[Map] cutec v1.6

[raiden]

Hi, I played a bit through your versions Dratz. My frames where similar to the 1-1-7-1 one. I also couldn't recognize big differences between 1-1-7-1 and the pvs version. New was the overall higher frames on these versions, caused maybe by the huge wall, which blocks the view to the opponents base. This absorbs the high peaks of the framerates. Otherwise there was nothing special to realize for me.
I started to build an alternative optimized version now, but there is nothing worth the download yet. I try to implement your plans of reducing the geometry and texturing, but in a more careful way, to keep as much details as possible at first. Surely I'll post this one here to give it to your hands (probably and hopefully next week).
I have also some questions to you: how did you reduce the lighting precission, respectively which command is necessary to do so ? What do you mean with PVS cull, what did you do here ?

greet you

PS: I could find the "/lightprecision" command, but a short description of your steps could still help (which values did you use).

[qreeves]

You increase the value of lightprecision to reduce the size of the lightmaps it produces. Try doubling it, then going back from there if you went too far.

[raiden]

I see, thank you for the concrete description.

[Dratz-_C]

raiden,
As qreeves said, I used a lightprecision of 64 to lower the detail of the lightmaps. Potential Visibility Set (PVS) and the console commands associated with it are described in the Cube 2: Sauerbraten - Editing Reference. To read it, visit http://sauerbraten.org/docs/editref.html#pvs_culling . It should explain the framerate gains that can be derived from exploiting occluders. Greetings right back to you.
Cheers

[raiden]

Thanks for the link. I read much interesting details there. I was searching for a site like this already. Now I know where I can quickly get usefull informations for that topic .

[raiden]

I could add many usefull improvements to the map now. They are looking tiny at first but they took much work. The aim was to speed up the gameplay of the map, mainly by reducing details. I did this and tried to keep many of the decorations, because I think, that they mirror the spirit of the map. Simple removing and deleting of the decoration didn't seem to be enough for me. Usefull reducing and/or changing was my intention here. However, the filesize could be reduced and the map runs a bit faster.
How told, a download will follow during the next week. Sadly I can't continue at the weekend . How Dratz's versions show, it is difficult to generate more frames on a maplayout like this, because the main structure of the map is very wide and open. The visibility range is high and the engine has to generate much data, if the player looks from predetermined points. The change of fog values could speed up the framerate very noticeabel, but that disturbs the fun with the sniper. So I'll let this for the moment.
...

[raiden]

Hello, I post my softly optimized version now and call it version 1.2. I hope we don't get a version chaos . My shortlist of changes is readable in the readme or in the changelogs on the top of the thread. I let the wpt. file out for the moment, to demonstrate that there are no hiccups. I couldn't get enough time to experiment with the waypoints until now. So Dratz, if you want, it's yours.
This version is close to the original, although I liked some ideas for layout changes by Dratz. I also let room for possible optimizations, but I post this one, because I like it in it's state and I don't want to remove more details than necessary. Lastly some of the decorations I made were build by making sketches before(, like the whole map layout). So I'm a bit spare by reducing these . The amount of triangles lies at 120k now (165k in the original). My frames jump between ~ 29 and 180 depending on the players view (from points you see both bases inclusive the middle platform the rates reach the lowest value).
Please tell me your impressions and if you think that the map still has to be optimized or not, or what ever else .

cutec 1.2

ahh: I didn't change any gamemodes here or added them, entities are the same as before.

[greaserpirate]

.
cutec 1.2

That is all I have to say.

It looks just as amazing as the original, loads in less than half the time, and plays like a dream, even on my not-so-top-of-the-line 2-year-old computer!
1.3 would not be complete without this.

BTW, would you consider putting this on Quadropolis? It is supposed to be a site for all the Cube games, and I can guarrantee at least ten points .

[raiden]

Too late, I'm to slow, because I have not the time .

[4rson]

I get 100 to 120 fps standing still on version 1.2 without other players, compared to 80 to 90 fps on Bath, and it still looks gorgeous. Great work.

[Dratz-_C]

Thanks for working on the map. The improvements are good. I tried it last night. It still is choppy for me in the game. Also, with bots moving around the map the long pauses are back. It is an improvement, but for me not enough to make it playable. For a coreference, I get a higher framerate on Bath than Cutec 1.2. Would you be willing to keep trying to raise the framerate? What about more weapons for the empty areas? How about tweaking the pushers? When you/we are done editing we should generate a PVS and see if it slightly helps people with old/slow computers like me. Also, we need to get qreeves on nailing those pauses.
Cheers

[raiden]

... Would you be willing to keep trying to raise the framerate? What about more weapons for the empty areas? How about tweaking the pushers? When you/we are done editing we should generate a PVS and see if it slightly helps people with old/slow computers like me. Also, we need to get qreeves on nailing those pauses.

To your questions: -yes,-yes,-yes and yes,but: since quinton said analogously that he couldn't take any of our versions, because of too much octrees(especialy my, but also your version), I don't know. He called the freezes a bug also (unrepairable ?). So I can't say if it makes sense to continue this project now. I'm at home on saturday again and will see than, what to do. ...
Thank you and greetings, we'll see/read us.

:::edit:::

Thx for the feedback everybody .

[qreeves]

Yeah, there are 10 million octree nodes (ond), and eihrul says this is way too much. I'm guessing it is a result of using a very small grid size a lot for all the tiny details.

[Dratz-_C]

raiden and qreeves,
I am glad to know what's wrong and how to adjust it. I can try come up with a version that doesn't have so many octrees by moving things around so that they align with larger grid sizes. I have done some of that already and now that I know it is essential I will see If I can do it some more.
Cheers

[Onemoar]

I have added this map to my server
va.pugbot.com

[Karoushi]

Yes 10 million octree nodes is too much, I was attempting to use the smallest grid cube for nivell, when putting up the skybox, but it would crash everytime I tried to save so I had to use the large cube to do it which made me change the map a bit.

You want to try and use the large blocks as much as possible and the little ones, as little as possible - lest you end up corrupting your maps save file.

[raiden]

Clearly we'll go on here with that project. It cost too much hours already .

[raiden]

Here comes cutec 1.3.
The maps filesize could be improved, the number of octreenodes was reduced from huge 10 millions to big ~1,5 mio. Big thanks to Karoushi here. Few details are reduced softly (as I usually do on this map). The waypoints come back. The most interesting point: although using waypoints I don't get the old hiccups. Surely the map isn't the fastest one and can't generate the biggest framerates, which isn't a surprise when you collate it with smaller maps which are less complex. But the loading speed and playability could be improved again compared to the values in the past.

Big thanks to Dratz, Karoushi and some other guys. Without your help I had probably lost the hope for this project sometimes .

(watch the top of the thread or the readme for a bit more detailed changelist)

[raiden]

cutec 1.4
The next round. I simplified many details here. The most noticable thing may be the new teleport devices. I gave a new geometry with less tiny cubes to them. This saved space and makes everything a bit more round. I tried to make use of a larger gridsize here. The rest of the clipping material was redifined by me to reduce the number of octreenodes again (before ~1.5 mio and after ~1,3 mio). There are also some other little changes, unnecessary to list them extra.

(~sorry for short update interval)

[Crone123]

Hello,
I have tested the V1.4 with Red Eclipse 1.3.1, it's perfect !
This map is beautiful
Good continuation

[Dratz-_C]

raiden,
I am so happy that you keep making this map better. My occasional pauses are still there but they are reduced in duration. The lowest frame rate I get on 1.4 is 20 a second playing with 11 bots. By all means continue to work on Cutec because I suspect that as you edit it in the future the pauses will be gone for me as well and my frame rate may also modestly improve. I congratulate and encourage you to persist in your revisions.
Cheers

[ballist1c]

Clearly we'll go on here with that project. It cost too much hours already .

those pictures.
THAT. IS. DISGUSTING.

What has dratz's whining made you doooo

[Dratz-_C]

Clearly we'll go on here with that project. It cost too much hours already .

those pictures.
THAT. IS. DISGUSTING.

What has dratz's whining made you doooo

ballist1c,
Yeah, I broke into raiden's house and detached his processor fan so that his Cutec map ended up in carebear land where players attack each other with cotton-candy, Skittles and M&Ms. Seriously though, whining is complaining without contributing; I have put the requisite effort into improving Cutec. You have just decided to complain about complaining without yet making a contribution. I will qualify that statement with the opinion that contributing by playing is helpful in itself and I remember your help on the editor tab paste helper. Thanks for it; I hope to see you in the game.
Cheers

[ballist1c]

I have not contributed to the map because it was raiden's solo project until you made your unofficial version. I was happy to not contribute because I saw no problem with the map. These freezes seem to only happen to you (with your OLD hardware), and in my opinion it is not worth ruining cutec's aesthetics for this. Improving FPS I am fine with but not to the level where that happened. However, it is raiden's map, and it is his choice.
And indeed, whenever someone makes a new map I playtest it at once, if you ever make a solo project I will be happy to playtest it I have recently started using the editor also to revamp old maps.

-Ballistic.

[Dratz-_C]

ballist1c,
According to the reading I had done, raiden was and is welcoming assistance.

... I'm not sure how to go about fixing this, but if you are done with the map I might try to make a more optimised version by removing/simplifying some of the details that aren't absolutely necessary.

Thank you for watching my map qreeves and for pointing out this triangle/render-problem. I thought as much, that this could be the case. I'll try to optimize the map by reducing the amount of triangles. If it shouldn't be enough - you can readily try to help. Probably I'll manage it by myself.

Is this going to be included? PLEASE TELL ME THIS MAP IS GOING TO BE INCLUDED. It's the best Red Eclipse map the community's ever come up with.

If we can optimise it enough, sure. I don't have that kind of time to invest in it at the moment.

Ok, first of all please excuse my long absence in these forums. I was very busy at the last time. The fact that I started the map too ambitious and maybe overloaded it with many details brought me in a difficult situation.
I had to reduce the details of the map. This is not really a problem, but for me it was difficult, because I spend so much work and minds in these details, that I had a too subjective point of view. I couldn't delete enough details of my work, because I liked it how it was. I wanted to preserve the spirit of the first versions and feared to destroy it through the optimisation. So, like you surely can imagine, I get stuck and stopped the work to become more distance. If I would start a new mapping project, I would learn from this mistake.
Sadly I have still not enough time to work on the map. I only can give my first step of an optimised version of cutec. I could sucessfully reduce the amount of triangles here. That saved some frames per second in the game, but that still seems not to be enough. So qreeves, if your offer to help still exists, I would be very grateful, when you could do so. Otherwise I would try to find an experienced mapper to help out or would work on by myself, which would need probably a longer time. If anybody likes to help, be warned: this is work . When you reduce geometrical details it could be the case that you have to rebuild the waypoints for bots. The main theme shouldn't be touched.
If these points don't scare you, you are invited to optimse this map, qreeves (no problem when you also don't have enough time) or anybody else. Feedback is desirable.

What I can give: the latest version of cutec, how I like it.
cutec v1.1

Here comes cutec 1.3.
The maps filesize could be improved, the number of octreenodes was reduced from huge 10 millions to big ~1,5 mio. Big thanks to Karoushi here. Few details are reduced softly (as I usually do on this map). The waypoints come back. The most interesting point: although using waypoints I don't get the old hiccups. Surely the map isn't the fastest one and can't generate the biggest framerates, which isn't a surprise when you collate it with smaller maps which are less complex. But the loading speed and playability could be improved again compared to the values in the past.

Big thanks to Dratz, Karoushi and some other guys. Without your help I had probably lost the hope for this project sometimes .

(watch the top of the thread or the readme for a bit more detailed changelist)

Because of the work of other individuals, Cutec has improved in ways raiden might not have thought of or desired to execute by himself and I am happy with the progressive process and emerging result.
Cheers