twistking

Hello, Since ED is currently prototyping new ATC and presumably other AI interactions, I hope I can provide some useful suggestions. I'm very critical of LLM "AI," but for video games - especially dynamic and immersive simulation environments like DCS - the technology is, in my opinion, extremely useful. To my surprise, the technology has now reached a point where it is also technically viable for a product like DCS, with very lightweight and fast models that can run locally on client machines, thereby solving a lot of economic obstacles. Also, most of these models are open source and relatively easy to implement. The cornerstone would be TTS/STT on the local machine, meaning speech generation and natural speech processing from the player (through a microphone). Models are now fast enough to run alongside DCS without significantly impacting performance. However, minimum system requirements would increase, as at least 2 CPU cores and a few GB of RAM would need to be dedicated to this task. This approach would not allow free-flowing conversations with the AI. The logic would still be scripted. A full LLM capable of handling open-ended conversation with the player would require a high-end workstation PC if run alongside DCS. However, models are becoming more efficient and average player PCs are becoming more powerful. ED should therefore anticipate the feasibility of locally running LLMs to further enhance the experience. In the meantime, ED could create an API for external LLMs to hook into their system, not to completely take over the logic but to augment the scripted logic with flair, personality and more dynamic conversations. Such an external LLM could be cloud-hosted (e.g., using the user's paid account with providers like OpenAI) or locally hosted, either on the player's machine or on a dedicated PC on the network. The proposed system would be fully multiplayer-compatible, since all communication with the AI is exchanged as text, while TTS and STT are handled locally on each client. If a larger LLM is desired for more free-flowing conversation, it could be run either on the server or in the cloud, with the server API managing the interaction between the game and the LLM. Thanks for reading. PS: One major advantage of this hybrid approach is that the LLM component could be easily maintained and expanded by the community. With a clean API separation, there would be no security risks for multiplayer or core gameplay, but plenty of room for creative extensions. For example, different LLM “personalities” could be developed - from a strict flight instructor helping beginners step by step, to a more casual RIO adding flavour and banter to the experience. Neither development nor maintenance of these LLMs would be on ED, as long as they provide a robust API for the external LLM to hook into.

what is AV in this context?

You are right. All i found was this image: https://de.wikipedia.org/wiki/Datei:Pont_de_Solferino_-_Les_Travaux_Publics_de_la_France.jpg Looks like every other Seine bridge from that time...

This bridge was constructed between 1997 and 1999. I'm confused. What am i missing?

Fair. Could be interesting as AI aircraft though...

I think only Nevada and Caucasus still run on baked lights to some degree. On both maps the airfield ramp lighting is baked. Not sure about street lights though...

Can we get this acknowledged, please? The parachute flares should illuminate the cockpit correctly Their colour tint should be corrected Their brightness should be reduced. They should also receive a little bloom/halo effect similar to flare countermeasures

I propose an editor "module" or function that would allow mission designers to auto-generate civilian air traffic for missions. Designers could designate airports or map edge points (portals) between which traffic routes would be automatically populated. The function should furthermore allow basic control over air traffic density, unit types, randomization etc. The "module" should also allow control through triggers to change parameters during runtime. With such a system mission designers should be able to create believable civilian air traffic with only a few clicks, while still allowing enough control for the system to adapt during mission runtime. If you like the proposed improvements, please rate this thread 5 stars: Thanks!

would also be useful for airfields to go dark when under air attack.

Oh, right. i forgot about the promised ATC upgrade. PAR would be a good fit for that indeed.

Things like this shouldn't be module specific imho. They should be rolled out for every AC similar to the canopy rain effects. That said, i agree it would be a very immersive feature.

That would be very nice to have indeed. I wonder though if PAR would make even more sense. I think in a military context these systems were also available as mobile units. They would also allow AC without ILS to do IFR approaches. Would only need a bit more work to implement unfortunately...

Can we get an update on one of the biggest issues with the current weather system, please? AI can still see through clouds, IR missiles can still track through clouds. None of the visually interesting cloud presets really work for singleplayer and cooperative play. WWII and modern ground attack are notoriously frustrating without clouds blocking AI's vision and weapons/sensors. Heatseekers (especially IR SAM) have an unfair advantage that flat out prevents certain scenarios to be playable in certain weather conditions. I would appreciate an update on the matter. I think it's one of the bigger long-standing issues DCS is plagued by.

A simple way to get some weather changes is indeed very much needed, but i'm not a fan of your particular solution. Wouldn't it be more elegant to define sizes of weather patterns instead of timings? For example you define two weather presets with their own cloud, temperature, pressure and wind settings. Then you select the size of each pattern's "tile". DCS would then move those tiles around the map depending on some simplified wind speed calculations. At the tile's borders the weather parameters would be interpolated between the two tiles. Also edges would not be straight lines but more fractal to hide a tiled appearance. So this would NOT be a weather simulation, but a simplified system that allowed 2 (or more) defined weather states at different points at the same time. Moving tiles would also allow auto-generated weather forecasts without doing a lot of calculations. It would just estimated tile position at requested position based on those simplified windspeed calculations. Surely not as good as a true weather simulation, but much easier to implement and imho more immersive and interesting than a timer-based global switch between systems.

I don't want to derail that thread, but in theory every light source gives you blooming. It should be directly proportional to perceived size of the source and perceived brightness of the source (and the condition of your cornea etc). Atmospheric conditions can of course further add to this. Even high quality photographic lenses struggle to render point light sources at their geometrically correct size. Your eye is optically inferior by a long shot. The issue in DCS is obvious. The implementation is just done without care. It's inconsistent between aircraft and even on aircraft where it's less bad it still scales weirdly and vanishes too early. Other light sources (city lights, the ships on PG maps, airport lights) look beautiful, have nice bloom, scale well and have acceptable visibility range - technically still too short, but "good enough" for gameplay purposes in most cases.

No, adding bloom to point light sources is the industry standard to show brightness on contrast limited displays. And you don't need bad eyesight to perceive this in the real world too: The sun is actually a rather small dot in the sky. So some "fake" blooming is needed in graphics to give a realistic impression of a point light source. The issue with DCS is, that aircraft nav lights are just badly done. That's all unfortunately. It's not just the weird bloom (that scales incorrectly with distance), it's also the lack of rendering distance. In reality the lights should be visible over 50km or so. I would be happy with half the distance already... currently lights from aircraft vanish after few hundreds of meters. It looks unrealistic, but most importantly it makes night flying unrealistically difficult and frustrating.

Select the "cold start" option and it should appear a "uncontrolled" checkbox. Click it and aircraft should remain parked. You can even activate it with a trigger at a later stage.

Back to topic: The thing that i want to see from a dynamic campaign above everything else is modularity. I'd be hyped to play a cool dynamic campaign that's delivered "as-is", but i get even more excited by the idea of building my own little dynamic missions/operations without needing to script advanced AI routines, C2 and logistics... Instead of setting up different enemy flights and creating a ton of triggers, i could hopefully just select an opposing strategic element (f.e. a fighter wing), give it a home airbase, set strength in aircraft numbers, weapons and manpower, set stance (aggressive, defensive etc.) hook it up to logistics (resupply) and give it one or multiple taskings. Then only add friendly forces and mission goals. Granted this would not yet make a grand strategic campaign, but it would allow everyone to create a dynamic mission with near endless replayability in a few minutes.

Fair enough. I still think that even static crew (no animations other than perhaps separate dead/alive states) would be much better than unmanned guns. Might look weird in promotional close-up videos, but would provide a much better game play experience.

Thanks for the info. I was suspecting that being the case, but good to have confirmation.

Alright. Got it!

I would love such feature. Starting the mission outside of the aircraft and being able to do a walkaround would be super immersive. Also getting out of the pit after a mission to inspect the battle damage you've taken would be cool. Or maybe you walk to the runway to see your buddies land after a MP sortie... Would also gel well with the super carrier... Proper animations for climbing into (and out of) the cockpit would be icing on the cake, but even if it would just fade to/from black would be ok, if that would significantly simplify the development of such a feature...

Not sure if joking or not, but since the release of those AAA guns, ED has released a lot of extremely high detail ground units; a lot of those being way less relevant for actual gameplay. Those man-hours should have been allocated to finishing what's already released. And this is a case where man-hours should be equivalent: I assume that modelling and animations are done by the same people. If ED had fulltime animators, we would not have that awkward situation in the first place...

After having seen Enigma's video i'm also a little bit concerned. I'd rather have a dynamic campaign of limited strategic scale, than have a feature-limited campaign engine that cannot be adapted to MP, different maps and different eras. To be more precise, i would be cool with a first iteration of such a "campaign engine" being limited to a tactical or operational level instead of a strategic one. So i guess it would be -strictly speaking- more of a dynamic operation than a campaign. This would mean in-game time-spans of days instead of weeks or months and it would free the AI from needing to do strategic decisions as the strategic situation could still be prepared manually by a mission designer. On the other hand the campaign engine (and related AI logic and tools) should be as flexible and robust as possible, including support for cooperative play and adaptability to other eras and all maps! With further iterations to the campaign engine, the scale of those operations could eventually increase from a tactical/operational to a strategic scale, similar to that other game. This would require the game engine to handle more units and the AI to react on a strategic scale. Only with that approach could that massively anticipated feature improve all types of DCS missions (through better AI and more robust game logic for logistics, C2, etc.). I think a narrowly designed campaign with "proprietary" systems without modularity and adaptability would be a humongous waste of time for ED and a big disappointment for the community.

The DCS ecosystem is lacking some civilian aircraft. I suggest to add at least one of the following AC to cater to most scenarios A vintage airliner (f.e. wide-body trijet of some kind) A vintage business jet and/or regional airliner A modern-ish narrow body airliner (with liveries to cover ~1970-today) A modern-ish wide body airliner (preferably a 747, again with liveries covering ~1970-today) A modern-ish business jet and/or regional turboprop airliner (again with liveries to cover different eras) With a smart selection of types, even a few models can cover a lot of eras due to the very long careers of some aircraft. Some types could also appear in military versions, like a 747 "Air Force One". Thanks for considering this suggestion!