View Full Version : DM-SealedDreadFinalSE

28th Jan 2004, 08:39 AM
As said in comments, I feel an urge to discuss the review greatly; no so much to criticize it, but to get some feedback, since this map had pretty lousy beta stage and there are numerous things that confuse me in the review. As well as to shed light on some decisions I made - I hope this will help any other mappers out there. Let's go point by point... this may get lengthy :P

First off, Homeslice is the only one still reporting problems with crashing; there were some other people that had similar issues with ver1.1, like you can see on Atari forums. SE version seems to fix this for everyone except Homeslice. And yes, all I did was rebuild the map (not mentioning couple of visual fixes, since they shouldn't have any impact), so it has to be an engine glitch. I inquired about this on numerous forums, but sadly noone seems to have a slight idea what could be wrong...

Let's go deeply into my planning procedure. I got the idea for making this map right at the end of 2002, after I played through UT2k3. I had no expirience with external modeling tools, so I decided that my first map will sport stock resources. It also had to be reasonably small and at the same time distinctive from the surge of stock-like maps that I expected to arrive on scene shortly. Since I didn't have many expirience with laying solid floorplans (I was working a lot with Ued2.0, but released nothing, mostly due to goals set too high, see here http://aggy.slohosting.com/ut/sshot ), I focused on the atmosphere and effects. Also, I was a bit disturbed about people saying BSP was passe, and defiantly made a decision to use it extensively. I was influenced by Doom3 a bit at that time and the first image of Sealed Dread was a noisy and working factory laden in a heavy mist, with narrow passages and sharp turns (yup, I am a flak monkey), a lot oz Z-axis and clasutrophobic feeling. Picture an interiror of a working World War 2 submarine, with brick walls and bathing in orange hues. Stock resources were ample in this department and I was pretty happy with the floorplan I scratched on the paper also. I had numerous ideas about special effects and was confident, that because people seemed to forget about sounds and dynamic environtment, my map has an opportunity to stand out. Also, a lot of maps, including Epic's own, had terrible flow problems due to bad smesh collision, whereas I had taken all that into account. But then problems started to arise.

Firstly, I had to drop the idea about really narrow passages, since movement in UT2k3 differs from that in its predecessor and substantial room had to be made for dodge-jumping. I then made a possibly crucial decision, to lower the height difference between the ground and first floor (mostly ramps along walls), to account for trick jumps - I can't agree with thunder on the lack of connectivity between floors, since you can dodge-jump onto higher floors at at least four different spots in the map (this means in four separate rooms - the long corridor, the furnace, the narrower path in the L room and in the main room between the center cube and the lowest ramp)! On the other hand, this seems to screw Z-axis action to some extent. The first concept had those ramps winding around ground floor on higher levels. Also, I had to drop the idea about heavy fog, since it's a DM map and players don't like to have their view obscured - I was participating in a lot of forum threads regarding fog issues. That's when the orignal concept started to transform into a desolate haunted place with severed surroundings. I was then struck by an image I probably saw as a kid in some kind of a thriller (yup, guys, don't let your kids wath TV :P ), with that general round light (those round ones in the big room) hanging above, red brick walls and.. a severed body layign in a kind of barrel. This picture became the permanent matrix for this map. Also, the story was more or less made at that time.

I can hardly accept that BSP geometry is basic, since there are 700 brushes in this map, which is a huge amoung for such a small map. But I agree completely that there are visual scenes that have a great impact in coneveying the boxy picture, like the upper half of the main room. Since players spend much of their time in the main room, they get the feeling of playing in a box, overlooking some delicate BSP deco, although I guess it's mainly my fault cause I didn't put enough emphasis on those spots (like ground corner of the L room and the furnace architecture). I admit that this part of the main room is underdecorated, but I'll get to that shortly. Also, it was my intention not to make rounded rooms, since you dn't seem them in industrial facilities. So what's the lesson here? Don't exert on delicate geometry if you can't make it stand out in the course of a play.

Thunder says that "realism does not equal good gameplay" and generaly speaking he's right. I possess a strange attribute - it seems I can't really enjoy the atmosphere if it isn't beleivable. In other words, in the course of playing other maps or mapping, questions about cohesion between general atmosphere and deco constantly pop up in my mind. Like, WTF is this machine doing here, if it's an elven castle I'm playin in?! Just that my thoughts are more compex, of course :P See, when I laid down the concept of this map upon a picture of last century factories, the amount and nature of deco became kind of "locked". Since I just couldn't enjoy mapping an environment that has inconsistencies present. Sealed Dread could probably look like another AugustMoon, laden with an almost ridiculous amount of smeshes, but the mapper building it prolly wouldn't be me. High time I started mapping for realistic mods, eh? The question is, what makes better atmosphere, pesudo-realistic environment or totally Unreal one?

Oh, the performance and optimization. But let's get a little back to when I was finishing the geometry. Since this is my first map with new engine, it's also the one I was getting expirience on laying smeshes with. I had several problems here, the biggest probably one that skinned smesh's material is also loaded into memory. That means if you want to skin a lot of stock smeshes, you have to duplicate them in mylevel adn change the material, else the invisible textures may painfully gobble you resources. The pure diamond are skyline smeshes with three 10241024 DXT5 textures on them, but only one shown. :))) So watch out. The other problem was, I put too many different smeshes into one room, especially the main room, which makes your rigs hiccup, because smeshes and textures are loaded into memory. The big thing was, of course, phase one of MSUC. Eventhough I started mapping this thing in January, it was made only to about a quarter till two weeks before deadline, mostly due to problems with studying (I study physics, hard as hell). I pulled up my socks and worked whole days, but only managed to release alpha version. Gui can tell you how it looked, since he provided me with some info. Then another round of exams came and I wasn't mapping until around October, when I decided to finish it for good. In the course of hasty mapping back in July, a stuffed the map with lots of smeshes, which killed FPS, so I had to remesh and retexture a huge amount of objects and his took another month on my duron900 machine. Due to incredible amount of actors (over 1000 lights), everything was going very slowly; I had only about three FPS in the editor and making any adjustments was a real pain in the ass. Now you undertand why it had taken so long and that the map couldn't look that much different. Well, now a have a power rig, so don't worry. :D

1000 lights?! Yes, I was very influenced by DavidM's and Hourences' lighting standards and took this very serious. I actually tried to light realistically, lighting only appropriate surfaces and that's why one light fixture usually has 4-5 light sources. Needless to say, it is an extremely laborious work and now I see that maybe it doesn't pay off. So if anyone out there thinks he may get through with it - think again. It seems this engine just isn't advanced enough to let you do it. I really don't get it why the engine doesn't support more elaborately implemented cone lights - usually just restricting the arc of a light would solve everything. I look forward to seeing such options in the next generation engines.

Thunder says that this map could benefit from darker illumination in places. Now that really struck me, because he's the first to mention a possibility od turning the brightness down (save for morpheus, but that's another story)! Mostly everyone that played this map in alpha, beta, and final versions, complained about dark areas. The only exception probably being the floor in the main room at the spot where you can dodge-jump up, where I did turn lower the brightness (morpheus claimed he saw greening there). I think what disturbs Thunder is a too uniform lighting. I admit that this kind of atmosphere would go better along with single player environments, whereas dynamic DM maps should use some more variation. If not from lighting, then maybe from texturing - those white brick walls span too high i.e. they cover too much surface. I planned to intoduce more color variation, but in the course of mapping, such intentions mysterioulsy dwindled away... and I still don't know exactly why, maybe I was fed up with this map and its orange hue already and lost power to make any bigger changes. Some lights are actually broken, but I should break some more, yes :P I was just too afraid of the lack of lighting in areas, since everyone was telling me the map is too dark.

I agree completely that in order to enhance atmosphere, more delicate geometry concepts, such as severed machinery, should have been included. Had I had knowlege of modeling at that time, you would haev seen such elements. Instead I aimed to use stock elements as much as possible and I think you should look a it this way. But as said, more than a year after release it's naive to consider any stock content still appealing. I was well aware of that, so bet on other effects and the layout. The dynamic lights should have been used better, I agree. It's just that nobody seems to use them really, so I guessed they haev some problems, like malfunctioning in the netplay or something. Eventhough I inquired about that on numerous forums, I got no feedback *at all*. That's why I only introduced them as a little spicey feature. The same goes for real-time procedural textures, that comprise pools of water. Whay I didn't add specialr maps? They don't seem to work with procedural textures. In this map I tried twice as more techniques as are actually present. Everything from opacity filters, reflections on oil pools, triggered sounds, dynamic projectors etc. - that is another reason why this map cam out late. The problem was my inexpirience with Photoshop, but that problem is now solved. :) I used redeemer trail because at that time I didn't know how to fix the menu problem in the emitter properties; I was led to solution not sooner than a month ago, so I improvised by modifying existing types. As for the particle density... Thunder, you're being a bit picky here, don't you think? :))) What I miss a bit in the review in this department is the metal sound effect used on grates and those imitation of a guy being closed into the cistern... And furnace definitely stays - it's by far the most proper representation of the place I wanted to make.

Thunder also says the main room is too open. I it's because undcorated upper half, which really makes everything look a bit empty. As for gameplay, this was one my basic decisions - to have several separated floors onto which you can double-jump, comprising a place for medium to small distance fights, where you can change levels quickly. Majority of people that tested the map found this a great and fun concept, that is ideal for UT2k3 moving nature.

And why is the main room so open, woth that darned ring of lights up there? It's a performance issue. The open nature of the room backfired, since there wasn't much room for any kind of optimization. the stuffed machinery makes polys jump to 120.000 in places, and combined with relatively numerous different smeshes and textures it really poses a problem. With sadness I came upon decision to remove deco from the ceiling and as a consequence I had to put out the lighting there also, so I ended with those lights up there. I hoped that this you-don't-know-what's-up-there feature would make just as claustrophobic feeling as would the closed ceiling, but it seems I was wrong, though there are people that find this well-chosen.

As for the song; I was really in two minds - should I choose an eerie, slow tune or a dynamic one, to complement DM nature? I still don't know if I have chosen correctly. Skaarj likes it :P

Weapons placement is where I have some serious questions. Thunder questions placements of shock and googun. As far as I play UT, shock has allway been considered as the best weapon to put on ground floors, because it's probably the best weapon to shoot upward with. I find such statemnts correct and followed it in my map, intentionaly puting it in a more voluminous room to account for combo possibilities. Whereas biorifle is, anyway, object of constant speculation. Some say it's better used in closed departments (that's why I used it there and testers acclaimed), and some are more fond of having it up high to pour down goo on teh enemy. Thunder seems to be more in the other group... Flak is, apart from RL, probably the most delicate weapon in this map. It became quickly apparent that it would reign and become the only weapon if put into an easily accesible place. So I made a.. err.. delicate decision to put it down there. At first there were no conveyor belts there, but a simple corridor and when I realized that this would make the path to long, I came upon an idea to add belts. Sadly, bots don't seem to be affected by physics volumes inf they stand on the firm ground.

And, hm, what do you think about lift jumps? Double Damage? Aggressorization (you did walk into spindle didn't you?)? Blood emitter in the corridor?

So, all in all, yes, it seems I set up my goals too high again. Thunder is right when saying that my inexpirience prevented me from fully taking advantage of the concept. I'll add my incompetent rig and lack of time as a reasons. There, I hope you find this essay :lol: at least a bit helpful.

P.S. 5.5? Gosh that means I'll have to DL and see every map that gets score above 5, since it might contain loads of orignal effects and moves.

28th Jan 2004, 05:58 PM
I too have a duron 900 and my maps always score lower than I`d like - must be a connection ;) (though my GF4ti pulls me through for now)

I remember checking this map out a while back and commented on how I liked the effort put into the machinery etc, trying to make something real or alive always impresses me. I guess as a PLAYABLE map though, people sometimes think differently. The main things they seem to `want` are flow, solidity and good looks. I usually fail in at least one of those areas because I`m devoting too much time to another area. It is very hard to balance everything up, and I`ve now released ... 7 maps for 2k3 (in various incarnations nearer 15) and around 6 for Unreal and UT (99) - and in each one I can feel myself getting better in some area - yet my `last` map scored lower than previous stuff - 6.5 here, which I`m happy with, though it took a LOT of frantic work on modeling and solidity I left out detail in the bases and better bot pathing.. something had to give for me then, cos I was `against the clock` (coming up to xmas - the map is Festive Worlds in case anyone is interested). Basically, it`s best to not have an opinion of what your map should score compared to other maps, because what can seem quite unimaginative, or basic in comparison to our `grand ideas` (this applies to me too) is often SOLID and FUN TO PLAY - 'grand ideas' can often fail in the execution. For instance, I know you must have spent a lot of time with the movers in your map, its a risk - its a short term `cool` effect (as I have learned) it does not affect playability overall so the player will not remember it as affecting his fun (unless he can USE the mover ;) ) that time you spent, and belive me I have been there too and wondered just why I bothered sometimes) *could* have been better used in creating a cohesive, solid level that relied less on `effects` and more on the less exciting stuff like `easy on the eye` lighting principles. Again I feel I can say this as I am guilty of it too. Starfall (another map of mine) got quite some stick from certain areas because it was a bit OTT in the colour department ;) - All I could see was, I`d done this interactive tooth plant, this moving `shroom scanner and an aquarium/crater with a lobster like creature in, and I`ve made a `unique` alien world. This was on top of building a lot of new meshes (I rarely use stock meshes) - the work I put into didn`t pay off because the planning was flawed. I may have put more work or creativity into than `map X` - same as you have done here - but it just needs to be solid, pleasing to the eye, bug free, well optomised and FUN to play. From that time on I have tried to concentrate on the basics more than the overly imaginative, because being totally creative is time consuming and corners will be cut in more 'standard areas' - if you want to compete with the 9+ scoring mappers then do like they do, stick to tried and trusted methods and practice - or if you want to be unique (which some pull off and still score highly - The winner of MUSC for one) then you are going to have to spend twice as long as most getting to grips with modeling (and texturing, which I was appalling at and hopefully am getting better) before you can `compete` on that level. Map reviews can`t be a scientific yardstick to hold every map to, I`m sure I could find maps that scored higher than yours on here, that I enjoy less - it`s down to taste in a large part - but techincal execution and consistency of theme (if there is one) still go a long way.

I say, don`t worry too much about it - you obviously have a lot of skill areas already - work on the ones that you feel let you/your map down this time and do it again - if you have the `fire in your belly` for level design then you will surely find this an exciting challenge rather than a depressing chore. And if the worst happens, and it may happen to me if I can`t `progress` then I`ll just stop doing it - no point letting it rule your life.

Good luck with your future maps - I apreciate the work you put in this - and I`ve had far more disparaging stuff said about some of my maps on forums and Nali city before now. :)


28th Jan 2004, 10:02 PM
Wow! Two lengthy posts! :D

First off, well said Sharpfish! A lot of good points in there.

Aggressor, it was interesting to hear of someone else's development nightmares for a change!:) I have no problems with the vast majority of your comments as they are opinion and could never be right or wrong. This brings me to my first point. I've said it many times, and I'll continue to say it: reviews are just peoples opinions. I've had some debate with a few UT stalwarts and reviewers who believe otherwise, but never doubt this for an instant. That's why nearly every review site worth its salt now has a marking schema in an attempt to reduce this effect. Unfortunately, whilst this may reduce personal interpretation, it doesn't eliminate it. Furthermore, our schema is divided into game play considerations and visual considerations (AVA). Due to the even distribution of scoring between these two facets, you will find in the mid-range scores (5-6.5) maps which look average/play average, maps that look quite good but play average to poor and maps that look average to poor but play quite well. This is why you might feel disillusioned looking at "every map that gets score above 5", as some simply won't look fantastic. This is why you shouldn't compare your level directly with them. A certain score doesn't mean a certain level of quality in a certain department (ie. another map scoring 5.5 may have better or worse lighting than your's, better or worse weapon placement etc.).

I'd just like to clarify for you (and any other level designers) a few questions you had, though. When I said "The room construction is reasonably basic, in terms of BSP geometry" I was refering to the type of brushes used in construction (ie. predominantly cubes of various sizes). Whenever we see a level which uses standard, non-vertex edited cubes as building blocks for a level, we normally mention that construction is basic or simple. This isn't refering to the number of brushes used, just the types of brushes used. The question of connectivity is always a curly one. However, when you state "you can dodge-jump onto higher floors at at least four different spots in the map " this actually highlights my concern, believe it or not. Not every player uses trick jumps to get around, and having to rely on trick jumps to cover level connectivity is never a good idea. The idea should be that you can get from one level to another easily. If a player can use a trick jump to get up there faster, then they are rewarded. You should never penalise a player for not trick jumping. If I played this level without trick jumping I would be penalised. That was the point I was making and is an important consideration for the game play aspects of any map. There wasn't enough normal connectivity.

Lighting is, again, a subjective issue, which is why you have received conflicting feedback during beta. I push for and encourage beta testing whole-heartedly. The art seems to be disappearing. This is largely due to the vast numbers of people trying level design with UT2k3 for the first time, such as yourself. Many level designers develop a propriotorial sensability towards their creations (which I completely understand!) and have no desire to have other people criticise their level. Logically this makes no sense if they plan to post it on the internet! People also try beta testing, and find they get conflicting feedback and get confused. The true art of beta testing is, firstly, being willing to accept criticism about your work (which Aggressor obviously is), but secondly to know what beta feedback to use and which to ignore. Learn to choose what feedback you agree with, any technical concerns, and suggestions which make sense to you. The rest is nice but unneccessary. Give every suggestion consideration, but never feel bound to satisfy everyone because you never will. Weapon placement is another personal preference, again highlighted by Aggressor's confusion over my comments. I always prefer the bio up high as it is less accurate lobbing goo upwards from the ground floor. Bio rifle professionals may disagree. Shock just seemed a little out of the way and I don't think I was ever capped by it. Certainly not often enough for me to notice, so too, the bio. I was capped often with rocket, flak and plasma and occassionally minigun. The need to make the flak harder to get is very valid, however I felt (again, opinion) that it was just a little too far away from the action.

I hope you haven't dispensed entirely with the Duron rig. It would be a great machine to test your new levels on, to see how much your smooth frame rates are your new rig and how much your smooth frame rates are kick arse optimisation! ;)

The final point is on the whole "realism" issue. This discussion hinges on whether you are talking about environments that already exist or could currently exist, versus levels that might eventually exist. Aggressor's comment on machines in elven castles is on the money. This type of thing is an thematic inconsistency, unless there are numerous such machines and the author is going for a "Thief" style pseudo industrial/medieval theme. I always make comment on such inconsistencies. I don't mind more fantastic levels as long as they look like they could exist. For example, I still like bridges to have supports, lights to come from light sources, etc.

Sharpfish, I hope you don't give up on mapping! You do have skill, it's just still developing. I'm sure even Gui would tell you he's still developing his own skill. It never stops (especially with a new engine every couple of years!).

Final words:
The debate of game play versus visuals will never be settled. Utlimately for me, I get more enjoyment out of a level that is fun to play on than one that looks snazzy. Of course, if I can have both I'm very happy! :D Playability is the hardest skill for a level designer to learn, as requirements for it vary from level to level and there are no hard and fast rules per se. Just practice. It sounds as though Aggressor has learned a lot from this exercise, and I agree with Sharpfish that he has some talent in level design which needs to be further developed, to find out what works and what doesn't or isn't worth the effort. Oh, and before I forget, never let anyone tell you that BSP is passe! Utter nonsense. The UT engine was built for BSP, and fancy static meshes were an addendum. BSP lights better (part of the compromise SM make for higher detail) and can, in many cases, still form the neccessary backbone of most UT2k3 maps. Use it wisely, and without fear! ;)

If you have any other questions or I've failed to answer some already posed, please make mention of them.

28th Jan 2004, 10:21 PM
True about BSP - it`s NOT a dirty word - most times it just feels more fun to walk on than a mesh.. something undefinable for me. Meshes are very good for tarting up a level - but are a pain with all the lighting considerations. The map I`m working on now (for 2k4) has lightmaps on virtually all the meshes (which are all custom) - its almost a requirement not an option anymore (for me). Bsp on the other hand lets you concentrate on gameflow and layout without getting caught up in the `modeling` part. And when used wisely, is hardly a penalty on modern rigs... obv dont use BSP for detailed modeling (BSP>SM in Ued is a no-no).

Either way, maps are taking longer and longer to make (good ones) due to all the content they need compared to using `just` BSP - if it`s worth doing, it`s worth taking your time on. At least you put some genuine 'soul' into sealed dread by not just doing the exact same theme even with limited resources that your lack of 3d app brought, this wont stop a more generic level scoring higher if the execution of that level is near perfect.

Oh and I agree about the Duron rig comment, Means most of my maps get optomised pretty well for general use simply because if I don`t I can`t run them myself ;)

29th Jan 2004, 05:50 AM
Thanks for feedback, Thinderstrike, specially on beta testing. I'll bring my next map right here into these forums :) I too am beginning to understand what sharpfish said - those fancy quirks ("grand ideas") are one timers that use up a lot of time, whereas good floorplan and flow are what majority of expirienced players want to *feel*. That doesn't mean that I regret the time I spent for them, not at all, I still think they were a necessity if I wanted to make a discernible map and I'll continue to use them; just the priorites will have to change.

I could see some floorplan inconsistencies in the process of building this map, but mainly due to a slow rig and restrictive timeframes I never decided to completely overhaul those questionable areas. It also seems I'll have to shift some of my beliefs of gameplay mechanics. Let's take that L room for instance. I intentionally put all health packs on ground floor so as to force players to switch levels downwards (which I still think is a good idea, altough I have a slight feeling that problems are due to that mentioned low height difference, which makes a psychological twist - when players see that higher level is very near, they get a subconsious urge to traverse height quickly, whereas if the first floor was ten meters above, they would reconcile with their "low position" and calmly seek a distant elevator; such is the case with VerticalFactory I think). Now it seems that either ground level is too boring to have players meditating :) there for more than a few seconds either there is an insufficient connectivity upwards. So, if I walk up from the belt area into the L room, I have an instant possibility of boost-jumping up, else I have to traverse a reasonably long distance around the corner to the stairs. Couple of months ago I'd find such concept good (and harcore), now I think that maybe it's a bit restrictive.

Thunderstrike says that boost-jumping should only enhance the flow, not comprise the core of it and I totally agree. I get this urge-to-get-higher feeling again; if you look at Nirvana, which is of comparable size, it also has four ways to get to higher floor, but because there is little visual connection between the two, noobody complains. And Victoria etc.. Also, when I looked at reviews or forum posts, most of the people found such jumping possibilities welcome and I thought they'd make my map more "hardcore". It seems I unwittingly chose the hardest-to-implment layout for my map, where I'd have to balance regular and "masterful" connections better. I did start everything with that "boost-jumps should be a reward" belief, it just seems I failed at implementation. Oh well, next time.

Oh, BTW, now I understand why ramps are useful. Apart from the "bigger height difference", there is another move to tackle the problem of coonectivity. If you look at maps that abundantly implement raps, you can see that they also have levels that don't differ in height very much. But the magic in ramps is that eventhough they usually represent the slowest way up, they also comprise a "battleable" floor pieces and make a feeling of moving somewhere, which means that eventhough the player is going up slowly, he has an illusion that because he actually is going up, everything is OK. Tricks to use in my future maps. :)

'I was refering to the type of brushes used in construction (ie. predominantly cubes of various sizes).'

I understand, though a substantial percentage of brushes was made via 2D editor, intersection and vertex editing. As said, I'll have to learn to put more emphasis on such geometry.

Yes, the shock is a bit out of way I know. Furnace room should be put more towards the center, not on the corner of the map, and have doors further towards the center of the room itself. There will be SealedDread][, but not until the next engine.

Oh, and the bots. I have to say that eventhough I used up whole day chasing bots in the game (with shodebug and other stuff), I never stumbled upon them being stuck somewhere. Damn it, my own game is playing tricks on me :D

30th Jan 2004, 03:52 AM
Some great points made in these mammoth posts!
It's good to see some mappers with a smart, realistic, and progressive approach to mapping rather than taking an argumentative approach you often see.

Sorry this post doesn't rival the novels that precede it ;)