TTLG|Thief|Bioshock|System Shock|Deus Ex|Mobile
Page 2 of 4 FirstFirst 1234 LastLast
Results 26 to 50 of 95

Thread: Reducing cell count?

  1. #26
    Member
    Registered: Feb 2000
    Location: Paris, France
    Does brush size have an impact on cell count?

  2. #27
    Dóttirin klęšist oft móšur möttli
    Registered: Apr 2015
    It's likely that big brushes interact with other brushes - more cells.

  3. #28
    Member
    Registered: Oct 2004
    Location: Ireland/Poland
    Does brush size have an impact on cell count?
    Sometimes it does, in conjunction with the texture scale. Large brushes have polygons split into smaller ones limited by the size of a lightmap. The more detailed texture (scale lower than default 16), the smaller resulting divisions and the higher polycount but also larger cell count.

    Polygons with sky texture should always have the maximum texture scale for that reason.

  4. #29
    Member
    Registered: May 2009
    Location: Germany
    Spilled some lemonade over your keyboard by chance?
    If we are talking about a regular PC keyboard, I would recommend disassembling it and giving the foil a good wash.
    My 20 year+ old Microsoft Natural keyboard appreciates that procedure every couple of years.
    Spare all electronics from water, though and let it dry thouroughly!

  5. #30
    Member
    Registered: May 2002
    Location: Texas
    Gnartsch I'm not sure who that was directed at, but I'm using an illuminated Saitek keyboard and I never spilled anything on it. When I try to straffe right with the D key sometimes it delays, and I have problems with the O and L keys. I'm not sure I can take it apart though.

  6. #31
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    Sorry for the thread necromancy, but I wanted to post on this thread, because it has the most apt title from my search results, if that makes any sense.

    I know we've discussed playing with the time stamp on some brushes and also using solids to fill in arches for better cell count....https://www.ttlg.com/forums/showthread.php?t=145894

    ....but I have another question. Psych0sis mentioned to me that he looked at The Power of Suggestion in Dromed once, and noticed that instead of carving a big airbrush, and then making buildings inside that airbrush out of solids, I tend to carve streets with fill air brushes in the solid, and texture one side of the street with building A texture and the other side of the street with building B texture. In other words, most of my building is carving the empty space with fill air out of a solid Dromed universe.

    He said I would probably get better cell counts by doing the opposite, making a big fill air brush and filling in solid buildings.

    I was looking at DCE mission 17 today to see if the red awnings in it were brush or object, so while I was in there, I checked, and yep, Yandros made his buildings out of fill solid.

    As far as lower cell count is concerned, have I been doing it backwards all these years?

  7. #32
    Member
    Registered: Oct 2018
    Location: Melbourne, Australia
    I made this mistake too, Don. The reason everybody advised using the 'carve-out' method in the OldDark days was because it resulted in fewer on-screen polys, and of course the game would almost always crash if you exceeded 1024. Considering the on-screen poly limit in NewDark is 20480, this is no longer an issue.

    Building with solid brushes is usually much better for the cell count, especially for building large cities and streets, but I've discovered that it isn't always better -- for example, sometimes carving out stairs is about twice as efficient as building them out of solid brushes, and sometimes it isn't. However, unless you want to exhaustively test making all your geometry both ways to see which is better for every single instance, it is better to err on making giant airbrushes and building everything out of solids.

    If you look at 20th Anniversary missions like Sound of A Burrick In A Room, Whistling of The Gears and Lost Among The Forsaken, they are almost entirely made out of solid brushes, including the terrain. That's partly how they managed to fit so much detailed architecture into ~32700 cells.

  8. #33
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    Ok, I guess for the OldDark it was ok. I'll just start using more solid architecture from now on.

  9. #34
    Member
    Registered: Apr 2016
    Not to mention that with fill air, you need fewer brushes. For example, making a room with trims at the top and bottom only requires 3 air brushes stacked on top of each other. For OldDark and its limit of ~7000 brushes, it was also a big deal. Nowadays, it's better to make those trims out of solid brushes.

    But the real cell saver is the time value. I recently reclaimed about 1000 cells but "to end"-ing just eight narrow but tall brushes. The general rule here, as far as I can tell, is that you want your biggest brushes to have the lowest values and the smallest one to have the highest values.

  10. #35
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    Quote Originally Posted by marbleman View Post
    Not to mention that with fill air, you need fewer brushes. For example, making a room with trims at the top and bottom only requires 3 air brushes stacked on top of each other. For OldDark and its limit of ~7000 brushes, it was also a big deal. Nowadays, it's better to make those trims out of solid brushes.

    But the real cell saver is the time value. I recently reclaimed about 1000 cells but "to end"-ing just eight narrow but tall brushes. The general rule here, as far as I can tell, is that you want your biggest brushes to have the lowest values and the smallest one to have the highest values.
    Yeah, Unna clued me into that last time I worked on this mission. I experimented, checking the number of cells on optimization each change. By manipulating the time on 114 brushes, I went from 32,585 cells down to 23,800 cells.

  11. #36
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    Tonight, I replaced some more fill-air brush textures against walls with Yandros' res_cube, changed some brush columns with nicked's marble column, brush awnings with object awnings, etc. Got it down so far from exceeding cell limit down to 22962 so far. I think I can do a lot more. Does anyone volunteer to make objects for authors? I have a couple of roof top air conditioning units made from a mess of brushes. I have a feeling if someone made object based on screenshots, that would save me a lot of cells.

  12. #37
    Member
    Registered: May 2017
    Location: USA
    Does relative brush timing matter, even for non-intersecting brushes?

  13. #38
    Member
    Registered: Oct 2018
    Location: Melbourne, Australia
    Quote Originally Posted by trefoilknot View Post
    Does relative brush timing matter, even for non-intersecting brushes?
    From a practical standpoint, it's often difficult to tell. The main thing you want to do is have large airbrushes as early as possible, and complex solids as late as possible. Other than that, I've found that usually*, you'll get the same reduction/increase in cells if you set time earlier or later than the original time, not the exact time you set. So long as changing the time doesn't visually affect/ruin your geometry, you should try experimenting with earlier/later timings on as many brushes as possible. You never know when an innocuous little brush you ignored might save you ~100 cells if you set it earlier or later than its original time.

    I'm sure somebody with more knowledge about csgmerge operations could describe how it actually works. Sometimes setting an airbrush TO END will actually reduce the cells, sometimes it increases it (without visibly affecting geometry, of course). It's impossible to intuit, even if you think you understand some general rule of thumb.

    *in some cases, this might not be true, and unless you're willing to try every single terrain brush time to maximize savings, I suggest you stick to earlier/later testing.

  14. #39
    Member
    Registered: Apr 2016
    Quote Originally Posted by DirkBogan View Post
    The main thing you want to do is have large airbrushes as early as possible, and complex solids as late as possible.
    From what I can tell, you want your largest solids to be placed as early as possible as well, pretty much right after the big airbrushes.

  15. #40
    Member
    Registered: Oct 2018
    Location: Melbourne, Australia
    Quote Originally Posted by marbleman View Post
    you want your largest solids to be placed as early as possible as well, pretty much right after the big airbrushes.
    I think it depends. If we're talking about a large solid cube brush, the rule usually holds. If we're talking multiple intersecting cylinders, it seems not to hold as well -- sometimes complex solids cost less cells when timed earlier, often they cost less when timed later. Although there is some relation between brush size and cell count/polygon, higher brush sides and intersections usually produce more cells than even the biggest simple cube. The only rule I found that always produces lower cells is having your large (exterior) airbrushes as early as possible. It's also more convenient for construction, so you don't accidentally erase low-time solid brushwork.

    Again, all these observations are totally conditional. It's best for authors to test for themselves, and experiment with these rules of thumb. You never know when an edge-case will save you a few hundred cells.

  16. #41
    Member
    Registered: Apr 2016
    Right, I only have giant solid cubes in the middle of a huge air brush in my mission so far, so that's where my observation comes from. Indeed, it is better for each author to test this themselves.

    Quote Originally Posted by uncadonego View Post
    Does anyone volunteer to make objects for authors? I have a couple of roof top air conditioning units made from a mess of brushes. I have a feeling if someone made object based on screenshots, that would save me a lot of cells.
    I think there was a thread for that somewhere but I can't seem to find it. You can also check the object repository.

  17. #42
    Member
    Registered: Apr 2011
    Location: Montpellier, France
    To see the difference between time numbers you need to activate the cell mode in DromEd, it'll show you all the BSP splits caused by brushes. Sometimes you will gain/lose cells even if the geometry is absolutely identical in both cases, just because the engine splits it in a different way.

    Another thing to take into account as well is that a brush's face that is flush with another face (like a solid face on an air face) will not produce any cells. This is also the case if you have several brushes stacked on top of each other that have the exact same X and Y dimensions for example. This is why almost none of the roofs have a protrusion in The Sound of a Burrick in a Room.

  18. #43
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    I have a dome made of a(i don't know right now off hand...)sided pyramid fill-air, then a solid cube part way , up, then another fill-air pyramid at a slighty different angle starting at the point where the solid cube truncated the first pyramid, etc, etc, until it looks dome-like. I thought I'd eliminate all that brush work and just use one fill-air cube and put a dome object instead.

    Seems weird, but I tried it, and cells jumped over 300.

    Would anyone have expected that?

    BTW, I checked a lot of object sites, including the repository, but didn't find an air conditioning unit that matches my brush work even remotely.

    Addendum: one thing that didn't occur to me to try with the dome situation was to change the time of the big air cube space for the dome to the beginning.......maybe that would have helped?
    Last edited by uncadonego; 20th Aug 2019 at 15:51.

  19. #44
    Member
    Registered: Apr 2016
    Found it.

    Here's the object request thread. It's worth trying your luck here.

  20. #45
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    Ah! Thanks!

  21. #46
    Member
    Registered: Oct 2017
    Like I said in the PM, perhaps just try to remake a side block or 2 out of solids within a big enough air brush for them, I can guarantee it would save quite a lot of cells , and get you ever closer to finished without needed to remove your wonderful brush furniture or build style

  22. #47
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    I've been working on it for a few days now. I'm down to 22842 cells.

  23. #48
    Member
    Registered: Jan 2001
    Location: Formby, NW England
    If you have any circular buildings or arches made with 10+ sided cylinders, try replacing them with 8 sided, aligned by sides.

    In the OldDark days, when 10 was the limit, I didn't like 10 sided cylinders because they looked like angular things pretending to be curved, whereas an octagon looks nice if aligned by sides.

  24. #49
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    On one fake doorway with an arch entrance, I deleted an air cylinder and replaced it with solid wedges. It saved 3 cells. Hard gains there. Every arch will save a lot, but wow...

  25. #50
    Member
    Registered: Dec 2001
    Location: Dromed Detention Room
    Small changes: I made a quarter unit high airbrush, long and narrow, textured to look like a drain at the bottom of a ramp. One face only touched the ground. Eliminating it reduces the cell count by four.

Page 2 of 4 FirstFirst 1234 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •