Jump to content

The BSA Project


dharh

Recommended Posts

that's a really good idea, packaging everything just ensures that files stay as they are supposed to. for instance, i have Pyfiied nearly every single mesh in vanilla oblivion. (some exceptions, thank's to tomlong's excellent tutorial on the matter), then rebuilt the stock meshes.bsa including all the unofficial patches pertinent, and REVWD highly optimized lod meshes. [then compress all, medium] so now i do not have to worry on mods overwriting these meshes. and causing lag.

 

next thing i did was to Pyfii and BSA all of FCOM resources (that's the absolute full package, with all optional mods.) all of my weapons, armors, clothes, are also Pyfiied,tinkered with (correct some stretched texture here and there, add egm to helmets needing them, correcting normal maps, etc.) and into a single BSA. things such as unique landscapes,region revive, trails of cyrodiil,roads and bridges, and such also get into a single bsa. castles (i have a LOT of those) and other houses also get packaged.

 

so why do that? to ease reinstalls! it is a real pain to reinstall oblivion when you are reaching 400 mods. the only problem i see about that is updates...

you have to merge the esp updates, and rebuild the bsa every now and then. but really really saves disk space and defragmentation time while maintaining the same average load times in game.

 

So what's left outside? QTP... and am really thinking on merging that one into textures-compressed.bsa too. of course, that leaves me with about 10 BIG BSA files, about 50 BIG esp and over a hundred small sub-megabyte mods and patches....plus the bashed patch. total space? 21gb.

total space before doing all that? 39gb plus more than 350 assorted mods and patches that were mostly causing lag and serious dificulties every time i rebuilt my bashed patch. (your load list is overloaded!, blah!)

 

 

but i can tell you this. what REALLY made a difference was 1) Undeleting and Disabling EVERYTHING in all of the mods installed. and 2) Pyfiing EVERYTHING from ALL mods!. I hardly ever have a crash any more. and average 30fps wether indoors or out. 20fps while in combat and i have not that great of a machine either!, i mean, it's not bad, but i could be much much better.

 

 

so good luck in your project. just wanted to let you know my remarks.

 

Thanks for the comment. That is exactly what I intend to do. Act like the unofficial patches are official and patch them into the BSA files. Sounds like I have a long ways to go before I get anywhere near your mod collection, but you're proof that using BSAs helps quite a bit. And yeah pyffing stuff seems to have been making a huge difference for me already and I haven't even done the official BSAs yet.

 

I'm curious how far you go with pyffing meshes as from the stuff i've read certain meshes can be optimized quite a bit (Architecture, Clutter, Dungeons, Furniture, Landscape, Plants, Rocks, Sky), others not so much, and even some (hair, helmet, some creatures, some clothes) can get bugged out.

 

luckily, pyfiing is no rocket science, you just right click a folder and select "optimize with pyfii", all the rest is (hopefully) managed by the script. as for obscure parameters and modifiers, i leave that to more well versed people, such as tomlong. but for the average mesh, i think it manages quite well on it's own (just make sure you have the latest version and you're good.) as it will automagically skip meshes with helmets and hairs, and "most" of the types of meshes that are known to create problems. for my experience, though, i can advise you to NEVER EVER pyfii landscape meshes. it will just create huge holes in the lanscape. (for reasons i am yet to learn). besides i was told to never pyfii animated meshes, such as bows, fires, doors, and some few other things, lest they lose any animation embedded... but regardlesss (hey, the best way to learn is skewing thing up) i have successfully cleaned animated meshes with no adverse results whatsoever... but i think other people can tell you better either if this is advisable or not.

regarding hair... hairs usually come very clean from their creators, so i have had no problems other than fixing the absolute texture paths they for some reason always use. helmets are another story... because not all helmets are clean (specially custom modeled ones) and they usually don't clean them because people say pyfiing helmets is a bad thing. i think otherwise. (specially after stripping 3mb of junk data off a helmet[which i will not name so the owner won't take offense or something]) so what do i do? i first check the mesh on nifskope. and do a "remove bogus nodes" if it does remove anything, then the mesh is dirty. so i delete the egm, pyfii and then re-create the egm with the conformulator. sure it takes some time, specially if you have the ammount of armor sets that i have but it does pay. consider i have crowded cities, crowded roads,cm partners,tamriel travelers, npc with jobs,tie4mods, and all the most well known companions. and ALL of them make use of the custom armors clothes and weapons i have installed. (uh yes, i do some modding of my own) and i'm always surrounded by quite a lot of npc, all with different armor sets. so having those a clean as possible assures that i will not run into any lag at the least desireable moment.

 

 

how small is the most i hava reduced a mesh by pyfiing, you ask? well... again i wouldnt want to point the dirtiest mesh i've ever cleaned for certainly its owner would not be pleased *cof,curp,cof*!...sorry. but i do can tell you i started with an 9Mb file, and after an hour or so i ended up with a 3.5Mb one with no noticeable loss in quality. sure there are some huge filesize drops here and there, but for the most part you want to look at the strip lenght. that (imho) will help your FPS the most. the smaller the strip lenght the faster that mesh will render. and when pyfii tells you that you should consider simplifying a mesh or breaking it up on smaller parts... do so! (unless, of course it is a helmet) otherwise that single mesh will start eating your fps like cornflakes.

Link to comment
Share on other sites

The latest versions of PyFFI skip files with EGMs (i.e., hairs, hoods, helmets) by default. You can read up on PyFFI-optimization in the opening post of my site's thread. (See the second link in my sig.)
Link to comment
Share on other sites

The latest versions of PyFFI skip files with EGMs (i.e., hairs, hoods, helmets) by default. You can read up on PyFFI-optimization in the opening post of my site's thread. (See the second link in my sig.)

 

do you think it is really unadvisable to pyifii a file with EGM? even when those are so easily re generated with comformulator? i ask because some of the helmets we download have no optimization whatsoever. they come directly from 3dsmax (often with outdated exporters that leave tremendous ammounts of junk on the files) so, do you know of any side effects on doing it this way?

 

 

i have had no problems so far with this, but you're the expert!

Link to comment
Share on other sites

The latest versions of PyFFI skip files with EGMs (i.e., hairs, hoods, helmets) by default. You can read up on PyFFI-optimization in the opening post of my site's thread. (See the second link in my sig.)

 

do you think it is really unadvisable to pyifii a file with EGM? even when those are so easily re generated with comformulator? i ask because some of the helmets we download have no optimization whatsoever. they come directly from 3dsmax (often with outdated exporters that leave tremendous ammounts of junk on the files) so, do you know of any side effects on doing it this way?

 

 

i have had no problems so far with this, but you're the expert!

I am by no means a PyFFI expert, haha. The things with PyFFI (or advertising its use) is that you have to take the experience level and the patience of the average user into consideration. Bad meshes may not show up right away, and many users would not want to spend hours testing all of them. Skipping EGMs requires the least work on the users part. If you are willing to optimize, test and then fix the issue, by all means, go ahead. I have not heard of PyFFI-optimization gone awry being really harmful. If the optimization of a certain mesh fails, replace it with the original, and keep going...
Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...