Code Review #3943
Change template dimensions
| Status: | Closed | Start date: | 2012-04-26 | |
|---|---|---|---|---|
| Priority: | Normal | Due date: | ||
| Assignee: | - | % Done: | 0% |
|
| Category: | - | |||
| Target version: | - |
Description
They're a little too wide in proportion to length, and have inconsistent height. See attachment for a visualization of an improvement.
History
#1 Updated by Eddi about 1 year ago
that is waaaay too narrow.
#2 Updated by michi_cc about 1 year ago
I definitely agree with Eddi here. IMNSHO Pikka's new finescale tracks are too finescale and not a good guide for vehicle width.
#3 Updated by oberhuemer about 1 year ago
- File template_test.png added
- File template_16_8bpp_normal_resized.png added
With the tracks, I didn't actually look at their size, but at the true width of the trains (3 m) and the ratio of diagonal width to vertical width (3/4 by default). The solution is 4.5, the first example had four pixels of width while this one has five as in UKRS2.
The height's also been corrected to eight pixels everywhere.
#4 Updated by Eddi about 1 year ago
In original Isometry definition, length/width/height should be in an 1:1:1 ratio, now the length has the game-induced scale problems, but i see no reason to apply the same distortion to the width scale as well. that means in the / view, a height of 4m (8px), a width of 3m would imply something closer to 6px.
#5 Updated by Eddi about 1 year ago
My math says 6px in the triangle side should translate to 3px in Y and 5.2px in X direction
#6 Updated by oberhuemer about 1 year ago
Which math now? If the side viewed diagonally is six pixels long, I really don't see how it could be shorter at a 90° angle.
If you switched number 1 and number 3, the newest template is correct.
#7 Updated by Eddi about 1 year ago
no, i meant within the / view the X and Y coordinates.
the really correct maths:problem: draw a line in the game's pseudo-isometric view:
- l = 6
- alpha = arctan(1/2) ~ 26.5°
- delta y = l*sin(alpha) = 6*sin(arctan(1/2)) ~ 2.6833
- delta x = l*cos(alpha) = 6*cos(arctan(1/2)) ~ 5.3666
#8 Updated by oberhuemer about 1 year ago
OK, then the template is fine now (5.36 ~ 5, 2.68 ~ 3).
#9 Updated by Eddi about 1 year ago
anyway, that still doesn't mean it "looks right"...
#10 Updated by oberhuemer about 1 year ago
- File Marnley Transport, 5th Feb 20000.png added
This doesn't?
#11 Updated by Eddi about 1 year ago
no, it does really not...
#13 Updated by Eddi 11 months ago
the | view doesn't work well with the default tracks, that's IMHO the biggest argument for not changing the width at all. the slanted view needs to become wider, and possibly the lower ent moved one pixel to the left (less offset to the | view). and in the / view the offset of the mail wagon doesn't match the offset of the passenger wagon. besides that, the white roof is very white, the green wagons are very grey,
i hope i didn't forget anything.
#14 Updated by oberhuemer 11 months ago
the | view doesn't work well with the default tracks, that's IMHO the biggest argument for not changing the width at all
Or for going to six pixels right away 8) We do have a track set for this anyway, theoretically.
The white is white because it was, and the green looks gray because of the white. Look closely and you'll see it's the same color used on the locomotive. Of course, the white wouldn't have stayed white for long, but it's a little difficult to model that without access to last_service_date through the Python code. I would have coded that this weekend but for the circumstances...
I'll take a moment to clarify things about the width of the roof.
#15 Updated by michi_cc 11 months ago
oberhuemer wrote:
Or for going to six pixels right away 8) We do have a track set for this anyway, theoretically.
Is six pixels the width of the newest Pikka finescale incarnation? Because at least in my opinion these tracks don't look good at all, neither the colours nor the geometry. The old finescale track was nice.
#16 Updated by oberhuemer 11 months ago
- File roofs.png added
I wasn't specifically talking about them, though their dimensions mostly work and they're really not bad.
Here's also a PNG showing diagonal roof widths. It's clear that if I made the new one (left) any wider, there would be no difference to the original size (right) at all.
#17 Updated by oberhuemer 7 months ago
- Status changed from New to Closed