OnlyTaffingCowardsHide

Oh neat, so can we disable depth writes for the heat haze without breaking things? The problem with the in game AO is that it's based on technology that is quite deprecated now, and could even be called performance heavy for what it is. Reshade can be heavy on Opengl, but I've since upgraded to an RX6600 and I'm seeing excellent results enabling modern shading techniques in TDM without significant performance loss. Not just better ambient occlusion but also global illumination, ambient light, and SMAA/Temporal AA solutions are available that work better and offer better performance. I can imagine higher end rigs can get 100+ frames no problem with Reshade and it does end up looking a lot better. I'm ready to upload the mod but if there was something that can be done to fix this last remaining issue...

Specifically around fires and torches in the game, they show up as strange cascading blocks/arrows in the buffer. It's making it very difficult to apply Ambient Occlusion and Global Illumination effects without it looking weird on flames.

Fair enough, and the shader pipeline while they're at it, cause the Ambient Occlusion shader looks right out of 2008. I tried replacing it with a proper modern method via reshade, but some dumbass turned the depth buffer around torches/fire pits into a cascading arrow, so reshade doesn't even work properly.

...

Fine, if people don't want nice things who am I to argue. Maybe I've got it all wrong, maybe TDM engine is very well optimized but the quality of the average mapper is the only issue. Maybe most of the mappers just need to git gud... You certainly wouldn't want an engine that's more forgiving considering the average hobbyist knows jack about optimizing anything...

Well, I have faith in greed so I don't see why that would ever happen. They want people making games because that's a potential product. Now that every corpo is doing the profit sharing model any corpos that opt out of that model are going to fail overnight. How is what I'm suggesting not a thief clone in Cryengine? Why have people learn a new editor if the files the current editor creates can be made compatible? I mean sure, we can start by just making a thief clone and letting people muck around in Cryengine to hack together missions, but I don't see why the scope should be limited to a hacked together solution.

Well, Unreal is generally trash without significant optimizations. Unity may be viable, as it's certainly shown promise in Daggerfall Unity. and yes there is a lot of work involved. I'm sure there was a lot of work involved in creating TDM in the first place, why not just stick with Dark Engine? It's not really about lighting, it's about making a contribution that everyone can benefit from. Doing the work now so that mappers can save a lot of time overall. Obviously cryengine has superior lighting techniques, but the more important aspect is being able to use more light and shadow casters without performance dropping off a cliff. Largely this is due to using modern gpu resources far more effectively rather than relying on cpu stencils or outdated hardware shadow maps.

It wouldn't be released under the TDM license, it would be a standalone tool that is merely compatible with TDM mission files, and would ideally use stock TDM assets. Whether those assets are packaged with the installer or pulled in from an existing TDM installation is another matter. The editor does exist... it's called Dark Radiant. The idea is to keep the editor and only replace the underlying engine. In the style of OpenMW or Oblivion Remastered, compatibility with existing files is key.

It's not a commercial project, what terms are even relevant to us? I'm literally suggesting a free to use tool that would convert all existing assets to a modern engine... and still all we can do is complain... I love this community, cynicism on a level that Garrett could only aspire to Yes there is a chance that Crytek could shut it down, if there are examples of such a thing happening to non commercial projects I'm happy to hear about it, but ultimately if it does get shutdown you would lose nothing, as mappers would continue to map in Dark Radiant anyways.

It depends on how much can be handled by a conversion tool. Ideally people can just keep working in Dark Radiant and convert it over. I don't think there would be much point if mission files couldn't be made compatible, as there is so much content available already...

I'm looking into it, no promises but maybe I can cobble together a proof of concept and get some other people on board. Mostly the benefit would be optimization and utilization of modern hardware. Modern engines have a lot more options to optimize large open spaces with high object counts. Currently mappers do a lot of extra work to get around this optimization problem using time consuming/outdated techniques that a modern engine can handle automatically, or by crippling their artistic vision to limit draw distances and object counts. Even with optimization, you can never really overcome engine limits without a massive performance hit. It would save experienced mappers quite a bit of time, and newbie mappers even more time and headaches, as they wouldn't need to pay as much attention to optimizing for engine limitations. There are some other fringe benefits, like access to more modern shading techniques. For example I recently hit a wall in trying to implement reshade for proper ground truth ambient occlusion and global illumination, as the Dark Mod's depth buffer includes some erroneous information around torches/fires. Having those shading techniques in engine would be quite an improvement, rather than hacking them in via reshade. Lighting in general should be a lot easier to program without tanking performance. I wholly disagree that TDM performs well enough. Claustrophobic mission design is the norm around here, with people who try to step outside these limitations encountering quite a performance train wreck if they don't know all the advanced tricks that Cryengine can handle much more efficiently. You can find examples of large open spaces with details, but these are the result of veteran mappers who know how to get the most out of very little, which just encourages gate keeping and fewer missions being made. There are plenty examples of TDM's pitiful default rendering of large areas, one in particular set on a bridge stands out...

Now that Cryengine is basically free to use for community projects, and has proven itself to be ideal for large open worlds with stealth gameplay like Kingdom Come Deliverance, is there anything stopping anyone from porting TDM over with support for existing fan missions? I mean legally speaking of course, as I'm sure it would be a lot of work, but there seem to be fewer barriers than ever before. It's incredible what people have done with an engine that was built to accommodate tiny corridors on ancient hardware, but imagine an engine purpose built to render large open spaces... No more compromised visions when mapping, no more endless optimization to fit a square peg through a circular hole... And if the mission files were to be compatible with the new engine, people could just keep on mapping in DR as if nothing even changed. Cryengine Community Edition https://engine.pterosoftstudio.com/

I don't care in the slightest what devs intend on the topic of subjective details. TDM devs wanted most of the lighting in the game to look oversaturated with piss yellow and nuclear blue lighting, but im changing that because i'm the one playing my copy of the game, and my subjective opinion is all that matters to me in that context. Just because you integrate sliders doesn't mean there won't be a default setting. I doubt you live under a rock, if you've read reviews you're obviously aware they praise the sound design with the caveat of the main characters inappropriate tap shoes...

Of course they can't get it right, there is no "right". The single most common complaint about Thief 1 & 2 is the footstep sounds

At this point most textures are looking pretty dated compared to any games released in the last 5-10 years, it's more the exception when a texture looks really good. I wonder if running all the base texture through an a.i upscaler would be worthwhile. Then it could be downloaded as an optional hd add-on.