You don't need to make the entity clipping full 3D in a game like GW2. The horizontal distance from the camera works just fine. And even a 3D distance calculation between two points is child's play on today's computers, not to mention most games use a cube instead of a sphere for culling distance.
And the fact that entities are visible "too late" (or not at all) has nothing to do with collision detections, you mix up two totally different problems. Collision detection can be done even if the two colliding entities aren't even displayed. Do you really think they make collision detections client side?
Whether computing a distance between two points is child's play depends on how many times you have to do it and how fast. A trillion times per second certainly is not.
For network code, yes, you're probably just looking at the distance between two points. But for some things, you're looking at the distance between two irregularly shaped objects, and having to ask how far one of them could move in a particular direction before they collide--and then which direction they should subsequently go if they do collide. That's computationally intractible if you have to redo it a huge number of times every single frame. That's why physics computations are completely independent of graphics computations and use much simpler shapes.
Most likely, all collision detection computations are done first client-side, and then some (but not all!) of them are verified server-side. Where to place the camera is done client-side only, as even if you verify it server-side and find that the client was wrong, it's too late to do anything about it as the client has long since moved on to the next frame.
You're still missing the point.
Collision detection has NOTHING to do with distance culling, and the problem in GW2 is distance culling. The problem is that some entities appear too late in high density situations, which is distance culling.
You are mixing up several things which are very different here.
Originally posted by Quizzical
To completely fix the culling bugs in Guild Wars 2 would likely cost ArenaNet millions of dollars.
I'm sure you had a look into their source code before posting such a bold expert opinion...
Let's be serious, will you?
Where to place the camera in a given frame is partially a matter of collision detection. If you place the camera in the "wrong" place, then you end up culling some things that you shouldn't. Guild Wars 2 has a rather bad case of this. There are several different culling bugs, and this is the cause of one of them--but not all of them. It is likely the one that would be the hardest to fix, however.
-----
That I said it would "likely" cost millions rather than "definitely" was a concession that I'm guessing. But fixing the near plane clipping problems would mean having to completely redo collision detection, and that means having to redo a considerable chunk of every single model in the game that has any physics attached. Whether or not that would cost millions, it certainly wouldn't be cheap.
Now, you can get partial fixes for much cheaper than that, and ArenaNet likely will. It's probable that ameliorating the network-based culling issues would be a lot cheaper than that.
I had a blast reading this post. It reminded me of my 3D computing courses and exams at Computer Science
And that's exactly how it sounds - like someone who has read about 3D computing theory but has little practical knowledge.
Hooray for ad hominem attacks?
Have you ever implemented culling in a 3D game engine--and not just licensed some off-the-shelf approach, but implemented your own from scratch? Considered how collision detection for a camera has to be different from collision detection for players running into a wall? Computed the maximum possible near plane clipping distance that was safe to avoid improper culling as a function of how your camera placement works, the size and shape of the game window, and the user-adjustable field of view constant? Debated the relative merits of a signed area versus normal vector approaches to back face culling? (If you've got geometry shaders, you can use either one, and don't have to rely on the fixed-function signed area approach.) Carefully considered when it was safe to cull entire patches in tessellation control shaders (signed area never works here, and sometimes normal vectors don't, either)? Tried manually setting the depth in fragment shaders rather than using the fixed-function depth buffer? Figured out exactly why 3D graphics needs to use homogeneous coordinates in RP^3 rather than ordinary rectangular, cylindrical, or spherical coordinates in R^3?
And as a result of all of that, seen a whole slew of client-side culling bugs that you've created and had to fix?
Because I have. And as a result, I recognize some of the culling bugs in Guild Wars 2 as things that I've done myself.
Wait can you clear up things and first start by explaining what the problem is?
Whats a culling bug?
Loosely, "culling" means that a game engine considers whether to draw something and decides not to. A culling "bug" means that it decides not to draw something in a situation where it should have drawn it.
Wait can you clear up things and first start by explaining what the problem is?
Whats a culling bug?
Here are some screenshots of the near clipping plane culling bugs, which are the only culling bugs that are easy to screenshot.
As you can see from radar, I'm staring right at a wall. You can actually see part of the wall in the lower left part of the screen. But you can also see right through most of the wall, to other walls behind it, and even mobs and the sky far behind it.
On radar, it's not immediately obvious that I'm again staring at a wall. But you can see part of the wall on the right side of the screen, as well as a bunch of stuff far behind the wall.
Again, on radar, it's pretty obvious that I'm staring right at a mountain. And that I can see right through part of it.
On radar, you can again see that I'm staring right at a mountain. You can only see the wall I'm up against in the lower right portion of the screen. You can see some internal portions of the mountain that have no good reason to be there because it should never be possible to see them. And you can also see the sky through a crack between those.
What you can't see on that last screenshot is that there is also a distant tree behind the crack that you can see the sky through. Rotating the camera a bit makes the tree disappear and reappear. That's a different culling bug that you a single screenshot won't show, as the tree is either there or not--and if it's not there, then it's not obvious that it should be there.
oh dam!i know what he talk about!its a setting issue,very simple to fix.it is related to the camera and disatance setting this happen only if you are too close to an out of bound object .say a wall or other object that are stationary ,i move away or turn and it fix the issue .but you are right this is one setting i wish they would tweak!lol,i didnt think it was what you wewre talking about !
Originally posted by drbaltazar oh dam!i know what he talk about!its a setting issue,very simple to fix.it is related to the camera and disatance setting this happen only if you are too close to an out of bound object .say a wall or other object that are stationary ,i move away or turn and it fix the issue .but you are right this is one setting i wish they would tweak!lol,i didnt think it was what you wewre talking about !
In the above screenshots, I was right up against a wall with the camera at max zoom distance. The near clipping plane only affects things that are close to the camera.
But there are other culling bugs in Guild Wars 2 that show up for objects that aren't necessarily close to you. I don't have screenshots of them because a screenshot doesn't really show the bug very well. I suppose that I could make multiple screenshots of a screen that are identical except with a slightly rotated camera, and suddenly some things vanished. But I haven't.
Network-related culling bugs are much harder to reproduce, and screenshots don't really prove that anything is wrong. A player who vanishes could have warped away. Or just walked away with a long time gap between the screenshots.
Isn't it obvious that culling is not an easy fix? They certainly would have addressed it in a patch by now if it was seeing as it is such a high profile issue. ( Actually they have done something it just doesn't, by users perception, seem to have changed much) . They have communicated with the players in a number of posts as to what the problem is and how they are trying to stage the fixes. Search the GW2 forums "culling" and you will find several ANET posts on the subject. Do you really think trying to client side interface debuging the code is going to give you the best idea of what is going on? Thats a pretty stone knife and bear skin approach. Not saying you cant make guesses just thats what it is guessing. (If I know the Linux OS inside and out it doesn't mean I can debug Windows issues from the user interface. ) ANET has claimed they have a plan. If you feel it should have been fixed by now (or never should have existed in the first place) then yes it is worse then you think. As to the half-baked partial fix stuff well time will tell. For right now, you could be right but your guess is as good as mine. (I don't have one)
Isn't it obvious that culling is not an easy fix? They certainly would have addressed it in a patch by now if it was seeing as it is such a high profile issue. ( Actually they have done something it just doesn't, by users perception, seem to have changed much) . They have communicated with the players in a number of posts as to what the problem is and how they are trying to stage the fixes. Search the GW2 forums "culling" and you will find several ANET posts on the subject. Do you really think trying to client side interface debuging the code is going to give you the best idea of what is going on? Thats a pretty stone knife and bear skin approach. Not saying you cant make guesses just thats what it is guessing. (If I know the Linux OS inside and out it doesn't mean I can debug Windows issues from the user interface. ) ANET has claimed they have a plan. If you feel it should have been fixed by now (or never should have existed in the first place) then yes it is worse then you think. As to the half-baked partial fix stuff well time will tell. For right now, you could be right but your guess is as good as mine. (I don't have one)
I'm guessing that you didn't read my original post and won't read this reply. Sure, that culling isn't yet fixed could be taken as evidence in itself that it's not easy to fix.
But my post has a lot more details than that. What's probably not obvious to most players is that there are at least three fundamentally different classes of culling bugs in the game that have nothing to do with each other. And that at least one of them is caused by the way ArenaNet structured their model data being all wrong, on the basis that it will necessarily create rafts of culling bugs.
No attacks, I just don't trust your ability to have any remotely valid opinion about what the ANet developers did with their engine, and saying it is not forbidden here from what I know. There's no insult, no attacks... just a serious and legitimate doubt about your real technical knowledge.
Have you ever implemented culling in a 3D game engine--and not just licensed some off-the-shelf approach, but implemented your own from scratch? Considered how collision detection for a camera has to be different from collision detection for players running into a wall? Computed the maximum possible near plane clipping distance that was safe to avoid improper culling as a function of how your camera placement works, the size and shape of the game window, and the user-adjustable field of view constant? Debated the relative merits of a signed area versus normal vector approaches to back face culling? (If you've got geometry shaders, you can use either one, and don't have to rely on the fixed-function signed area approach.) Carefully considered when it was safe to cull entire patches in tessellation control shaders (signed area never works here, and sometimes normal vectors don't, either)? Tried manually setting the depth in fragment shaders rather than using the fixed-function depth buffer? Figured out exactly why 3D graphics needs to use homogeneous coordinates in RP^3 rather than ordinary rectangular, cylindrical, or spherical coordinates in R^3?
And as a result of all of that, seen a whole slew of client-side culling bugs that you've created and had to fix?
Because I have. And as a result, I recognize some of the culling bugs in Guild Wars 2 as things that I've done myself.
I've been working for computer generated real time special effects for video and movie industries for over 15 years (actually started doing 3D on a Commodore Amiga if that tells you something), even though lately I'm working in other domains. I'm not saying you're wrong on some aspects, but you're mixing up apples and oranges. You've learned a lot of technical terms you mix up into a soup which may impress the random forum reader. The distance culling problem in GW2 has nothing to do with collision detection, or even with visibility culling (frustrum, contribution, occlusion, etc...). Additional proof, the screenshots you've posted have nothing to do with the WvsW "invisible entities" problem GW2 has, it's something totally different.
And if the screenshots in your profile are any indication of what you ever did in 3D graphics, well... no big comment, I will just say the guys you critic the work of are a team of veteran professionals with several game engines in their resume.
Originally posted by jpnz
The irony when you post that the camera doesn't need collision detection.
The camera colliding with the landscape and static items so that it doesn't end inside of the floor or a wall has NOTHING to do with distance culling. Nothing.
Try reading the original post. There are at least three entirely separate classes of culling bugs in the game with totally different causes. Completely fixing one of them would probably leave the others unaffected entirely. You seem to want to focus on just one particular culling bug, and then argue that any discussion of other culling bugs in the game is automatically wrong because it doesn't apply to the particular culling bug that you have in mind.
Isn't it obvious that culling is not an easy fix? They certainly would have addressed it in a patch by now if it was seeing as it is such a high profile issue. ( Actually they have done something it just doesn't, by users perception, seem to have changed much) . They have communicated with the players in a number of posts as to what the problem is and how they are trying to stage the fixes. Search the GW2 forums "culling" and you will find several ANET posts on the subject. Do you really think trying to client side interface debuging the code is going to give you the best idea of what is going on? Thats a pretty stone knife and bear skin approach. Not saying you cant make guesses just thats what it is guessing. (If I know the Linux OS inside and out it doesn't mean I can debug Windows issues from the user interface. ) ANET has claimed they have a plan. If you feel it should have been fixed by now (or never should have existed in the first place) then yes it is worse then you think. As to the half-baked partial fix stuff well time will tell. For right now, you could be right but your guess is as good as mine. (I don't have one)
I'm guessing that you didn't read my original post and won't read this reply. Sure, that culling isn't yet fixed could be taken as evidence in itself that it's not easy to fix.
But my post has a lot more details than that. What's probably not obvious to most players is that there are at least three fundamentally different classes of culling bugs in the game that have nothing to do with each other. And that at least one of them is caused by the way ArenaNet structured their model data being all wrong, on the basis that it will necessarily create rafts of culling bugs.
I assume that alpha + beta should have revealed the issues prior to release and another 6 months post-launch makes a significant amount of time without it being solved. This is why I personally don't think it will get solved anytime soon. It doesn't seem like a priority to them.
Obviously we don't know how much time or money it will take them to fix it, but there is a limitation to every customer's patience when it comes to gamebreaking issues.
Enter a whole new realm of challenge and adventure.
On a technical focused thread like this, always look for who attacks the poster rather than the tech information / facts. If you know your stuff you don't have to attack just post the knowledge and see the other side squirm and try to attack you.
I love that the pro-gw2 is classifying culling as a 'minor bug'. I guess Anet is lying then.
Gdemami - Informing people about your thoughts and impressions is not a review, it's a blog.
Originally posted by itgrowls Agreed and so is the Loot nerfs/permaDR bug. They've had 4-5 months now to fix those and they still haven't done it yet. Been watching. There's a new forum post about the loot problem and the culling problem daily (and I do include the posts about how thieves need nerfing because they have omgwtfbbqstealth GASP in that count) people keep thinking its the class it's not. The only thing I keep seeing is them saying either the problems don't exist (as they do with the loot, thousands of people's experiences are wrong despite video and screenshot evidence to the contrary you see) or that they are endlessly looking into the problem (like they've done with culling.) Oh well onward to a game title that has better management methinks. So much for GW2.
Bye then? /shrug
You go on back to STO now, ya hear?
By the way, how long did it take Cryptic to fix the Klingon cloaking bug with text popups?
Originally posted by jpnz On a technical focused thread like this, always look for who attacks the poster rather than the tech information / facts. If you know your stuff you don't have to attack just post the knowledge and see the other side squirm and try to attack you.
I love that the pro-gw2 is classifying culling as a 'minor bug'. I guess Anet is lying then.
Arenanet didn't classify it as a bug but as a problem.
Arenanet also said the culling is server-side, not client side.
Arenanet also said that the number of characters the server report to the client is limited.
That is where the OP started from.
There is also the delay related to loading assets that mean the enemies will take time to be drawn and might be on top of you durinf the interval between the server reporting the enemy and your client render it.
All this is said in the OP and what is said by Anet.
Then the OP go on talking about issues that exisit but that don't exaclty relate to the "invisible armies".
In sPvP there also used to be problems with enemies popping after you were dead and that doesn't happen now.
So it is a question of believing Anet or not. I see no advantage of giving specifics that are lies - they could just be vague and they said they were fixing it.
Currently playing: GW2 Going cardboard starter kit: Ticket to ride, Pandemic, Carcassonne, Dominion, 7 Wonders
Originally posted by Gaia_HunterThen the OP go on talking about issues that exisit but that don't exaclty relate to the "invisible armies".
And this is what some here, including people 3D graphics development backgrounds, disagree with. Most of what the OP told holds some truth when taken separately. Trying to amalgamate it all together in some kind of doom theory about GW2's graphic and client/server engine becomes total nonsense.
Well since you seem to be the expert here on the culling issue then, can you please explain its exact cause and give us the steps to rectify the issue and an ETA on how long such a fix should take?
If you dont do stupid things while youre young, youll have nothing to smile about when youre old.
Sorry if this has been addressed, I didn't read the entire thread.
Umbra. GW2 uses Umbra 3 occlusion culling in GW2. Any discussion of the culling problems in GW2 that doesn't focus on Umbra 3 seems to be missing the biggest piece of the puzzle.
Umbra 3 handles visual and auditory culling for GW2 by modeling a virtual representation of the game world using propriatary algorithms and systems.
I don't know if the problems here are flaws in Umbra 3, or flaws in the way Arenanet has integrated Umbra 3 into GW2, or both, but examination of the Umbra 3 tech is crucial for understanding the culling issues in GW2.
I hope Quizzical will do a bit of digging on Umbra 3 and then do a new analysis of the possible causes/solutions for the culling issues in GW2.
During beta, you could turn off Umbra and use GW2s original culling solution. Performance hit, for me, was about 10% to FPS, but I ran with Umbra off during most of the beta events and with it off, I never had culling issues in WvW or elsewhere. Near the end of Beta, they removed the shortcut command line toggle and the game fully transitioned to the use of Umbra for culling. At the switch, there were massive problems with culling in the game, with entire landscapes vanishing under certain conditions. These issues seem to have been manually fixed, for the most part, in time for launch, but the remaining issues have been persistant or have even become worse since launch.
Even during the first month after launch, WvW culling seemed to be an infrequent issue, despite the fact that the queues were full almost 24/7. At around the same time culling started to become a bigger and bigger issue in WvW, sound culling and load time issues for certain locations, most notably Lion's Arch, also started to become constant.
The smoking gun has Umbra 3 branded on it's side, but I'd have to ask people with more knowledge of the technical details to tell me why exactly it's smoking so profusely!
About dynamic objects…can you talk a bit about how the culling of dynamic objects was integrated into your engine’s rendering pipeline? How much work and issues did this bring along?
Guild Wars 2 typically has thousands of dynamic objects active at any given time. These include characters, spell effects, particles, lights, etc… Each has variable lifetimes, complicating interaction with Umbra 3. To simplify this, we added a new management system to track object lifetime and bounding volume data. This system serves as the primary culling interface to the rest of the renderer.
At a high level, we first process static object results from a Tome query at the start of our render pass. This generates a small depth buffer used to cull dynamic objects. The object manager provides lists of dynamic objects to test against this buffer, along with any supporting data. Objects that pass the culling test are queued for render.
Our greatest challenge was developing this system within the framework of a large, mature code base relatively late in production. The results were worth it! The Guild Wars 2 client regularly processes tens of thousands of unique renderable objects while maintaining high frame rates on a wide range of hardware.
Originally posted by Gaia_HunterThen the OP go on talking about issues that exisit but that don't exaclty relate to the "invisible armies".
And this is what some here, including people 3D graphics development backgrounds, disagree with. Most of what the OP told holds some truth when taken separately. Trying to amalgamate it all together in some kind of doom theory about GW2's graphic and client/server engine becomes total nonsense.
Well since you seem to be the expert here on the culling issue then, can you please explain its exact cause and give us the steps to rectify the issue and an ETA on how long such a fix should take?
Without access to the code no one can.
But someone with access to the code, Arenanet already pointed the general cause and an estimate to have fixes in place.
Of course it is convenient to try to make Anet incompetent (not the OP btw) since it is cool it to bash it since it has no sub and all while ignoring the complexities inherent to having a physic engine like the one GW2 use, have such a huge amount of players per map in WvW and each character have a ton of customization (for example PS2 another MMO with a pretty sturdy physics engine has less charracter customization) opposed to the large majority of the MMORPGs that simply lack collision detection and projectile physics.
Currently playing: GW2 Going cardboard starter kit: Ticket to ride, Pandemic, Carcassonne, Dominion, 7 Wonders
Originally posted by Gaia_HunterThen the OP go on talking about issues that exisit but that don't exaclty relate to the "invisible armies".
And this is what some here, including people 3D graphics development backgrounds, disagree with. Most of what the OP told holds some truth when taken separately. Trying to amalgamate it all together in some kind of doom theory about GW2's graphic and client/server engine becomes total nonsense.
This discussion in this thread is a bit above my head, but I will say that over the years Quiz has proven himself to be a very knowledeable poster with no real hidden agenda for or against any particular game.
Just trying to live long enough to play a new, released MMORPG, playing New Worlds atm
Fools find no pleasure in understanding but delight in airing their own opinions. Pvbs 18:2, NIV
Don't just play games, inhabit virtual worlds™
"This is the most intelligent, well qualified and articulate response to a post I have ever seen on these forums. It's a shame most people here won't have the attention span to read past the second line." - Anon
Originally posted by Gaia_HunterThen the OP go on talking about issues that exisit but that don't exaclty relate to the "invisible armies".
And this is what some here, including people 3D graphics development backgrounds, disagree with. Most of what the OP told holds some truth when taken separately. Trying to amalgamate it all together in some kind of doom theory about GW2's graphic and client/server engine becomes total nonsense.
Well since you seem to be the expert here on the culling issue then, can you please explain its exact cause and give us the steps to rectify the issue and an ETA on how long such a fix should take?
Without access to the code no one can.
But someone with access to the code, Arenanet already pointed the general cause and an estimate to have fixes in place.
Of course it is convenient to try to make Anet incompetent (not the OP btw) since it is cool it to bash it since it has no sub and all while ignoring the complexities inherent to having a physic engine like the one GW2 use, have such a huge amount of players per map in WvW and each character have a ton of customization (for example PS2 another MMO with a pretty sturdy physics engine has less charracter customization) opposed to the large majority of the MMORPGs that simply lack collision detection and projectile physics.
Pretty much this ^
The main issue with Culling (and Anet has NOT been shy about explaining this), is how the game prioritizes it's data. There are definitely some optimization issues with the game engine (i.e. getting only 20 buddies to display at once, is not impressive), but without knowing exactly and all variables the game is checking for and processing, no one but Anet can know that much detail. And it would be unreasonable to ask them for such.
I have my own doubts about Anet's ability to fix the issue, but they've only ever promised they'd work on it. Which they have been. Furthermore, there is a lot more going on in this game than most. When you look at any other MMO, what goes on under the hood is very stripped down by comparison, in favor of being able to process more data. This is why the older games tended to be more 'massive'. They had less variables to compute at any given time, so they could process a lot more of them. There may come a time when Anet decides to reduce / remove some minor feature in favor of performance, but for now I don't think that'll happen.
Comments
Where to place the camera in a given frame is partially a matter of collision detection. If you place the camera in the "wrong" place, then you end up culling some things that you shouldn't. Guild Wars 2 has a rather bad case of this. There are several different culling bugs, and this is the cause of one of them--but not all of them. It is likely the one that would be the hardest to fix, however.
-----
That I said it would "likely" cost millions rather than "definitely" was a concession that I'm guessing. But fixing the near plane clipping problems would mean having to completely redo collision detection, and that means having to redo a considerable chunk of every single model in the game that has any physics attached. Whether or not that would cost millions, it certainly wouldn't be cheap.
Now, you can get partial fixes for much cheaper than that, and ArenaNet likely will. It's probable that ameliorating the network-based culling issues would be a lot cheaper than that.
Hooray for ad hominem attacks?
Have you ever implemented culling in a 3D game engine--and not just licensed some off-the-shelf approach, but implemented your own from scratch? Considered how collision detection for a camera has to be different from collision detection for players running into a wall? Computed the maximum possible near plane clipping distance that was safe to avoid improper culling as a function of how your camera placement works, the size and shape of the game window, and the user-adjustable field of view constant? Debated the relative merits of a signed area versus normal vector approaches to back face culling? (If you've got geometry shaders, you can use either one, and don't have to rely on the fixed-function signed area approach.) Carefully considered when it was safe to cull entire patches in tessellation control shaders (signed area never works here, and sometimes normal vectors don't, either)? Tried manually setting the depth in fragment shaders rather than using the fixed-function depth buffer? Figured out exactly why 3D graphics needs to use homogeneous coordinates in RP^3 rather than ordinary rectangular, cylindrical, or spherical coordinates in R^3?
And as a result of all of that, seen a whole slew of client-side culling bugs that you've created and had to fix?
Because I have. And as a result, I recognize some of the culling bugs in Guild Wars 2 as things that I've done myself.
C
N
Wait can you clear up things and first start by explaining what the problem is?
Whats a culling bug?
Philosophy of MMO Game Design
The irony when you post that the camera doesn't need collision detection.
Gdemami -
Informing people about your thoughts and impressions is not a review, it's a blog.
Loosely, "culling" means that a game engine considers whether to draw something and decides not to. A culling "bug" means that it decides not to draw something in a situation where it should have drawn it.
Here are some screenshots of the near clipping plane culling bugs, which are the only culling bugs that are easy to screenshot.
As you can see from radar, I'm staring right at a wall. You can actually see part of the wall in the lower left part of the screen. But you can also see right through most of the wall, to other walls behind it, and even mobs and the sky far behind it.
On radar, it's not immediately obvious that I'm again staring at a wall. But you can see part of the wall on the right side of the screen, as well as a bunch of stuff far behind the wall.
Again, on radar, it's pretty obvious that I'm staring right at a mountain. And that I can see right through part of it.
On radar, you can again see that I'm staring right at a mountain. You can only see the wall I'm up against in the lower right portion of the screen. You can see some internal portions of the mountain that have no good reason to be there because it should never be possible to see them. And you can also see the sky through a crack between those.
What you can't see on that last screenshot is that there is also a distant tree behind the crack that you can see the sky through. Rotating the camera a bit makes the tree disappear and reappear. That's a different culling bug that you a single screenshot won't show, as the tree is either there or not--and if it's not there, then it's not obvious that it should be there.
In the above screenshots, I was right up against a wall with the camera at max zoom distance. The near clipping plane only affects things that are close to the camera.
But there are other culling bugs in Guild Wars 2 that show up for objects that aren't necessarily close to you. I don't have screenshots of them because a screenshot doesn't really show the bug very well. I suppose that I could make multiple screenshots of a screen that are identical except with a slightly rotated camera, and suddenly some things vanished. But I haven't.
Network-related culling bugs are much harder to reproduce, and screenshots don't really prove that anything is wrong. A player who vanishes could have warped away. Or just walked away with a long time gap between the screenshots.
Isn't it obvious that culling is not an easy fix? They certainly would have addressed it in a patch by now if it was seeing as it is such a high profile issue. ( Actually they have done something it just doesn't, by users perception, seem to have changed much) . They have communicated with the players in a number of posts as to what the problem is and how they are trying to stage the fixes. Search the GW2 forums "culling" and you will find several ANET posts on the subject. Do you really think trying to client side interface debuging the code is going to give you the best idea of what is going on? Thats a pretty stone knife and bear skin approach. Not saying you cant make guesses just thats what it is guessing. (If I know the Linux OS inside and out it doesn't mean I can debug Windows issues from the user interface. ) ANET has claimed they have a plan. If you feel it should have been fixed by now (or never should have existed in the first place) then yes it is worse then you think. As to the half-baked partial fix stuff well time will tell. For right now, you could be right but your guess is as good as mine. (I don't have one)
I'm guessing that you didn't read my original post and won't read this reply. Sure, that culling isn't yet fixed could be taken as evidence in itself that it's not easy to fix.
But my post has a lot more details than that. What's probably not obvious to most players is that there are at least three fundamentally different classes of culling bugs in the game that have nothing to do with each other. And that at least one of them is caused by the way ArenaNet structured their model data being all wrong, on the basis that it will necessarily create rafts of culling bugs.
Try reading the original post. There are at least three entirely separate classes of culling bugs in the game with totally different causes. Completely fixing one of them would probably leave the others unaffected entirely. You seem to want to focus on just one particular culling bug, and then argue that any discussion of other culling bugs in the game is automatically wrong because it doesn't apply to the particular culling bug that you have in mind.
Guessing the key word of the post.
Obviously we don't know how much time or money it will take them to fix it, but there is a limitation to every customer's patience when it comes to gamebreaking issues.
Enter a whole new realm of challenge and adventure.
The only way it would cost millions is if they have to build an engine from scratch using hundreds of engineers.
Having a couple of guys working on it exclusively for a few months won't be millions.
Currently playing: GW2
Going cardboard starter kit: Ticket to ride, Pandemic, Carcassonne, Dominion, 7 Wonders
I love that the pro-gw2 is classifying culling as a 'minor bug'. I guess Anet is lying then.
Gdemami -
Informing people about your thoughts and impressions is not a review, it's a blog.
Bye then?
/shrug
You go on back to STO now, ya hear?
By the way, how long did it take Cryptic to fix the Klingon cloaking bug with text popups?
At least 10 months, right?
Arenanet didn't classify it as a bug but as a problem.
Arenanet also said the culling is server-side, not client side.
Arenanet also said that the number of characters the server report to the client is limited.
That is where the OP started from.
There is also the delay related to loading assets that mean the enemies will take time to be drawn and might be on top of you durinf the interval between the server reporting the enemy and your client render it.
All this is said in the OP and what is said by Anet.
Then the OP go on talking about issues that exisit but that don't exaclty relate to the "invisible armies".
In sPvP there also used to be problems with enemies popping after you were dead and that doesn't happen now.
So it is a question of believing Anet or not. I see no advantage of giving specifics that are lies - they could just be vague and they said they were fixing it.
Currently playing: GW2
Going cardboard starter kit: Ticket to ride, Pandemic, Carcassonne, Dominion, 7 Wonders
Well since you seem to be the expert here on the culling issue then, can you please explain its exact cause and give us the steps to rectify the issue and an ETA on how long such a fix should take?
If you dont do stupid things while youre young, youll have nothing to smile about when youre old.
Sorry if this has been addressed, I didn't read the entire thread.
Umbra. GW2 uses Umbra 3 occlusion culling in GW2. Any discussion of the culling problems in GW2 that doesn't focus on Umbra 3 seems to be missing the biggest piece of the puzzle.
Umbra 3 handles visual and auditory culling for GW2 by modeling a virtual representation of the game world using propriatary algorithms and systems.
I don't know if the problems here are flaws in Umbra 3, or flaws in the way Arenanet has integrated Umbra 3 into GW2, or both, but examination of the Umbra 3 tech is crucial for understanding the culling issues in GW2.
I hope Quizzical will do a bit of digging on Umbra 3 and then do a new analysis of the possible causes/solutions for the culling issues in GW2.
During beta, you could turn off Umbra and use GW2s original culling solution. Performance hit, for me, was about 10% to FPS, but I ran with Umbra off during most of the beta events and with it off, I never had culling issues in WvW or elsewhere. Near the end of Beta, they removed the shortcut command line toggle and the game fully transitioned to the use of Umbra for culling. At the switch, there were massive problems with culling in the game, with entire landscapes vanishing under certain conditions. These issues seem to have been manually fixed, for the most part, in time for launch, but the remaining issues have been persistant or have even become worse since launch.
Even during the first month after launch, WvW culling seemed to be an infrequent issue, despite the fact that the queues were full almost 24/7. At around the same time culling started to become a bigger and bigger issue in WvW, sound culling and load time issues for certain locations, most notably Lion's Arch, also started to become constant.
The smoking gun has Umbra 3 branded on it's side, but I'd have to ask people with more knowledge of the technical details to tell me why exactly it's smoking so profusely!
Want to know more about GW2 and why there is so much buzz? Start here: Guild Wars 2 Mass Info for the Uninitiated
This link may give a little insight on the integration of Umbra 3 into GW2.
http://www.umbrasoftware.com/clients/case-studies/guild-wars-2-case-study/
A possible clue here:
About dynamic objects…can you talk a bit about how the culling of dynamic objects was integrated into your engine’s rendering pipeline? How much work and issues did this bring along?
Guild Wars 2 typically has thousands of dynamic objects active at any given time. These include characters, spell effects, particles, lights, etc… Each has variable lifetimes, complicating interaction with Umbra 3. To simplify this, we added a new management system to track object lifetime and bounding volume data. This system serves as the primary culling interface to the rest of the renderer.
At a high level, we first process static object results from a Tome query at the start of our render pass. This generates a small depth buffer used to cull dynamic objects. The object manager provides lists of dynamic objects to test against this buffer, along with any supporting data. Objects that pass the culling test are queued for render.
Our greatest challenge was developing this system within the framework of a large, mature code base relatively late in production. The results were worth it! The Guild Wars 2 client regularly processes tens of thousands of unique renderable objects while maintaining high frame rates on a wide range of hardware.
Want to know more about GW2 and why there is so much buzz? Start here: Guild Wars 2 Mass Info for the Uninitiated
Without access to the code no one can.
But someone with access to the code, Arenanet already pointed the general cause and an estimate to have fixes in place.
Of course it is convenient to try to make Anet incompetent (not the OP btw) since it is cool it to bash it since it has no sub and all while ignoring the complexities inherent to having a physic engine like the one GW2 use, have such a huge amount of players per map in WvW and each character have a ton of customization (for example PS2 another MMO with a pretty sturdy physics engine has less charracter customization) opposed to the large majority of the MMORPGs that simply lack collision detection and projectile physics.
Currently playing: GW2
Going cardboard starter kit: Ticket to ride, Pandemic, Carcassonne, Dominion, 7 Wonders
This discussion in this thread is a bit above my head, but I will say that over the years Quiz has proven himself to be a very knowledeable poster with no real hidden agenda for or against any particular game.
Now you on the other hand, not so much.
"True friends stab you in the front." | Oscar Wilde
"I need to finish" - Christian Wolff: The Accountant
Just trying to live long enough to play a new, released MMORPG, playing New Worlds atm
Fools find no pleasure in understanding but delight in airing their own opinions. Pvbs 18:2, NIV
Don't just play games, inhabit virtual worlds™
"This is the most intelligent, well qualified and articulate response to a post I have ever seen on these forums. It's a shame most people here won't have the attention span to read past the second line." - Anon
Pretty much this ^
The main issue with Culling (and Anet has NOT been shy about explaining this), is how the game prioritizes it's data. There are definitely some optimization issues with the game engine (i.e. getting only 20 buddies to display at once, is not impressive), but without knowing exactly and all variables the game is checking for and processing, no one but Anet can know that much detail. And it would be unreasonable to ask them for such.
I have my own doubts about Anet's ability to fix the issue, but they've only ever promised they'd work on it. Which they have been. Furthermore, there is a lot more going on in this game than most. When you look at any other MMO, what goes on under the hood is very stripped down by comparison, in favor of being able to process more data. This is why the older games tended to be more 'massive'. They had less variables to compute at any given time, so they could process a lot more of them. There may come a time when Anet decides to reduce / remove some minor feature in favor of performance, but for now I don't think that'll happen.