I took a look at it and did a critique: http://imgur.com/a/ktNrP It may focus on the negative but I really do like the map, I think the jagged walls look way better than the usual subdivided blocks. As for the name drygulch is a little awkward. Personally my go-to name would be emp_wadi: https://en.wikipedia.org/wiki/Wadi Edit: I almost forgot, be sure to add at least one soundscape of some sort, probably just one of the HL2 ones will work, I think emp_duststorm uses one from HL2's coast levels. Making your own isn't that hard though, it's just a script basically and really makes the map feel complete.
first thank you for running around on it and such detailed feedback - more than i ever expected. detailed and helpful, i must be on the wrong forums :D the extremely bad shadowcontroller is because i tried a night theme before the final compile and i didnt look well enough. also i dont really get how it works (i mean i basically know how to use it but the angles are off). just "fixed" it - unless you can help me out maybe? they way i did this cliffs, "zcontrast" is only possible for a final (or very late) version or i dont know how to alter the displacements without it getting out of hand. atm i could move walls around or add new stuff and its just vetex editing the brushes. but yes, it needs more variation. i did tiny bits at the start base and for some plateaus. that should give you a rough idea. theres only a single floor texture atm. id deffinately want a bit more variation. probably a road or something. but thats eyecandy. its only meant to mark the resnodes in 3d space. im still searching for ideas. maybe ill put a rock there in the end - or i build some geometry like for the start refinary - do you have suggestions for a prop i could use? that there is no sun is odd because there was one a few versions ago. i just checked must have deleted it when i tried the night theme. readded, fixed. the super large lightmap settings is due to that it refused to compile with anything lower. or well it compiled but it threw a hunk overflow exception when i tried running it. im not happy with that either. any ideas? missed that platform, thanks (and fixed). thats no overlay, thats the shadow (plus the wrong shadow-control color) of the refinary prop crack thing. and it also is only there to mark ref-spots in 3d space. had just forgotten to disable as for the rest, fixed. res nodes are all 1 res and there is six per side, so 12 in total. but mark me what refs you want to give how many resources ill try that. lol optimization is something ive spend like 2minutes on, its basically giant nodraws with a hint cutting them off at the top - everything above is a huge viscluster. i could do more by setting draw distance for all the props - once they are at their final position. but the s-shaped layout deffinately helps. how did you make that screen? yeah lighting settings ... ill have to play around with them a lot. but compiling takes its time ... fixed the leaking displacement thing. yeah it needs a 3d sky and i should put something there. ive already copied some rock right behind the start bases, theres something supposed to lead into the canyon. i want to block it off differently aswell. dunno how though. rocks probably, they are a solution for everything. should i block it with a cliff face in the meanwhile? mm. i will create better fitting nowdraw blocks. such things bug me aswell. thats something i really forgot about. probably happens a lot, i didnt account for that at all. ive put it "on the list". im honest, i was to lazy to add info_lightings - theres a lot of rocks. but yes, i should motivate myself even for a playtest version. tomorrow. as for textures thats something i was going to ask for a future version. i think i understand what you mean, if not ill ask if thats ok. good point. forgot that aswell. was a rather quick fix, done already. so if you have any ideas concerning the lighting - you can have the vmf aswell if you want - please tell me. if not ill put it on my list for the very next version since shadows are important, but gameplay is more. i tried to answer in the correct order. could i have answered on that site you linked me to? sry if so ...
testing 5min tweaking 2min compiling 15min (not noticing its done, add 20min) this slows progress so much -.- getting there though. still some minor issues with lighting, but for a playtest its probably sufficient. also something weird happened, i suddently could reduce the lightmap scale back to 32 on displacements, yesterday the map threw a hunk engine overflow exception unless i used 128 (thus why it was so shit). i have no idea what i changed, but im happy it works
fixed most things varbles complained about. theres still some rocks that need info_lighting and stuff (its a lot better though, i used different compile options), but id really like people to try out first. https://dl.dropboxusercontent.com/u/331067/map/emp_drygulch_a2.zip and the vmf https://dl.dropboxusercontent.com/u/331067/map/emp_drygulch_a2_vmf.zip
I can confirm A2 has everything present and working. The hunk overflow is something that used to fuck me up a lot, it's the engine saying the data chunk with the lightmaps in it is too big to handle so it crashes. It only happens on massive maps because a 1/2 km^2 of 16 units/luxel terrain generates a really large size hunk of lighting data. Type r_hunkalloclightmaps 0 before loading your map to disable it so you can load the map just to see, but to fix it you should reign in lightmap scale starting with faces and displacements where it won't matter. For instance I would leave your cliff walls and sand floor at 16 but then select just the walls and floor sections that will not be recieving shadows, either directly facing the sun or away from the sun or just the middle of the open area, and turn those all to 64. The flat tops of the highest cliffs would also drop to 64. Backs of cliffs or things that will never be seen from infantry perspective are good candidates too. Honestly that should be enough, even on big maps it doesn't usually take much to get rid of the hunk overflow. As for compile time, my test compile settings for slaughtered only take 2 minutes total and I'm working on a laptop. With regular lightmap settings VRAD takes 2 minutes and fast vvis is instant. It used to be much higher, like 20 minutes, because i had a large tool texture face on the entire bottom of the skybox cube, which was having lighting calculated for it. Heres my compile settings: Don't forget that you can do an onlyents compile to test different settings on entities that are already in the map (timers, shadow_control, fog settings, dynamic models like the refinery props etc.) Only make sure you do a final compile for release as an onlyents compile isn't meant to be on servers. As for the shadow_control, it looks fine in the A2 version. I've never had a problem with the angles being off but it wouldn't surprise me if it just does that sometimes. On slaughtered my shadow_control angles are set to 30 318 0 and the light_environment is 0 318 0 with -30 pitch
oh its a full vis compile and it only takes a few seconds, vbsp is done faster then the console prints it out my vrad opitons (for the final compile) are -ldr -final -bounce 100 -threads 3 -LargeDispSampleRadius -staticproppolys -staticproplighting -game $gamedir $path\$file and -hdr -final -bounce 100 -threads 3 -LargeDispSampleRadius -staticproppolys -staticproplighting -game $gamedir $path\$file thats what takes 10minutes - i have another lower setting for test compiles, no worries :D if i use -both instead of ldr and hdr seperately it hangs up at a point. -threads 3 is to be still able to use the computer during the compile (only got 4 cores and no HT) :D edit: btw varbles, what do you think about the start bases?
good luck i think there were afew low pop maps in the past like emp_oasis, emp_batos(i think), emp_tms, emp_badlands and emp_lake i may of missed afew
version a3: fixed the hurt and unintended ninja rax spot. this plateau is only meant to give defenders an advantage. now i can only hope this will come up in rotation a few times so i get good feedback. https://dl.dropboxusercontent.com/u/331067/map/emp_drygulch_a3.zip
if you dont want it to be used for ninjas then raise the place up since it look like you could easily wallride there
As per request, here's version 1 of my briefings for drygulch. Let me know what you think: After finally pushing past the stalemate at Arid Desert, command has sent a small force to enter a long-abandoned gulch, establish a small base, and resume mining the abundant natural resources along the riverbed. It seemed like a simple job, but right after cutting comms with the main army, an expeditionary Imperial contingent was spotted entering the gulch from the north. These resources will come at a price. The Third Contingent was double-timing it to the Arid Desert when news of the rebels' victory came through. Given the agility of Jekotian armor, a retreat was likely to end in them catching up to the much slower supply trucks. Instead, the commander has ordered for the units to continue forward and take shelter in a long, narrow gulch, where mobility isn't nearly as important. However, upon approaching, air reconnaisance reported a Jekotian force heading in from the gulch's other entrance.
Here's a demo of the game we played on NUBS today, it was kind of an awkward round and I wasn't playing very well but it still is probably useful. https://dl.dropboxusercontent.com/u/2644783/gulch1.dem It felt a little odd in the beginning because of the starting vf being there but still having to save up for the refs, like on the old emp_money. All the open space behind the front line is why NF lost ultimately, BE destroyed our main and then starting attacking our forward base from behind.
more like two guys fucking around with your cv and (general) you only built jeeps at start but no tanks afterwards. also no turrets, no walls and a lack of rax behind the frontlines (1 forward rax isnt enough - never) but it kinda worked out for the start at least. ill wait for a bit more feedback, i already noticed that center needs a bit more love. in the direct center is a tad too open for my taste. sadly it was less representitive - looked a lot like your comm fucked it up - but the low player count assumption seemed to work - players stuck together in the same area most of the time (wonder, wonder) - thats something to start with (dont forget, its a full-sized map). sadly it neither let me open tab-menu to see how many players in total (smthg like 10-12?) nor does it show res flow - how about income? oh and ramming and killing lights apc at start with your jeep was just epic - i lol'd :D
This was our team minus boovdawg who commanded for half the game and lilquan who fucked around a lot. As for res it seemed kinda low for us but BE didn't seem to have a problem rolling meds and AFVs. We had a lot of res at the end because they destroyed our main but left most of the refs intact as they slowly finished off our last rax. As far as empires games go it was pretty lackluster, due to the team balance and people on the server mostly.
No offence to varbles, but just from that screenshot one can tell that those teams were a bit imba in favor of BE. I wonder how the late game would have panned out if they had been more even. That polak I have never actually seen is a bit of a sweeper though, daym.
a bit higher income might be good - i just fear its gonna turn into a ton once it reaches the multi-cap. guess ill prebuild the 3rd base ref aswell so that noone has to stay behind that far. i dont think up-ing start res will help, it will only lead to fast radar or more vehicles right away. but i also dont want people to run minutes to reach frontlines or resnodes. running is boring. if i prebuild everything its gonna be dull for the comm (i mean placing refs isnt the funnest thing to do, but at least its something)
