So, if you started a flight sim from scratch, a new project, what are your must have features?

[Mysticpuma]

6 hours ago, krraatumi said:

A feature I REALLY want for SP is the ability to switch to an AI plane and take control of it, like in DCS.

And Cliffs of Dover does this too. 

[Livai]

On 3/20/2023 at 12:43 AM, BlitzPig_EL said:

• All at an affordable price too?

 

On 3/20/2023 at 1:01 AM, BraveSirRobin said:

• With smooth frame rates on a 2010 PC?

 

Previously it was quite normal for new games with groundbreaking graphics to run smoothly after 1-2 chip generations!!!
But today the developers say to us that it's not worth the cost. Seriously?? - With this attitude we would still be with the stone tablets....

[BraveSirRobin]

5 hours ago, Livai said:

 

Previously it was quite normal for new games with groundbreaking graphics to run smoothly after 1-2 chip generations!!!
But today the developers say to us that it's not worth the cost. Seriously?? - With this attitude we would still be with the stone tablets....

I have no idea what you are talking about.  You need to add some context to link sentence 1 to the rest of your post, because to someone who can’t hear the voices in your head they appear completely unrelated.

[Livai]

On 4/24/2023 at 10:54 PM, BraveSirRobin said:

I have no idea what you are talking about.  You need to add some context to link sentence 1 to the rest of your post, because to someone who can’t hear the voices in your head they appear completely unrelated.

 

My Context was my first post what you quoted, plain and simple.

Is this your context -> Nowadays nobody feels like playing below 120 fps. Unplayable, totally laggy! Ultra HD and playable on a 2010 PC must of course also be. And pretty, because graphics are important. But woe, the game costs money or stutter, was this what you wanted to say with your Quote "With smooth frame rates on a 2010 PC?"

What are smooth frame rates anyway - 24,30,60,120,144,165,240,360,540 fps. 
You watch Blurays with 24 FPS. Console Gamers enjoy games with 30 FPS.

Previously it was quite normal to lower the screen resolution for more FPS because games on LOW/Medium Settings looked awful.
Nowadays you lower the Graphic Settings for more FPS because Games look already good at LOW/Medium Settings.

Example for run smoothly after 1-2 chip generations = Crysis 1 released with groundbreaking graphics you needed two 8800 GTX Ultra to play it with 60 FPS @ 1080p.

 

When a sim release with groundbreaking graphics where you need two 4090 RTX, where is the problem? Playing Games is our hobby and if you have nice games with nice graphics it is very normal to buy better hardware. Games with groundbreaking graphics support also all the Hardware producer because Gamer need new faster Hardware for these groundbreaking graphics...................

 

[BlitzPig_Bill_Kelso]

Back in the day I had a 69 Ford PintoR  totally tricked out, Tach by Timex, Harmon Kardon Carb kit, slicks on SS rims and those bitchin dangly balls.

Anyways it could go 0-60 in 3.9 seconds on nitroeTM  boost mode. I had a mechanic that really tricked out points hack with a nail file...Mind blown!

Must have boosted my PSI by 13mm.

People used to say "That car can fly Targ" And that's when I first really got into flight sims. 

If I were to have one thing in a new flight sim? Easy!

 

Cupholders in the cockpits!

[Majakowski]

1) If it was from scratch then I would first gather all the things I have missed in earlier flight sim iterations and implement or at least make provisions for them. Such as auxiliary duties like transport, liaison, spotting and reconnaissance and blah blah blah

 

Retrofitting a feature is tedious and prone to hit limitations that were set in stone by neglecting such features in the first part. And in implementing it from the ground up, once it is there, it's there. Also it broadens the future usability of the product.

 

When I measure the game on how historically accurate it is (which is said to be the case for Il2), I can't just ignore vital parts of operational history and not implement certain aspects due to "oh only 15% percent of the players will use it so we leave it out" altogether, this is how we get no drop tanks for example. And yes maybe there were not 15% of all planes in WW2 of the liaison type but they existed nonetheless and did their part for the whole effort.

 

Also some "side aspects" of operations like communications should be modeled for example throwing messenger bags to ground- or sea units to get them to act in a certain way. Catapult functionality for seaplanes comes to mind. Maybe a cruiser won't be in the first release but who knows where the future brings us? To have this feature then will ease up implementation tremendously. Same of course for recovery routines.

 

So it isn't the whole world that needs to be implemented but in a game that runs in a technical timeframe from say 1914 to 1945, at least all the operational realities of aviation at sea and in the air should be provisioned for before even thinking of which planes to implement or how the map should look.

 

We aren't living in 1993 anymore where specialty knowledge of military operational details was confined to and spread within the narrow realm of books in the local library. All the knowledge is readily available (and for the most part quite cheap) through established online sources and even in the player base. Gathering special knowledge today is easier than picking up pebbles from a rocky beach.

 

That said, a combat flight simulation in my eyes would be a framework of routines and structures that allow for most or everything that you would read in the boring parts of a pilot's biography. Flying over the channel and spotting a raft? Fine, circle him, make a report by radio or messenger bag and a rescue unit will catch him. Provisions for this would be made within a powerful communication system. Make other units interactable, that makes for immersion.

 

Spotted a submarine instead of a raft? Fine, report it and the next catapult plane in reach will take off with depth charges through a sleek system of dynamic "mission creation" inside a running mission. The coordinates of sighting will be passed and automatically a search routine for the new unit created. So an entire cycle of flight operation can run autonomously because all the necessary variables are in the system already.

 

If you have covered these cases of what realistically happens in a war on a daily basis, there will never be any discussion of the usefulness of a certain plane, if resources are available, then it just can be done. Because no matter if you decide to introduce a Walrus or an unarmed requisition of a civilian type totally unsuited for dogfight, all the foundations for putting it to a good use will be laid in advance, nobody would have thought of the U-2 when Il2 GB began and yet we have it along with a very narrow spectrum of uses as liaison and artillery spotting aren't implemented, or at least only under great pain. And missions then could be created easily through implemented basic functions instead of mind numbing placement of highly complicated trigger webs that cease to function when you make the slightest unexpected move.

 

So my first and foremost must have would be a solid foundation of what we encounter in reality. Cockpit and graphical fidelity are important too but also they are quite independent from the first as the sober variables do not care for graphic fidelity and can be ported over to a new engine if necessary without loss.

 

2) The second most important must have for me is making it easy for in house developers and the community to make content. Especially missions, so the internal content creation tools should be designed with intuitive usage in mind. Not only does this save time in making native content for experienced staff but also newcomers can be put into operation quite fast. No need for overly complicated make-shift tools that are a pain to use and can only be even remotely comprehended when using it everyday (looking at you, Mr. "create linked entity" although you are one of the easy ones). Here some inspiration can be taken from other content editors like silent hunter where setting up a convoi within randomized spawn circles is as easy as chewing a slice of toast with the crust removed while still having all teeth and a healthy saliva production.

 

Only after on a spherical water map an untextured box floats when specified as floatplane, sinks when not, and flies when the manipulation of all the variables in the background, that a real pilot also must adhere to, tell it to and a data request shows me different speed values of that box, each depending on air densities and altitudes and temperatures, would I care to even think of any plane type or map region to implement.

 

Because from that point on, anything would be possible and not a question of "can I bring this into the game?" but of "When can I bring this into the game".

 

3) After that but certainly of equal importance come things like net code for multiplayer functionality. The topic of singleplayer-playability is covered with the all-encompassing approach of section 1) because then you have no enemies that have you as a waypoint but a dynamic environment, easily configurable through simplified access to the inner workings.

 

4) How then to pay for competent provisions like servers for MP hosting? I think a small fee for online functionality could be introduced, given that many games work on subscription models and buying content. Either monthly subscription or per an on demand basis for players that aren't online all the time. It should not be in the range of 30 or 40 Euros per month, that would be excessive especially for thise that are not fortunate enough to have a ton of money but maybe 5 or 7 Euros and coupled to some extra features that a player then can use like special skins that can be seen by others or some functionalities like reporting enemies on the map for others to see so that the most casual furball crowd still has basic usability cost free and the more invested players can do a little more but no p2w of course.

 

5) Now comes the physical scope. Where does it play? What time? If section 1) is implemented thoroughly, the map would be on a certain part of a huge virtual sphere that represents the earth (not all of it traversible though, but always in relation to that globe). And depending on latitude and longitude for this map would be the range of expected weather or magnetic deviation that comes with all the data provided when laying down the framework.

 

I think that once all the variables necessary for aviation that one encounters in reality are put into the framework, the game can be turned into anything and what is deemed undoable today because of massive expenses that come with retrofitting entire systems and sets of functionalities that aren't remotely privided for, would only be a question of "when is there some free space to quickly put it inbetween".

Edited May 17, 2023 by Majakowski
exchanged circle for sphere

[RossMarBow]

On 5/11/2023 at 6:15 AM, Livai said:

 

My Context was my first post what you quoted, plain and simple.

Is this your context -> Nowadays nobody feels like playing below 120 fps. Unplayable, totally laggy! Ultra HD and playable on a 2010 PC must of course also be. And pretty, because graphics are important. But woe, the game costs money or stutter, was this what you wanted to say with your Quote "With smooth frame rates on a 2010 PC?"

What are smooth frame rates anyway - 24,30,60,120,144,165,240,360,540 fps. 
You watch Blurays with 24 FPS. Console Gamers enjoy games with 30 FPS.

Previously it was quite normal to lower the screen resolution for more FPS because games on LOW/Medium Settings looked awful.
Nowadays you lower the Graphic Settings for more FPS because Games look already good at LOW/Medium Settings.

Example for run smoothly after 1-2 chip generations = Crysis 1 released with groundbreaking graphics you needed two 8800 GTX Ultra to play it with 60 FPS @ 1080p.

 

When a sim release with groundbreaking graphics where you need two 4090 RTX, where is the problem? Playing Games is our hobby and if you have nice games with nice graphics it is very normal to buy better hardware. Games with groundbreaking graphics support also all the Hardware producer because Gamer need new faster Hardware for these groundbreaking graphics...................

 

 

 

I would like to explain some of these concepts better.

 

#1 rule of computing is HARDWARE can not fix SOFTWARE issues

Your assertion that games run better after chips have got faster is just wrong.

Games that run poorly now will always run poorly.

Take this game or Hunt: showdown (crysis engine) for example
Both are CPU limited - throwing a 4090 at the problem won't solve it
And guess what throwing a faster CPU at the problem won't solve it either 
It's a fundamental flaw in the programming

IL-2 was released in 2012?
Hunt showdown is over 5 years old and built by the same people who built Crysis (crysis has always been a performance meme)

You also completely misunderstand what smoothness is 
smoothness is nothing to do with the average frame rate
smoothness is the variation in frame time 

 

For a game to be smooth effort has to be expended to adjust the timing of the code so that spikes in frame times don't happen

For a game to be Esports grade the frame time must be smooth
If you throw dual 4090s at an esport game you will get 2000 smooth fps 

If you throw dual 4090s at ? code you won't get jack

Read my thread here if you want more information

 

[IckyATLAS]

On 5/26/2023 at 5:13 AM, RossMarBow said:

You also completely misunderstand what smoothness is, smoothness is nothing to do with the average frame rate, smoothness is the variation in frame time 

I think we overlook this. You can have an average frame rate that is good but say if you mix frames with some at half the frame time and half at double the frame time then the average will stay the same but it will be very choppy if it jumps here and there.

To ensure same frame time or say constant frame time, you must once you have decided what is your target, say 60/sec then you must know what is the CPU-GPU capable to load or build in a frame time or around 1/60 sec (I make it simple here). So if you have less then you keep the frame in the cache until it is released. But if you need more time then you are in trouble. This means that you must calibrate the scene so that it will not take more than the allowed frame time.

No idea how you can do that besides testing you game scenarios and measure all frame times and get a red flag when one frame exceeds the time. 

You play the game and no red flag is fine. One red flag means you must reduce the CPU-GPU load. It is a combination so you have also to know and measure separately CPU and GPU load to know where you are limited. Is it graphic load or game processing load. 

This means investing a lot of time for these kind of code or scene optimizations. I suppose this will be very costly and this is why they do not do it in general. Maybe only proven blockbusters (I mean in their second or third iterations) can afford doing it with their big buck budgets but this is not always true. 

 

Edited June 1, 2023 by IckyATLAS

[Justicier]

First and foremost... highest priority...

 

1: Dynamic Campaign System. (Or more literally a real-time, fully simulated, dynamic war)

 

Something on the level of Falcon 4.0, or Microsoft CFS3. Often the excuse given by developers these days is that a true dynamic campaign or fully simulated war is too difficult to implement... I tend to believe that is an excuse given for lack of priority. Remember that Falcon 4.0's dynamic campaign, which was more literally a fully simulated war, was designed by a full-time INTERN, for the duration of a summer, with the intention of overlaying a large scale strategy game element over the core Falcon flight simulator. If one person, in that short a period of time, with the limited hardware and software capabilities of the time, can implement the best dynamic campaign system made to date... I'm reasonably sure that a comparable or even better dynamic war can be simulated with modern tech. I still play Falcon 4.0 to this day because the campaign system is unmatched (and Falcon BMS provides a great face-lift)

 

 

2: Dynamic/unpredictable AI.

 

Put simply, AI whose routines do not appear "on-rails" and predictable, but are highly dynamic depending on situational factors... very difficult to implement well, but from a single-player or Co-Op standpoint having a relatively "realistic" AI is pretty important.

 

3: Realistic flight dynamics.

 

Pretty self-explanatory. I would still play CFS3 to this day due to its dynamic campaign system... IF it had flight modeling that did not feel "on-rails"... bad flight physics/dynamics kills the experience.

 

 

4: The pretties.

 

Yeah, graphics are important... but also some of the easiest elements to incorporate into a sim... AI, gameplay, flight dynamics are key with graphical fidelity supporting the core game.

 

Those are my big must-haves... unfortunately they seem to be polar opposites, in priority, to the modern methodology.

[Livai]

On 3/19/2023 at 9:12 PM, Livai said:

1. Unparalleled visual fidelity
2. Fully instrumented world simulating everything
3. Recreating the world with real data using multiply layers of information
4. Fully simulated cockpits and systems
5. Competitive Ai adversaries

 

"One Picture is Worth a Thousand Words"

 

[BraveSirRobin]

54 minutes ago, Livai said:

 

"One Picture is Worth a Thousand Words"

 

You want crappy looking blurred aircraft?

[BlitzPig_Bill_Kelso]

Beauty is in the eye of the beholder they say!

[Cynic_Al]

On 1/9/2023 at 4:48 PM, Mysticpuma said:

My short list of core features. 

 

1) Code and Netcode that can handle large formations of aircraft. 

 

2) CPU multi-core support

 

3) PBR graphics (for future proofing) 

 

4) VR support

 

5) Clouds and environment like DCS or an enhanced version of Truesky

 

5) 4 engine bombers

 

6) A detailed Quick Mission Builder much like @SYN_Vander has created for BoX. 

 

7) Ai of a coding level like Buddeye created many years back. 

 

I have lots more but as a starting point, those are a few points. 

 

What would yours be? 

 

 

 

1) An accompanying forum in which only established facts can be discussed, and nobody is allowed to ask a hypothetical question.

[gimpy117]

More detailed engine simulation and an accurate damage model. 

[Livai]

19 hours ago, BraveSirRobin said:

You want crappy looking blurred aircraft?

 

No sense for motion.................

You never ever used graphic settings slider?

Turn on Anti-aliasing cause the same effect

 

Anyway nonsense comment -> because motion blur effects always can be disabled in every game.

[BraveSirRobin]

1 hour ago, Livai said:

 

No sense for motion.................

You never ever used graphic settings slider?

Turn on Anti-aliasing cause the same effect

 

Anyway nonsense comment -> because motion blur effects always can be disabled in every game.

Sorry, but a screenshot doesn’t say “a thousand words” about a game.  It actually says nothing at all.

[Livai]

41 minutes ago, BraveSirRobin said:

Sorry, but a screenshot doesn’t say “a thousand words” about a game.  It actually says nothing at all.

 

- Your post before backfired now you try it again - - about what screenshot you talk anyway - wrong Quote - I guess -

Edited June 19, 2023 by Livai

[RossMarBow]

On 6/19/2023 at 7:08 AM, Livai said:

 

"One Picture is Worth a Thousand Words"

 

looks ai generated 

On 6/2/2023 at 1:25 AM, IckyATLAS said:

I think we overlook this. You can have an average frame rate that is good but say if you mix frames with some at half the frame time and half at double the frame time then the average will stay the same but it will be very choppy if it jumps here and there.

To ensure same frame time or say constant frame time, you must once you have decided what is your target, say 60/sec then you must know what is the CPU-GPU capable to load or build in a frame time or around 1/60 sec (I make it simple here). So if you have less then you keep the frame in the cache until it is released. But if you need more time then you are in trouble. This means that you must calibrate the scene so that it will not take more than the allowed frame time.

No idea how you can do that besides testing you game scenarios and measure all frame times and get a red flag when one frame exceeds the time. 

You play the game and no red flag is fine. One red flag means you must reduce the CPU-GPU load. It is a combination so you have also to know and measure separately CPU and GPU load to know where you are limited. Is it graphic load or game processing load. 

This means investing a lot of time for these kind of code or scene optimizations. I suppose this will be very costly and this is why they do not do it in general. Maybe only proven blockbusters (I mean in their second or third iterations) can afford doing it with their big buck budgets but this is not always true. 

 

Lots of low budget games run well.
Basically the greater a x3d chip helps performance the worse the code is

Edited July 1, 2023 by RossMarBow

[IckyATLAS]

On 7/1/2023 at 4:39 PM, RossMarBow said:

Basically the greater a x3d chip helps performance the worse the code is

This is also very true. Give plenty of power and memory as well as bandwidth and nobody cares much and the code becomes bloated and eats up the gain.

 

I remember when I had 16k on a cpu board with a Zilog CPU. I was using assembly language and implemented a 3D graphics program. Primitive line drawing but still 3D with matrix multiplications for translations/rotations etc. I implemented a very fast line drawing algorithms like the Bresenham one playing with registers etc. I had to scratch my head to use the ram optimally. When came the dos with a maximum  memory adressability around 640k that is 40 times 16k. Everybody said that such an enormous memory limit would last decades. It lasted just a few years.

 

[No_85_Gramps]

4 hours ago, IckyATLAS said:

I was using assembly language and implemented a 3D graphics program.

 

Ah...assembly language, probably a long lost art, but I believe MASM still has a following.