New swamp

Production of artwork for the game by regular contributors takes place here.

Moderator: Forum Moderators

User avatar
Sapient
Inactive Developer
Posts: 4453
Joined: November 26th, 2005, 7:41 am
Contact:

Post by Sapient »

the new variations are quite nice.
However, I'd like the submerge point to be halfway between those two choices. You don't want it to look like they are standing in thick grass.
http://www.wesnoth.org/wiki/User:Sapient... "Looks like your skills saved us again. Uh, well at least, they saved Soarin's apple pie."
Darth Fool
Retired Developer
Posts: 2633
Joined: March 22nd, 2004, 11:22 pm
Location: An Earl's Roadstead

Post by Darth Fool »

Sapient wrote:the new variations are quite nice.
However, I'd like the submerge point to be halfway between those two choices. You don't want it to look like they are standing in thick grass.
I concur.
User avatar
Eleazar
Retired Terrain Art Director
Posts: 2481
Joined: July 16th, 2004, 1:47 am
Location: US Midwest
Contact:

Post by Eleazar »

Very nice.

However with the heavy reed coverage, i think the color of the background swamp-water tile should be changed to something much closer to the normal water tile. Otherwise the fact that there's water underneath the reeds can be missed. Yeah i know some swamps look like a field of grass, but some don't, and what's more important, it should be easy for a player to tell the difference.

There are some overlaping issues, (which. i know. are a pain to fix) such as the swamp incorrectly overlapping the north terminal end of the bridge.
Feel free to PM me if you start a new terrain oriented thread. It's easy for me to miss them among all the other art threads.
-> What i might be working on
Attempting Lucidity
User avatar
Jetrel
Posts: 7242
Joined: February 23rd, 2004, 3:36 am
Location: Midwest US

Post by Jetrel »

Eleazar wrote:Very nice.

However with the heavy reed coverage, i think the color of the background swamp-water tile should be changed to something much closer to the normal water tile. Otherwise the fact that there's water underneath the reeds can be missed. Yeah i know some swamps look like a field of grass, but some don't, and what's more important, it should be easy for a player to tell the difference.

There are some overlaping issues, (which. i know. are a pain to fix) such as the swamp incorrectly overlapping the north terminal end of the bridge.
These comments may well be valid, however in the spirit of "release early, release often," I would like to see what has already been done get committed as-is (if it hasn't been already).

Doing so doesn't really impede us making any further tweaks, it just makes sure that this very deserving stuff is in the 1.2 release.
mog
Inactive Developer
Posts: 190
Joined: March 16th, 2006, 2:07 pm
Location: Germany
Contact:

Post by mog »

Eleazar wrote:Very nice.

However with the heavy reed coverage, i think the color of the background swamp-water tile should be changed to something much closer to the normal water tile. Otherwise the fact that there's water underneath the reeds can be missed. Yeah i know some swamps look like a field of grass, but some don't, and what's more important, it should be easy for a player to tell the difference.
I just tried some color variations, but anything close to the water tile has too little contrast or looks like a pond, not like swamp. One variation that looks quite good (imho) and makes the swamp look less like grass is a brownish tint:
Image
There are some overlaping issues, (which. i know. are a pain to fix) such as the swamp incorrectly overlapping the north terminal end of the bridge.
This can be easily fixed by using small reed graphics next to bridges. Unfortunately, this can lead to gaps in the swamp:
Image
So I'd like to keep the number of tiles next to which the small overlay is used as small as possible.

Is it somehow possible to draw the bridge over the reed-overlay?
Aurë entuluva!
User avatar
Jetrel
Posts: 7242
Joined: February 23rd, 2004, 3:36 am
Location: Midwest US

Post by Jetrel »

A suggestion that got PMed to me - like anything else from the general public, take it with a grain of salt:
khamul wrote:New swamp in Art Contrib forum is great. Had a suggestion for a tweak, but can't post in the forum.

The peat bogs I'm familiar with contain these incredibly reflective black pools, like a dark mirror - they're quite distinctive. A variation of the swamp tile that had a black pool, perhaps with a point of reflection, rather than blue ones might help to set the new swamp apart from forest.

It would be great with a half-submerged blacked branch poking out of it - but that's maybe pushing the scale available.

Just a suggestion, anyway.
Darth Fool
Retired Developer
Posts: 2633
Joined: March 22nd, 2004, 11:22 pm
Location: An Earl's Roadstead

Post by Darth Fool »

Darth Fool wrote:
mog wrote:Ok, submitted bug as https://gna.org/bugs/index.php?7274
Excellent. I have assigned the bug to myself, for now. Do you (or anyone else) have a pathologically designed map where it is clear that this effect is happening.
This bug should now be fixed and there is a crude test for it in the test scenario. The WML for terrain that was adjusted to try and compensate for this bug should be reverted to the normal probabilities.
User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Post by JW »

Major props for creating incredibly awesome, realistic swamp tiles!

When I ran 1.1.11 I was highly surprised and impressed!
mog
Inactive Developer
Posts: 190
Joined: March 16th, 2006, 2:07 pm
Location: Germany
Contact:

Post by mog »

Darth Fool wrote:This bug should now be fixed and there is a crude test for it in the test scenario. The WML for terrain that was adjusted to try and compensate for this bug should be reverted to the normal probabilities.
I'm afraid it isn't. It will only fix the case with #matches==#rotations, but when e.g. there are only 2 matches like in this case:
[map]
hwwh
hwwh
hwwh
hwwh
hwwh
[/map]
the probability will be too low. The new behaviour will probably get it wrong even more often than the old one, as most rules have six rotations but match only once.

I looked at the code, and the problem is that the rotated rules are generated at parse time. So one cannot determine later if two rules originally belonged to the same terrain_graphics tag. So the possible solutions would be:
1) generate the rotations in terrain_builder::rule_matches after the probability check.
2) generate rotated contraints and image names at parse time, but store them inside the building_ruleset. When matching, check the probabilty and then iterate over the rotations.

Solution 1) would be really easy to implement, but might be too slow. 2) would be a bit more complicated. So both solutions are probably too disruptive for 1.2, but one should keed it in mind for trunk.

[edit]
I just noticed that this would break cases where you want the rule to match several times (e.g. transitions). That could probably be solved by only applying the aforementioned solutions when the image-tags are identical. Or introduce an attribute to choose the behaviour.
[/edit]
Aurë entuluva!
Darth Fool
Retired Developer
Posts: 2633
Joined: March 22nd, 2004, 11:22 pm
Location: An Earl's Roadstead

Post by Darth Fool »

mog wrote:
Darth Fool wrote:This bug should now be fixed and there is a crude test for it in the test scenario. The WML for terrain that was adjusted to try and compensate for this bug should be reverted to the normal probabilities.
I'm afraid it isn't. It will only fix the case with #matches==#rotations, but when e.g. there are only 2 matches like in this case:
[map]
hwwh
hwwh
hwwh
hwwh
hwwh
[/map]
the probability will be too low. The new behaviour will probably get it wrong even more often than the old one, as most rules have six rotations but match only once.
hmmm... good point. Just goes to show how having a good test for any bug is important.
I looked at the code, and the problem is that the rotated rules are generated at parse time. So one cannot determine later if two rules originally belonged to the same terrain_graphics tag. So the possible solutions would be:
1) generate the rotations in terrain_builder::rule_matches after the probability check.
2) generate rotated contraints and image names at parse time, but store them inside the building_ruleset. When matching, check the probabilty and then iterate over the rotations.

Solution 1) would be really easy to implement, but might be too slow. 2) would be a bit more complicated. So both solutions are probably too disruptive for 1.2, but one should keed it in mind for trunk.

[edit]
I just noticed that this would break cases where you want the rule to match several times (e.g. transitions). That could probably be solved by only applying the aforementioned solutions when the image-tags are identical. Or introduce an attribute to choose the behaviour.
[/edit]
Well, I would say try solution 1 first. Terrain building is generally done only at the beginning of a scenario, so speed is not crucial.
There are other options as well, but unfortunately, I am disappearing for a week, so I won't be able to work on it again until next week. Of course, someone else can feel free to fix it while I'm gone. :)
mog
Inactive Developer
Posts: 190
Joined: March 16th, 2006, 2:07 pm
Location: Germany
Contact:

Post by mog »

I noticed some changes to the swamp lying around in my working copy, namely making the swamp a more brownish colour:
Image

What do you say (I like the new one more)?
Aurë entuluva!
User avatar
Sapient
Inactive Developer
Posts: 4453
Joined: November 26th, 2005, 7:41 am
Contact:

Post by Sapient »

The one on the left... yes, that's exactly the kind of transition I was trying to re-create earlier. Well done.
http://www.wesnoth.org/wiki/User:Sapient... "Looks like your skills saved us again. Uh, well at least, they saved Soarin's apple pie."
Boucman
Inactive Developer
Posts: 2119
Joined: March 31st, 2004, 1:04 pm

Post by Boucman »

I like the one on the left more... colour differenciation is a good thing for terrain
Fight key loggers: write some perl using vim
User avatar
Jetrel
Posts: 7242
Joined: February 23rd, 2004, 3:36 am
Location: Midwest US

Post by Jetrel »

I like the new swamp color more. Not so sure about pulling the reeds back from the edge - how they blended into the grass was something I liked, but the color change is definitely good.

Your judgement is probably better than mine, though. :)
mog
Inactive Developer
Posts: 190
Joined: March 16th, 2006, 2:07 pm
Location: Germany
Contact:

Post by mog »

Jetryl wrote:I like the new swamp color more. Not so sure about pulling the reeds back from the edge - how they blended into the grass was something I liked, but the color change is definitely good.
I didn't change anything about the reeds, it's just that the new colour makes the edge more distinct (not neccessarily a bad thing though). Also it's less visible in most other tile combinations, but I will probably someday change the reeds to make it blend better.

For now though, I committed the change to trunk and 1.2 (I also adjusted the submerge depth to 0.35 as discussed in this thread).
Aurë entuluva!
Post Reply