Page 1 of 1
Color mapping problems
Posted: 24 Jan 2009 16:10
by Uwe
Hi,
for the GermanRV project (see sig) I'm currently implementing lots of special stuff which relies on appropriate usage of recolor tables. For example, vehicles will provide "real" liveries which are achieved by recoloring a basic wrong-colored sprite using recolor tables. Now the code for that works fine in TTDPatch, but not in OTTD, all I get to see is the sprite as it is drawn in the underlying graphics file, without any recoloring applied.
The code for it looks like this:
Code: Select all
10 * 5 0D 01 00 00 00 // set var 01 = var 00
11 * 9 0D 00 00 00 FE FF 08 00 01 // reserve 256 tables
12 * 5 0D 01 02 00 01 // set var 01 = var 00 - var 01
13 * 44 06 01 82 03 01 82 06 01 // add var 01 to the base of each table
14 * 44 0A 0E 80 00 00 80 80 00 // define 2*128 new sprites
...
here come the recolor tables, real sprites and other stuff
...
// check refit cycle to determine usage of 2CC sprites or false color sprites
1887 * 14 02 01 02 81 F2 00 FF 01 //
03 00 00 00 // 2CC case, goto 03 00 (points to 2CC sprites)
04 00 // else goto 04 00 (points to false color sprites)
// CB 2D chain, select color table (Berlin)
1889 * 17 02 01 20 85
7F 00 20 FF FF 00
1A 00 00 00 // use recolor table 0
00 00 80
// CB 2D chain, select color table (Bochum)
1890 * 17 02 01 21 85 // colour mapping select
7F 00 20 FF FF 00
1A 00 01 00 // use recolor table 1
00 00 80
// CB 2D, select color mapping
1891 * 18 02 01 12 81 F2 00 FF 02 // check var F2 (refit cycle)
20 00 01 01 // goto 20 (Berlin)
21 00 02 02 // goto 21 (Bochum)
02 00 // else fail (no recoloring shall be applied to 2CC sprites)
// standard chain, check CBs
1894 * 30 02 01 01 81 0C 00 FF 05
01 80 11 11 // CB 11, make vehicle shorter by one 1/8th
10 00 16 16 // CB 16, articulation
11 00 19 19 // CB 19, cargo subtype display
12 00 2D 2D // CB 2D, select color mapping
13 00 33 33 // CB 33, sound effects
02 00 // goto 02
// action 03
1896 * 10 03 01 01 46 01
FF 00 00 // buy menu, goto 00 00
01 00 // default, goto 01 00
... here come action 0, enabling all used CBs
Somebody with helpful advice? A test file can be provided if necessary.
Re: Color mapping problems
Posted: 24 Jan 2009 22:04
by DaleStan
Code: Select all
10 * 5 0D 01 00 00 00 // set var 01 = var 00
11 * 9 0D 00 00 00 FE FF 08 00 01 // reserve 256 tables
12 * 5 0D 01 02 00 01 // set var 01 = var 00 - var 01
...
// check refit cycle to determine usage of 2CC sprites or false color sprites
1889 * 17 02 01 20 85
7F 00 20 FF FF \2+
1A 00 00 00 // use recolor table 0
00 00 80
1890 * 17 02 01 21 85 // colour mapping select
7F 00 20 FF FF \2+
1A 00 01 00 // use recolor table 1
00 00 80
What happens to those var2s the second time the file is activated? Param 0 is now (usually, at least) 0, which is not what you need to be adding to get to recolor tables 0 and 1.
Also, from your comments, I assume you have 256 total recolor tables -- 128 for Windows and 128 for DOS. If this is the case, as a style/sanity issue, only reserve 128 sprites, and replace them twice, with two action As. Once for DOS and again for Windows. Whichever one comes second should be skipped for the other palette is in use. Otherwise, each CB 2D call needs to check both the refit cycle and the DOS/Win state.
Re: Color mapping problems
Posted: 24 Jan 2009 23:19
by Uwe
I'm afraid I can't completely follow you here. Are you hinting at the information given in the wiki on
Action 06, pointing out the fact to skip the whole GRM stuff during the test stage? If that's the case, then yes, that is indeed missing from the file. It is also missing from the example given in the wiki on the very same page... How would I be able to check if this missing test could lead to problems? I just tried repeated activation and deactivation ingame, and it always worked in TTDPatch, I couldn't produce any corrupt sprites. OTTD wasn't impressed at all, always ignoring the recoloring. To me it seems that OTTD just completely ignores callback 2D.
DOS/Windows is no concern here, I am in fact using all 256 recolor tables for Windows only.
Re: Color mapping problems
Posted: 25 Jan 2009 01:54
by DaleStan
Oh. Nevermind. I assumed your code was the way I'd do it the first time through; I thought sprite 12 was " 12 * 5 0D 00 02 00 01 // set var 00 = var 00 - var 01"
Which is, obviously, not the way to do it. Now that I've read the code correctly, I do not see a problem. You might get useful information with -d grf=9 or whatever it is that OpenTTD uses to generate gobs of GRF-related debugging output.
Re: Color mapping problems
Posted: 25 Jan 2009 07:30
by michael blunck
Uwe wrote:[...] To me it seems that OTTD just completely ignores callback 2D.
No, it works alright in the DBXL 0.9 here.
regards
Michael
Re: Color mapping problems
Posted: 25 Jan 2009 09:47
by Uwe
Ok, I started OTTD with debugging enabled just as DaleStan suggested. I could not see anything that would resemble an error message, but I have attached the logs as well as the grf file that I use for testing. Hopefully that helps tracking down the problem.
Re: Color mapping problems
Posted: 25 Jan 2009 10:41
by planetmaker
Uwe wrote:Ok, I started OTTD with debugging enabled just as DaleStan suggested. I could not see anything that would resemble an error message, but I have attached the logs as well as the grf file that I use for testing. Hopefully that helps tracking down the problem.
Excerpt from your log:
Code: Select all
dbg: [grf] [recolortest.grf:1] DecodeSpecialSprite: Handling action 0x08 in stag
e 0
dbg: [grf] LoadNewGRFFile: Reading NewGRF-file 'trg1r.grf'
dbg: [grf] LoadNewGRFFile: Custom .grf has invalid format
dbg: [grf] LoadNewGRFFile: Reading NewGRF-file 'trgcr.grf'
dbg: [grf] LoadNewGRFFile: Custom .grf has invalid format
dbg: [grf] LoadNewGRFFile: Reading NewGRF-file 'trghr.grf'
dbg: [grf] LoadNewGRFFile: Custom .grf has invalid format
dbg: [grf] LoadNewGRFFile: Reading NewGRF-file 'trgir.grf'
dbg: [grf] LoadNewGRFFile: Custom .grf has invalid format
dbg: [grf] LoadNewGRFFile: Reading NewGRF-file 'trgtr.grf'
dbg: [grf] LoadNewGRFFile: Custom .grf has invalid format
dbg: [grf] LoadNewGRFFile: Reading NewGRF-file 'openttdd.grf'
dbg: [grf] [openttdd.grf:1] DecodeSpecialSprite: Skipping action 0x07 in stage 0
Assuming that your newgrf is the only one loaded, it fails with the action 0x08 in a few places. Though unlikely, alternatively you just might have faulty base graphics. With clean trunk I don't get such errors.
Re: Color mapping problems
Posted: 25 Jan 2009 11:33
by Uwe
Just rechecked that, even without any newgrfs loaded, it keeps telling me those messages about the base graphics. Quite weird, as these are straight from a clean TTDX CD archive I keep on my harddrive. Suggestions?
Apart from that, how could an action 08 possibly fail? And what would happen then? I'd assume the newgrf wouldn't be loaded at all, which isn't the case here. The newgrf runs just fine, only the part with the refitting doesn't work.
Re: Color mapping problems
Posted: 25 Jan 2009 11:43
by Rubidium
It's just telling you that it found the TTD (not New)GRFs and that it determined that they are not a NewGRF. Absolutely nothing to worry about.
Re: Color mapping problems
Posted: 25 Jan 2009 14:20
by DaleStan
Bother. That was not the debugging information I was hopping to see.
I was more interested in the type of data that would be found in grfdebug.log. Does anyone know how to get that sort of information from OpenTTD?
Re: Color mapping problems
Posted: 27 Jan 2009 19:06
by Uwe
Ok, I think I found the problem. It seems CB 2D does indeed not work - for road vehicles. Since Michael Blunck declared the CB would work I changed the vehicle to appear as a train, and behold, recoloring worked like a charm without further fiddling. Should this then reported as bug, and if so, where?
Re: Color mapping problems
Posted: 27 Jan 2009 19:30
by Eddi
Re: Color mapping problems
Posted: 27 Jan 2009 19:48
by frosch
Fixed in r15286.
The colour map cache was not invalidated on refitting.
Re: Color mapping problems
Posted: 29 Jan 2009 15:25
by Roest
michael blunck wrote:
No, it works alright in the DBXL 0.9 here.
Is that one released already and i'm just too stupid to find it?
Re: Color mapping problems
Posted: 29 Jan 2009 16:12
by michael blunck
Roest wrote:michael blunck wrote:
[DBXL 0.9]
Is that one released already and i'm just too stupid to find it?
No, not yet. I was speaking about my own working copy which you won´t find elsewhere, regardless of being stupid or smart.
regards
Michael