merlin
Public Area Guest
Posts: 19
|
Post by merlin on Jun 16, 2004 14:15:41 GMT
Just a quick note on Chris' mention that all scale 7 graphics would be at an offset of 100: that's not necessarily the case. There are in fact 18 different radius values for viewable terrain features if you follow the landscaping technique on www.icemark.com. Some of the furthest ones will merge together in actual application, meaning they will all seem to be on the horizon, but you still get more than 8 actual offset values, so the result is that sometimes you will have the same scale but differing offsets. This will happen with the smaller scale features, where the offsets are closer together. Yes, you are correct, I misread my document. The following shows 3 same size but pulled forward graphics. I think the fact they were in the middle and thus draw order wrong as you mention, is what really threw me. { 5, 4, 86, 64, 7 }, { 4, 5, 114, 64, 7 }, { 6, 2, 41, 64, 7 }, { 6, 1, 21, 64, 7 }, { 6,-1, -21, 64, 7 }, { 6, 0, 0, 64, 7 }, { 5, 3, 69, 63, 7 }, { 3, 5, 131, 63, 7 }, { 4, 4, 100, 63, 7 },
|
|
|
Post by sparrowhawk on Jun 16, 2004 15:04:34 GMT
If I understand your code correctly the order in which the items are drawn is determined by the order of the arrays as you have listed them. However, what seems obvious from that is that items which have a higher offset - and are therefore further away - are sometimes drwan after items with a lower offset, which are presumably nearer the observer. That's probably it then. COuld you point out which ones you've noticed (sorry, head stuck in battle script when not actually doing real work!)
|
|
merlin
Public Area Guest
Posts: 19
|
Post by merlin on Jun 16, 2004 15:10:56 GMT
For the draw order just look down your list... for size 7 you go something like 99,99,99,100,99,99... it should be all 100 then 99 etc..
If that all fails, I have sent you a full list of coordinates as used in LOM. However lom uses a 256x192 screen with horizon at 128.
You appear to be using 350x181 with horizon at 63. If you confirm that this is your window then I will generate numbers for this and then you can really compare.
|
|
|
Post by celebaglar on Jun 16, 2004 15:29:54 GMT
I'm inclined to think so too. From my observations, the abnormalities seem to match the area in which the order appears to be wrong. No problem, I'll post some figures tonight. I think some of your offsets may match the wrong locations, so it's not just the order in which things are drawn, but their actual offsets too. It may be that Peter's tables were wrong in the first place (must dig out his Amiga port and find out). I've actually calculated the real distances to each feature, so you can't go wrong if you take your order from that. At the moment it looks like some of the features that should be at 100 are drawn at 99 and vice versa, with the same issue repeating through the bigger scale groupings. Need to do some shopping first, so I'll post figures when I get back...
|
|
|
Post by sparrowhawk on Jun 16, 2004 16:02:53 GMT
thanks to you all
|
|
|
Post by celebaglar on Jun 16, 2004 21:22:15 GMT
Here's my version of the NE offsets: Note that there are no chnages to the x-offset (I haven't checked those at all), and that all the changes are to Size 4 and smaller. In fact, the offsets for sizes 4, 5, 6 and 7 seemed to be reversed, with the nearer objects actually having the bigger offset. I have also re-ordered the arrays in those sizes so that the terrain farthest away gets drawn first, gradually moving inwards. The N will follow shortly. I have a bit more to do on that, as I haven't yet sorted my own data, and my "polar" viewpoint is based on the E view rather than the N one. P.S. I've added the commas at the end of the array lines, which I'd missed out earlier, so copy and paste should work OK.
|
|
|
Post by celebaglar on Jun 16, 2004 22:09:09 GMT
Just a quick mention on subject of the y-offset for the armies while I'm here. For sizes 0,1,2 and 3 you can afford to put the army slightly below the terrain feature, which gives a better visual result. For the other sizes, this doesn't work because of the very small difference between offsets, so for sizes 4 through 7 the armies should have the same y-offset as the terrain they are standing on.
|
|
|
Post by sparrowhawk on Jun 16, 2004 22:17:04 GMT
excellent, thanks . will try to get it done tomorrow, but I have a recacitrant asp website that is misbehaving and a client demo on friday, so...
|
|
|
Post by celebaglar on Jun 19, 2004 18:27:25 GMT
Better late than never, so here's the revised NORTH view array list:
|
|
|
Post by sparrowhawk on Jul 1, 2004 14:44:33 GMT
thanks (again!)
|
|
|
Post by sparrowhawk on Jul 7, 2004 9:22:55 GMT
Bill, I've replaced the old arrays with your amended ones. Using the same location as shown in your example earlier (Amarin looking N from the corelay corridor to the domain of Dawn), here's what I get: Moving one square N gives: Turning NW at this point gives: The downs still seem slightly off, but could that be due to the gfx themselves (ie larger than village so looks nearer?)
|
|
|
Post by celebaglar on Jul 7, 2004 9:58:51 GMT
Looks like the village and downs are at the same displacement level, which is probably a side-effect of the approximation process. As I said previously, there are 18 (actually 19!) different distances for objects in the standard view, and this would mean 19 displacement levels. Clearly you don't have that, nor would it work at this scale to try and implement it, so there are places where different theoretical displacement levels need to approximate to the same actual level. I think this is such a situation, but it's still much more acceptable than the previous result. I'll post my theoretical tables later tonight, so you can have a look for reference. Whether they'll help you make any changes remains to be seen.
|
|
|
Post by celebaglar on Jul 7, 2004 10:03:22 GMT
Incidentally, I think a little lack of clarity for items at a distance actually helps the atmosphere, because things become clearer as you get closer, and this applies to distances as well as terrain. As such, I think the above views work while the earlier ones didn't, because they actually gave outright misleading views rather than merely unclear ones.
|
|
|
Post by sparrowhawk on Jul 7, 2004 10:04:59 GMT
Thanks for the explanation Bill, and you'll be glad to know that you are now credited in the MU source code! I agree that the effect is acceptable. Mind you, I never noticed the bug in the first place, despite wandering around MU for 18 months now (is it really that long...)
|
|
|
Post by sparrowhawk on Jul 7, 2004 10:06:14 GMT
Thanks for the explanation Bill, and you'll be glad to know that you are now credited in the MU source code! Here you are: // N View data
//Portions (c) 2004 Jean-Yves Rouffiac. All Rights Reserved.
//From data originally supplied by Peter Armstrong //Corrections/adjustments by Bill Hoggett
|
|