Bouben
-
Content Count
1203 -
Joined
-
Last visited
-
Medals
Everything posted by Bouben
-
Yeah, nice one. It is not that easy unfortunately :)
-
Yes, it is certainly more flexible.
-
Shouldn't the tail remain only part of its frequencies based on a distance? Shouldn't the relationship between frequencies of a tail change based on environment? Some of the frequencies should be absorbed even for tails depending where the "ear" is located in relation to the source sound (in front of or beyond the critical distance). But I am just speculating as sound design is only a free time hobby of mine which I don't know much about in the area of actual physics rules. There will be definitely a huge change with tails in the game as relationship between source sound and echo creates huge distance information for human brain. I wish you success as this is an immense task from my point of view. So many calculations or clever workarounds will be needed. Oh and thank you for the papers. I will read it. EDIT: If you implement the stuff in the papers, I will give you my deepest respect as this is currently beyond my brain capacity. So many factors... ---------- Post added at 18:38 ---------- Previous post was at 18:34 ---------- Sorry, but the current filter implementation is very simple. The more distant the more HF attenuated the sound becomes. That is a very basic and only partial implementation of how distant sounds behave. Filters are seamless but in real world there is an immense amount of filters and their relationships immensely complex. A filter is definitely better than no filter at all but that does not mean it is an implementation ready-to-go. Also, there needs to be some algorithm (or workaround) calculating how long individual frequency bands will be as this is also a non-uniform thing in real world. Also, additional samples are much more performance friendly than real-time filter calculations. Samplers have been always easier on HW than virtual synthesizers. Samplers need mainly RAM, but real-time audio processing needs CPU. But if BIS can manage to get believable effect from filters and other real-time calculations then it would be much better than samples of course. It would also be much more difficult to achieve as synthesis is a complex art.
-
Currently, it is in a state I was afraid it would be. The same samples just attenuated with a very simple LP filter applied to them. That is not how it works in real world and it is not how it sounds in real world. The filters cannot be uniform. There is a lot of non-linearities to distant sounds. Distant gun sounds have much more mid frequencies than it's currently in the game. There is also MUCH less bass frequencies in real world to distant gun sounds. There should also be an echo but I know the sound team said it will be implemented later. Also, echo cannot be uniform for all frequencies too as some of them will be absorbed much more quickly than others based on the environment. There should also be a very pronounced higer-mid frequency volume peak for distant gun shots. Listen to BF3 or Half Life 2 distant gun shots to hear how it is done properly, yet in aesthetically pleasing way in a game. Real world footage is a more preferable source material though. Now, I am not saying it is easy to do. I can imagine it must be super-difficult to program sounds in such way they behave like in real world. So every effort is valued but currently it sounds just bad and needs a lot of work. Will hear what implemented echo tails do to the sounds once they are in the game.
-
I really hope the distant shooting will be as close to real sounds as possible and not just samples being lower in volume with some basic filters on it. It is so iconic and immersive to every firefight.
-
Yes, it is a good change.
-
Weapon Resting & Deployment Feedback
Bouben replied to solzenicyn's topic in ARMA 3 - DEVELOPMENT BRANCH
I would also like an indicator. It could be a small dot or something similarly unobtrusive just to know I am ready in the best possible position (and you feel your resting in the real life so there has to be a feedback). -
Agreed. This is one of the main factors that make AI in Arma look so dumb.
-
Yeah, that is probably because the combat routines are separate and needs to be tweaked individually.
-
I see. I agree with what you want but it is a separate thing that would not solve what I mean.
-
That would not be sufficient unfortunately. You would have AI going prone in situations when it is better to run fast (bCombat AI mod has this problem). You would also have AI going prone when they just should lean back behind a corner.
-
And then AI dynamically popping up from behind obstacles...I keep dreaming :-)
-
[OPEN BETA] [SP] bCombat infantry AI Mod
Bouben replied to fabrizio_t's topic in ARMA 3 - ADDONS & MODS: COMPLETE
I am interested in this too. -
You have to think about AI. If this was a strictly multiplayer game project, things would be much easier.
-
I am very excited that AI is getting some love recently. Good effort, devs!
-
Good point.
-
I would guess it's preliminary yet.
-
DAC alters vanilla AI. If you want to post relevant feedback, you have to test things in vanilla. There is no way around.
-
[OPEN BETA] [SP] bCombat infantry AI Mod
Bouben replied to fabrizio_t's topic in ARMA 3 - ADDONS & MODS: COMPLETE
Excellent, Fab, thanks for the explanation! -
[OPEN BETA] [SP] bCombat infantry AI Mod
Bouben replied to fabrizio_t's topic in ARMA 3 - ADDONS & MODS: COMPLETE
So you think there is currently no way for AI to use correct stances according to their needs and LOS (i.e. not going prone when in need of shooting from a tower)? I am thinking of Men of War series where AI is obviously aware of their LOS and stances and can adapt accordingly (I am not talking about predefined "cover positions"). I haven't noticed Men of War being too resource heavy. But I am not well educated in this field. And how is bCombat benefiting from the "massive LOS real-time checking" you have mentioned? Could you please elaborate a bit what is going on? -
[OPEN BETA] [SP] bCombat infantry AI Mod
Bouben replied to fabrizio_t's topic in ARMA 3 - ADDONS & MODS: COMPLETE
fab, this is probably not strictly related to your mod but do you know if AI is vertically aware of obstacles in their LOS and if they are capable of adjusting their stances in order to gain LOS over the obstacles again? -
I would say that this is simply a badly designed mission. I agree with the rest that it would be better to have it as a behaviour module.
-
Hello, I have just recently created a ticket about AI turret gunners' unreal behaviour and precision. I consider this to be a major flaw of the game since the OFP. It has considerable impact on the gameplay. The link is in my signature: 0022531: AI vehicle crew gunners are too quick/too good at aiming
-
[OPEN BETA] [SP] bCombat infantry AI Mod
Bouben replied to fabrizio_t's topic in ARMA 3 - ADDONS & MODS: COMPLETE
OK, now my config isn't working. I can change whatever in the config.sqf but it does not change stuff in the game. I downloaded bCombat from Armaholic. EDIT: No matter what I change in the config.sqf it does not work accordingly. I am testing it on the latest DEV branch but I cannot imagine configs not being respected just because of running the DEV branch. I will try it on stable just to be sure, though. EDIT2: OK, the problem was caused by the -nofilepatching parameter. Sorry for that. -
That's unfortunate but thank you for the answer. Do you know if BIS vanilla explosions and dust effects affect AI LOS? Thank you!