Jump to content

SCO

Members
  • Content Count

    29
  • Joined

  • Last visited

About SCO

  1. BTW, here is the paper and repository: https://github.com/xinntao/ESRGAN
  2. Yeah. I also notice that even on 'good' results, the NN can get confused about what material some of the textures are supposed to be. I think this sort of stuff will have to lean heavily in the training sets and many unusual environments or perspectives or multiple depth scenes might get screwed by a one size fits all approach. For instance on that longest journey screenshot above, the bricks and vine on that building on the background appear not to be very brick or vine like on one side of the wall, then the wall turns into obvious bricks again on the other. I'm actually surprised
  3. It's not a runtime method. That's why i mentioned EE. Or gemrb, now that i think about it, but i don't think gemrb has a zoom in function yet. This method appears to work better if the image is already reasonably detailed and 'true life' related (because the extrapolations attempt to match textures). If it's small enough resolution or 'artistic' enough or both the results can be disastrous: But if it's 'orderly' enough, they appear to work great (i wish scummvm and dosbox had generic texture replacement): http://screenshotcomparison.com/comparison/127742 (this is a lon
  4. It seems like the new methods of texture upscaling with neural networks are significantly better at extrapolating details than the previous geometric methods. I'd love if the enhanced editions got this applied to their area and npc textures (with some QA ofc). Especially because of the zoom in. These comparisons work better if you zoom in to see what happens at non-standard zooms (like the EE). (original) vs
  5. Well i don't think you have to merge it yourself even if you are a github 'collaborator'. Just that you can. I'd advice to have everyone involved (or at least one person) you trust to be collaborators because even if it's policy to have a code review before merging, having multiple persons with the ability might prevent unacceptable delays, say if the 'owner' is on vacation without the internet, or other less benign reasons.
  6. I think you should simply add the 'people you trust' who are collaborators to the list of people who can merge pull requests. In effect, this is Roxanne. If someone else wants to take up maintenance, say a author, they can ask to be added. It's on 'settings->collaborators' and add her github account id or email if you own the project, iirc. I don't know if other people can be made owners, or if normal collaborators can invite others, or if that requires them to fork the project into their own page.
  7. Who wants to do work on weidu itself? Honestly. For posterity, and because i don't like to only bitch (in spite of not being able to install BWS at all, because i'm on linux) this is my strategy to make a verification script work if weidu gets the ability to verify compatibility constraints on a list of tp2. 1. download the mod tp2 file in order not to have to download the whole mod. This only really works for github or similar mods, but i'd gladly be restricted to that over the jungle of terribly packaged mods nowadays. Github has a ways to get the raw file from a repository, and no,
  8. BWS simply needs to be replaced. Various people wrote good suggestions for it. I for one think that the rules on it should be translated to native WEIDU tp2 directives of the mods, put on the fix pack, or another 'verificationpack' - if upstream doesn't accept patches - and optionally verified by weidu itself with a command. Then it would be pretty easy to replace it. The parts that are actual code and verification for EET, BGT etc, can be coded elsewhere better, it's this crucial feature and data that needs to be preserved. I was thinking of: weidu --verify [listofselected (mod_tp2, com
  9. I think i was one of the ones that originally requested something like this (or it was that mod to auto resurrect if ressurect is available, end game if not). Pity about the dialog bugs, i was expecting biff the understudy to take over without problems. I wonder if the feature can be brought into the EE engine.
×
×
  • Create New...