or units with more segments which would require an additional tile (or tiles).
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 29, 2014, 09:55:37 PM
The position seems to be that, unbeknownst to me, the Croydon Tramlink tram was intended to be the scale determiner of the pakset at 30m/tile; I have based the scale of the vehicles that I have produced (and there are a great many of those) on the BR Mk. I carriages, on the assumption that they were as good a vehicle as any other on which to base the scale. It turns out that the scale of the BR Mk. I carriages is slightly different (larger) than the Tramlink trams. It does not seem practical to re-do all the vehicles that I have drawn (including ships, boats, trams and trains) to change the scale very slightly. Should the Tramlink trams then be in a (slightly) different scale to all the other vehicles?
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 29, 2014, 10:54:19 PM
Tramlink wasn't the scale basis for the pakset. BR Mk.1 is indeed the scale basis for the pakset. However as I indicated above basic trams should have scale modified slightly to ensure they always fit in one tile.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 29, 2014, 11:20:00 PM
Thank you for the clarification: that is helpful.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 29, 2014, 11:46:19 PM
Just to add - in terms of infrastructure required for trams remember that trams are only single track so automatically require twice as much as buses anyway. Also buses within cities will run mainly over city roads which don't cost anything in maintenance for the player.
Title: Re: pak128.Britain scale problems
Post by: ӔO on January 30, 2014, 01:04:47 AM
I don't think it would be bad to increase the length on those very long trams.
I think, with the updated length calculation code, the sheffield supertram (34.5m) doesn't fit inside 1 tile anyways, on some rotations.

And then there is also the Edinburgh Tram, which is slated to be 42.85m long, which definitely doesn't fit inside one tile.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 30, 2014, 08:29:40 PM
Looks like I was mistaken on the length of trams and what defines pak scale then. Sorry for any confusion caused there. But in any case I'd agree with Kieron about fitting the Flexity Swift and Metrolink T86 onto a single tile. Old graphics should suffice for this.

Otherwise, where are we up to on this debate? Are we now in agreement? Who is going to rescale the various trams and old steam locos to the new scale?
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 30, 2014, 08:40:14 PM
I think that we have found a workable compromise as follows:

(1) all road, tram, rail, narrow gauge, maglev, monorail (if any) vehicles are scaled by reference to the BR Mk. I carriages, and should be 1.25x as high as long and wide, save for some modern trams (it would help to have a definitive list) which should be slightly shorter to enable them to fit into a single tile per tram car;
(2) all water vehicles are scaled by reference to the BR Mk. I carriages save for those longer than 15m, which are scaled on the basis that length additional to 15m is scaled using a square route formula the details of which I now forget (it should be buried in a thread somewhere);
(3) (I assume that aircraft do not need to be further rescaled - Giuseppe has been using his own internally consistent scale for those); and
(4) buildings are to retain their existing scale, save for some which are too large compared to the others.

As to who is to rescale the early steam locomotives and trams, I had planned to do that eventually, but accumulating bug reports in Experimental currently take priority, so it might be a while before I get around to that; the best thing is for whoever is free first to start on the work and post on this thread that the work has started on a particular set of vehicles in order to avoid duplication.

It would also be useful., I think, to have a stickied Pak128.Britain scale guide in the relevant subforum.

One final thing: the ship and boat graphics that I produced were scaled at 1:1:1 rather than 1:1.25:1.25 as trains and road vehicles - do these need rescaling, too?
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 30, 2014, 10:37:47 PM
A useful summary. I'd leave the boats as they are.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 31, 2014, 12:53:57 AM
Quote from: ӔO on January 30, 2014, 01:04:47 AM
I don't think it would be bad to increase the length on those very long trams.
I think, with the updated length calculation code, the sheffield supertram (34.5m) doesn't fit inside 1 tile anyways, on some rotations.

And then there is also the Edinburgh Tram, which is slated to be 42.85m long, which definitely doesn't fit inside one tile.
I already mentioned that the Edinburgh Tram would be more than one tile. Sheffield Supertram can also be more than one tile as there should be a Manchester Metrolink (29m) available at a similar time which does fit in one tile.

In summary trams which should fit in one tile (at most) are:
DLR stock (all varients 8+8)
Croydon CR4000 (8+8)
Manchester T-68 (8+8)
Manchester M5000 (8+8)
Birmingham T-69 (length 7+7)
Tyne and Wear Metro (length 7+7)
Units of each type should be able to be coupled together to form longer trains.

Ones which should be scale length are:
Croydon Variobahn (prototype is 18=4+4+2+4+4, 4+4+2+4 could be allowed fitting into one tile?)
Edinburgh Tram (prototype is 24=4+4+2+4+2+4+4, 4+2+4+2+4 could be allowed fitting into one tile, maybe allow 4+4+2+4+2+4+2+4+4=30?)
Birmingham Urbos 3 (prototype is 19=4+4+3+4+4, 4+4+3+4 could be allowed fitting into one tile?)
Sheffield Supertram (prototype is 20=8+4+8 - no alternatives)
I've given prototype lengths for these with suggested alternatives that we could let players choose to enable them to fit in one tile. Should players be given this choice or is that just a needless complication?

Some of these trams haven't been drawn yet it should be noted. Also for the Tyne and Wear Metro I can't get a source for a unit length, however from http://en.wikipedia.org/wiki/File:Tyne_and_Wear_Metro_Capacity_and_Saloon.jpg (http://en.wikipedia.org/wiki/File:Tyne_and_Wear_Metro_Capacity_and_Saloon.jpg) it appears to be around 24m long (assuming a width of 2.6m), so easily fitting within one tile.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on January 31, 2014, 09:59:57 AM
I know that some of the trams that are actually three articulated vehicles, such as the Nottingham vehicles, are actually drawn as two vehicles, with the central vehicle part of the graphics of the leading vehicle; I suspect that that was done to reduce awkwardness (and I did something similar with the new Blackpool trams for the same reason). I wonder to what extent that this system is applicable for the other trams that you mention?
Title: Re: pak128.Britain scale problems
Post by: kierongreen on January 31, 2014, 01:44:43 PM
Nottingham (new and old) and Blackpool (new) each have 5 segments. For Nottingham lengths would be 19=4+4+3+4+4 Blackpool I think 18=4+4+2+4+4. As for whether segments should be combined, that depends on whether there are unresolvable graphical issues if they are separate, if not I think the preference should be for separate vehicles.
Title: Re: pak128.Britain scale problems
Post by: The Hood on January 31, 2014, 01:59:47 PM
Probably worth seeing what it looks like in separate segments.
Title: Re: pak128.Britain scale problems
Post by: The Hood on February 01, 2014, 05:01:56 PM
Trying to rescale these but kierongreen I'm not sure how you're converting from lengths in m to lengths in the simutrans length parameter. I'm doing 16*length(m)/30 - 16 parameters per tile and 30m per tile... I get NET tram of 33m = 17.6 (round up to 18) and Blackpool 32.3m = 17.2 (round to 17)
Title: Re: pak128.Britain scale problems
Post by: kierongreen on February 01, 2014, 07:21:22 PM
I was using MK1 lengths - that's 28m per tile. Therefore NET (33m) = 18.9 (round up to 19) and Blackpool 32.3 = 18.46 (round down to 18). There isn't information on how long individual segments are, so I've given reasonable values based on photographs and what would combine to give prototypical lengths.

As I said, trams are a bit of an exception as a few simple 2 segment units seem to be just over 28m. These would be difficult to balance if they required 2 tiles so I thought it would be better shrinking these slightly to fit into 1 tile. With articulated units with more than 2 segments players generally have the option as to how long the trams should be - therefore they should be to scale with the rest of the pakset. The slight difference in scale lengths won't be that visually noticeable I think. There is a potential economic distortion of favouring 2 segments over 3 or more, however this could be countered by slightly reducing the capacity of 2 segment units (make it 28/30=93% of prototypes?).

2 segment units that would require slightly altered scales are Croydon CR4000 (30m) and Manchester T-68 (29m).

2 segment units that would already be the correct scale are Manchester M5000 (28m), Birmingham T-69 (24.4m), Tyne and Wear Metro (24m) and DLR (28m).
Title: Re: pak128.Britain scale problems
Post by: The Hood on February 01, 2014, 07:27:00 PM
That makes sense now - 28m per tile rather than 30 as I was using.


EDIT: Just redone the Flexity 2 as 5 vehicles 4+4+2+4+4. Works rather well. Have coded it so 7-car (and indeed 9,11,13 etc car) variants are available as indeed used in Europe.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on February 02, 2014, 11:23:55 AM
Quote from: The Hood on February 01, 2014, 07:27:00 PM
EDIT: Just redone the Flexity 2 as 5 vehicles 4+4+2+4+4. Works rather well. Have coded it so 7-car (and indeed 9,11,13 etc car) variants are available as indeed used in Europe.

Interesting!
Title: Re: pak128.Britain scale problems
Post by: kierongreen on February 02, 2014, 01:05:41 PM
Excellent :)
Title: Re: pak128.Britain scale problems
Post by: Ves on February 19, 2014, 08:23:55 PM
Excuse me for jumping in this thread, but the statement made on the first page in this thread......
QuoteIf you try and make railways and roads the same scale then either trains will end up very long with lots of clipping errors,
...... leaves me wondering:

What kind of clipping errors are you refering to? Train getting outside the 128x128square in the PNG, or is it something else?
Just asking for reference.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on June 01, 2014, 08:39:25 PM
I am doing some work re-scaling the earlier steam locomotives. Data for the very early locomotives being hard to come by, I used the information about the Hackworth locomotive and applied it to the Patentee as The Hood did earlier, noting scale factors of 0.6y, extrapolating that to 0.6z on the assumption that the y and z dimensions were already in correct proportion to each other, and, assuming an 8ft width over frames in these earlier locomotives and multiplying by the factor of 1.25, 0.89x. I have then been using the same scale factors for the other very early locomotives for which no specific data are available, but I seem to be missing a few .blends. They might well be in a .zip file at home somewhere, so I will check that, but currently, I have noted the following missing from the repository:

* SDR 1001 (inc. tender)
* LMR Rocket (inc. tener)
* LMR Lion (inc. tender)
* LMR Patentee tender
* Puffing Billy tender
* SDR Locomotion (inc. tender)
Title: Re: pak128.Britain scale problems
Post by: The Hood on June 01, 2014, 09:23:15 PM
Just added them to my GitHub blends repository.

For info:
- SDR 1001 Tender can't find blend. It's the same as the LNWR DX Goods but recoloured, so easy enough to replicate.
- LMR Lion, Patentee both use same tender graphics as LMR Planet
- Puffing Billy Tender is the same as SDR tender (I think, may need recolouring or swapping back to front?)

Hope that helps.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on June 07, 2014, 07:35:56 PM
Thank you for these - working on this now. When you refer to the "SDR tender", do you mean the SDR 1001 tender, which is the same as the DX Goods tender, which in turn is the same as the LNWR Bloomer tender (in each case, bar livery)?
Title: Re: pak128.Britain scale problems
Post by: The Hood on June 08, 2014, 12:12:08 PM
Sorry meant SDR Locomotion tender. Brain was frazzed when I wrote that!
Title: Re: pak128.Britain scale problems
Post by: jamespetts on June 08, 2014, 12:19:12 PM
Ahh, yes, I had proceeded upon that assumption. Thank you for the clarification!
Title: Re: pak128.Britain scale problems
Post by: jamespetts on October 13, 2015, 10:46:24 PM
I have been working recently on re-scaling quite a number of pre-grouping railway locomotives and carriages which were too wide/high (and, occasionally, too short/long) even taking into account needing to be scaled 1.25 in width and height compared to length, and the results I have been posting to my Github repository. I shall not enumerate every one on this thread to avoid clogging it up, but the Github commits should be self-explanatory. I recommend that these changes be applied to Standard, too, as they are just as relevant there as here. I still have a few more to do (1870s 4 wheelers, Midland and LBSCR carriages of the 1880s-1900s and some miscellaneous locomotives), so watch out for furhter Github commits. Also, although I have not measured yet, I suspect by look that many of the early railway wagons are likewise too large and will need reducing in scale to match the locomotives and carriages of the pre-grouping era.
Title: Re: pak128.Britain scale problems
Post by: kierongreen on October 13, 2015, 11:23:35 PM
Only thing I'd say is with some of the earlier vehicles there's the danger of them becoming so small to be unrecognisable.... It's a difficult balance to get right.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on October 14, 2015, 09:18:50 AM
I don't think that any of them are unrecognisable at their correct scale, and we already standardised on a model of applying the same scale to all vehicles except those over 15m in length over 18 months ago. What always looks particularly bad is different vehicles of adjoining eras being out of scale in comparison to one another; you will recall, no doubt, that the inconsistency of scale was discovered and became a problem after I added a great many more early railway vehicles to fill out the early time-line, creating a smoother continuum of vehicles making such disparities in scale much more noticeable.
Title: Re: pak128.Britain scale problems
Post by: Junna on October 14, 2015, 10:50:03 PM
Goods vehicles would surely need rescaling as well? Well, goods vehicles ought really be expanded and improved upon in general, too. Times at which they become available are largely arbitrary and variation is very limited. One wishes that Simutrans had some of that feature - as in OTTD - where a certain vehicle might have a randomised livery and slight variation in appearance. That would be wonderful for goods wagons.
Title: Re: pak128.Britain scale problems
Post by: jamespetts on October 14, 2015, 11:23:12 PM
Re-scaling early goods vehicles and adding variety to them is on my very long list of things that need doing. If you'd like to accelerate that process, as I know that you're handy with Blender, that would be very gratefully received. I assume that you have access to the Blender scale ruler .dae used for scaling? If research is a problem, I could lend some assistance with that much more quickly than I could actually produce the vehicles themselves.