|
| 1 person marked this as a favorite. |
I have no idea about the game you were referencing. But what I would like to see is a system where you can mix the ingredients in different ratios to find different abilities (one mix might have higher dmg, while another might be cheaper to enchant, while another might have higher durability, etc) and to have each step be a mini game to produce the item with a very minor buff depending on your skill in the minigame.
Goblin Squad Member
|
I feel that I should point out I am only referring to the crafting portion of the game, not the combat portion. I am in no way trying to advertise the game, just had no availabe examples otherwise.
I'm glad you were able to sift through that and see part of my point. If that comes off as snide it wasn't intended that way. It was a comment on my rambling, not your cognative abilities.
It's okay, relax, you're allowed to make references and share your opinions. Wait till someone actually accuses you of something before defending yourself. *pat pats*
|
Goblin Squad Member
|
| 1 person marked this as a favorite. |
I agree with you Sabatour that character customization is very important for many people (myself included). I have a different view on what should be early priorities though:
Rather than blowing too much of their time and resources (delaying start of EE and OE) to create a million different hairstyles I would suggest that they make sure they create a system that can handle plenty of customization, thus allowing them to expand character customization options as they go along. Players will never stop asking for customization options for their characters, there can never be "enough" for some people. I think it would be a good idea to keep adding stuff as they go but this works only if they have a system in place that allows a lot of customization. Rewriting code at a later stage when possibly the original coders have been replaced could turn out to be a huge hassle. Thus, a flexible system rather than a lot of options available from the start sounds like the proper way to prioritize to me
In TSW, when players asked for more customization options the devs were seemingly at a loss because the system in place did not allow for the kinds of customizations asked for. They tried to change parts of their systems to accomodate for this but it was a slow and painful (and, I bet, expensive) process. If they had spent some more effort on creating a flexible system from the start, I believe they would have benefited greatly.
I am neither a programmer, game developer nor 3D-artist so its possible my reasoning on this matter is flawed but it makes intuitive sense to me.
(EDIT) I haven't read the entire thread since its an old one, sorry if these things have already been discussed in-depth before.
|
Goblin Squad Member
|
Slider scales coupled with generic variable types (like hair color 6H, hair style 02) takes all the information for your unique character and transforms it all into digits making it very fast. That is what is being sent back and forth between server and client. The translation on the client of those digits into individual characters and world objects is a function of the user's client running on their unique computer system.
Every character body will have to be rendered anyway, whether using strings of numbers that are reasonably unique combinations or identically numbered clones.
Whether to use a fully featured character creation system or a simplistic cartoony character cloning system is a question of project management.
If project management has to choose between dedicating limited resources to building the world or to building and using a character creation system, which should the developer prefer?
My hope is that the Unity engine has enough pre-built skins, motion modeling, and wireframe coordinate tools to make that never be an either/or decision.
Neverwinter has a good character generation system but the camera position used in-game is always a set distance from the character, so shortcuts could be taken on which of the character's unique numbers (character appearance values) will display. This 'shortcut' led to so many player characters looking like clones. Just as objects disappear when far away, so do unique features disappear from a certain distance, leaving all looking very much alike. This is exacerbated by limitations in available clothing/gear.
So the depth available in the Neverwinter character creation system is irrelevant except only in that the players cannot complain that the character creation system does provide enough options for a custom appearance. It provides plenty of options, but those options are not part of what the game actually renders in the game world
|
Goblin Squad Member
|
| 1 person marked this as a favorite. |
The character has to be given priority at EA because the playing experience starts with the character. It's the first system in the game. It's the one players will have to live with for their entire PFO career. I'd say it's the one and only thing that has to be robust and complete when the curtain goes up.
Heck, forget the game, give me a great character creator and I'm entertained for two weeks.
|
Goblin Squad Member
|
Neverwinter has a good character generation system but the camera position used in-game is always a set distance from the character, so shortcuts could be taken on which of the character's unique numbers (character appearance values) will display. This 'shortcut' led to so many player characters looking like clones. Just as objects disappear when far away, so do unique features disappear from a certain distance, leaving all looking very much alike. This is exacerbated by limitations in available clothing/gear.So the depth available in the Neverwinter character creation system is irrelevant except only in that the players cannot complain that the character creation system does provide enough options for a custom appearance. It provides plenty of options, but those options are not part of what the game actually renders in the game world
...yuck. That's just turrible.
The character has to be given priority at EA because the playing experience starts with the character. It's the first system in the game. It's the one players will have to live with for their entire PFO career. I'd say it's the one and only thing that has to be robust and complete when the curtain goes up.
I wholeheartedly agree with this sentiment. I don't want to be hamstrung with an ugly generic character for my entire PFO "career" because I was part of EE and less options were available to me at that time.
|
Goblin Squad Member
|
| 1 person marked this as a favorite. |
My very sincere hope is that we'll have access to a Character Creation tool well before Early Enrollment starts, so that we can spend a lot of time customizing our characters.
Vanguard did this and I thought it was great, until they repeatedly deleted all my templates due to bugs.
|
Goblin Squad Member
|
If project management has to choose between dedicating limited resources to building the world or to building and using a character creation system, which should the developer prefer?My hope is that the Unity engine has enough pre-built skins, motion modeling, and wireframe coordinate tools to make that never be an either/or decision.
Neverwinter has a good character generation system but the camera position used in-game is always a set distance from the character, so shortcuts could be taken on which of the character's unique numbers (character appearance values) will display. This 'shortcut' led to so many player characters looking like clones. Just as objects disappear when far away, so do unique features disappear from a certain distance, leaving all looking very much alike. This is exacerbated by limitations in available clothing/gear.
Good points about the Neverwinter system. NWO has a good enough creator, but the overall art style isn't that good. It's like a poor man's Dragon Age. That said, NWO is F2P. PFO is not, so it has to look good. GW2 set a pretty high bar for that. A pay to play MMO that launches in 2015 is going to have to look pretty good.
PFO EA will open with only 3, maybe 4 races. The sucker punch they haven't told us yet is that only the basic 4 "class paths" will be available on day 1 (fighter, cleric, rogue, wizard). We can handle all of that, but a limited/bad artwork creator? I will consider my pledge money wasted if we get that.
|
|
@Being
The strings of numbers are different sizes depending on how unique they need to be, that is why clones are so often used because a clone with a few options can be transmitted in a few digits, but a unique character with a thousand different options many of which are not simply yes-no, would require tens of thousands of digits. How much is clone vs unique and the methodology used makes a HUGE difference on bandwidth and for something the is purely cosmetic. Other stuff is at least as important and can't be trimmed down the way cosmetics can.
That is not to say I am fine with clones, just that I understand the limitations and won't hold grudges for avoiding the creation of a game that is gorgeous but has a frame rate of 1 per second because they used all their bandwidth on prettiness. The game has to play well, and frankly, for me at least, the gameplay takes priority. Yes I still want all those awesome character options and customization just not at the expense of the gameplay (but I have found that I am unusual in that regard. Many of my friends won't even TRY a game if the graphics are old. I wonder what they would think of muds)
|
Goblin Squad Member
|
@Being
The strings of numbers are different sizes depending on how unique they need to be, that is why clones are so often used because a clone with a few options can be transmitted in a few digits, but a unique character with a thousand different options many of which are not simply yes-no, would require tens of thousands of digits...
'Tens of thousands' of digits is, I believe, an extreme exaggeration, especially in a world where the byte allows for 64 bits. Where your bottleneck will be is in the local client translating those bytes into images, and not in the network through-put. That is why the shortcuts used in NW for abbreviated rendering of the characters is a shortcut at all.
|
Goblin Squad Member
|
I don't want to be hamstrung with an ugly generic character for my entire PFO "career" because I was part of EE and less options were available to me at that time.
They could allow players to recustomize their toons once new options become available, same as for when they release new playable races.
|
|
'Tens of thousands' of digits is, I believe, an extreme exaggeration, especially in a world where the byte allows for 64 bits. Where your bottleneck will be is in the local client translating those bytes into images, and not in the network through-put. That is why the shortcuts used in NW for abbreviated rendering of the characters is a shortcut at all.@Being
The strings of numbers are different sizes depending on how unique they need to be, that is why clones are so often used because a clone with a few options can be transmitted in a few digits, but a unique character with a thousand different options many of which are not simply yes-no, would require tens of thousands of digits...
There are 8 bits in a byte. 64 bit processors refer to processing 64 bits at once not how many bits in a byte.
The one thousand options was the exaggeration and the bottleneck for bandwidth is the bottleneck because the server has a limited bandwidth to communicate to every player that is online at the time. That is the big slowdown, communication. Ryan (I think it may have been one of the other developers, this was a while ago) even said so during one of the debates about arrows magically flyinf through trees to hit (said that there wasn't enough bandwidth to comunicate that well for collision detection)The translation between the incoming info and the drawing of graphics onscreen has far more to with the power of the clients computer and that is why games have requirements.|
Goblin Squad Member
|
There are 8 bits in a byte. 64 bit processors refer to processing 64 bits at once not how many bits in a byte.
Thank you for the correction, but 64 bit data paths and data storage blocks are not only in the processor. Incidentally while a byte is most commonly 8 bits, less commonly a byte can be anywhere between seven and twelve. Each transmitted byte can identify eight blocks of data, each of which may contain 64 bits. Each bit in your transmitted byte, once the packet is received, carries more information than you are allowing for. A 64 bit register can store 2^64 or more than 18 quintillion different values. That means it can take few bytes transmitted from server to client to mean quite a bit (pun intended)to the client.
If everything were in the data stream we wouldn't have to download megabytes of data with every patch. What is transmitted from server to client is a very small part of the vast amount of information processed on the client.
The one thousand optionsfwiw I balked at your 'ten thousand' not the one thousand.
was the exaggeration and the bottleneck for bandwidth is the bottleneck because the server has a limited bandwidth to communicate to every player that is online at the time.That is certainly not the only bottleneck, especially if you are downloading a movie, running an active instance of Skype, TeamSpeak, FRAPs, and all while streaming Beethoven's Seventh. Players doing these and similar then complaining about their lag is laughable and also tragic.
That is the big slowdown, communication. Ryan (I think it may have been one of the other developers, this was a while ago) even said so during one of the debates about arrows magically flyinf through trees to hit (said that there wasn't enough bandwidth to comunicate that well for collision detection)The translation between the incoming info and the drawing of graphics onscreen has far more to with the power of the clients computer and that is why games have requirements.
|
|
You are only thinking of the player but the server side is just as important.
Consider for a moment, ten slots for options, each slot has 100 options, each slot has primary and secondary full color options.
That is very light for character customization, but lets look at numbers, each color takes three bytes (8 bit), each slot takes a byte, so 30 bytes times 8 bits each times how many people are on each persons screen.
That is just reference info not the options themselves.
Now for a more modern game, you need each of hundred sliders just for the body plus twenty or thirty slots for options plus primary secondary and sometimes tertiary colors for each, plus a symbol made of one to three slots.
Each slots meaning, hair options is one slot, shirt options is one slot, etc.
Goblinworks Executive Founder
|
It doesn't matter how you arrange it; if there are 2^n possible character models, you need to communicate n bits once per character to the client as network overhead. That's trivial in terms of network overhead.
The rendering is more complicated, but it isn't significantly harder to render 25 different models doing different things than to render 25 of the same model each doing a different thing.
|
Goblin Squad Member
|
You are only thinking of the player but the server side is just as important.
Consider for a moment, ten slots for options, each slot has 100 options, each slot has primary and secondary full color options
What I think you aren't factoring is that what is actually being transmitted can be a character appearance key which is translated by the client into each unique character image.
If you played the Mass Effect series you may remember when you imported a character model from an earlier game (importing a mass effect 2 character into mass effect 3 as an example) the character had a string of hexadecimal digits that represented the various possible unique appearance details. If that string of hexadecimal values is all that is transmitted from the server to the client to render a player character for your client it represents very little bandwidth. The whole image isn't transmitted, the appearance key is transmitted and your client builds the image from that key.
|
|
That key is exactly what I am talking about. But the key you see in mass effect is only the face, now add the key for each of the fifteen item slots, the three colors for each of them, body types, the players unique symbol, name, current action/pose, current location, current facing, each currently visible item and their colors styles and designs, and any visible magic effects, and send that data for each character on screen to each of the thousands of players connected to the server at any one time.
Oh and not only that but as the number of options for each slot increase so does the number of bits. Even hexadecimal requires two digits (8 bits) for each red, green, and blue for each color for each item. That's 135 bytes for the color alone not including colors for held/sheathed items or magic effects.
Lets say an average of ten people onscreen for a thousand players, 135*10*1000=1,350,000 that is over a million bytes, or over a gb, just for the colors.
|
Goblinworks Executive Founder
|
64-bit color is pretty high up in the realm of color possibilities. It won't be transmitted as hex-equivalents-in-ASCII, it will be as binary data. A basic RGB values in the form of #CC0044 takes 24 bits to communicate (Likewise for hue, saturation, and brightness); selecting from a palate of 4096 colors would take only 12 bits to communicate.
The network traffic required is trivial next to the hardware requirements to be able to interpret and render that information.
The server side issue is with the number of interactions
if N is the number of bytes per character
and Y is the number of characters then the formula for bytes the server has to transmit
is (N*(Y-1)) * Y)
so lets plug some numbers in assuming N is a small number say 100 bytes
Y ... total bytes
1 ... 0
2 ... 200
3 ... 600
4 ... 1200
5 ... 2000
6 ... 3000
7 ... 4200
8 ... 5600
9 ... 7200
10 ... 9000
....
20 ... 38000
30 ... 87000
40 ... 156000
It soon mounts up and this let us not forget this is not per second but several times per second
The server has to transmit to all characters from the server to the client. Given how often people move emote etc while in sight it is often easier and quicker for the server to transmit it all every time than try and work out what has changed on each character and transmit only that info
|
Goblin Squad Member
|
That key is exactly what I am talking about. But the key you see in mass effect is only the face, now add the key for each of the fifteen item slots, the three colors for each of them, body types, the players unique symbol, name, current action/pose, current location, current facing, each currently visible item and their colors styles and designs, and any visible magic effects, and send that data for each character on screen to each of the thousands of players connected to the server at any one time.
Oh and not only that but as the number of options for each slot increase so does the number of bits. Even hexadecimal requires two digits (8 bits) for each red, green, and blue for each color for each item. That's 135 bytes for the color alone not including colors for held/sheathed items or magic effects.
Lets say an average of ten people onscreen for a thousand players, 135*10*1000=1,350,000 that is over a million bytes, or over a gb, just for the colors.
Occlusion is one's friend...
|
Goblin Squad Member
|
| 2 people marked this as a favorite. |
People are going to be really, really disappointed when the character creator comes out. I suspect character model options are going to be extremely limited. Character customization requires a large investment into the art assets. Best case scenario that I see is that they set themselves up with a system with a large number of potential points of customization, though each point has only a few options to start. That way they can extend it with new options as the character art guys find time to produce new assets.
|
Goblin Squad Member
|
People are going to be really, really disappointed when the character creator comes out. I suspect character model options are going to be extremely limited. Character customization requires a large investment into the art assets. Best case scenario that I see is that they set themselves up with a system with a large number of potential points of customization, though each point has only a few options to start. That way they can extend it with new options as the character art guys find time to produce new assets.
That sounds sensible - minimum viable product and all that - and as long as we get free recustomizations when the new stuff is added I have no issues with it. I doubt I'll ever get to put Deianira's butterfly tattoo on her in-game character face, but I can live with that! :)
I'm still hoping for a character creator to play with prior to Early Enrollment. I have fond memories of playing around with City of Heroes' character creator for entire play sessions.
|
Goblin Squad Member
|
@Dario
Depending on the methods and systems GW takes, even limited character options could produce a good range of customisation.
If GW goes for the Swap the model or a Swap and minor Morph method it'll provide the least return on customisation options.
Where a near full morph based system provides the most customisation per morph.
Look in the thread above, you'll find some Character Creation systems;
The Dragon Age one is a hybrid of Swap and Morph, with Swapping of the base bodies (which is also the clothes), base heads and hair. The heads had morphing to help in making them look as you wanted.
The Phantasy Star Online 2 one is nearly all morph based, the only swappable parts are the base body (which includes the head) the hair, and clothes (all clothes have morphing). The customisation in it is very high even with a few morphs.
|
Goblin Squad Member
|
The server has to transmit to all characters from the server to the client. Given how often people move emote etc while in sight it is often easier and quicker for the server to transmit it all every time than try and work out what has changed on each character and transmit only that info
Once a character is in range its values must be transmitted. Subsequent transmissions need only respond to changes, such as position and emote. You will surely have maxchar rendered settings. By implication each character need thereafter only be numbered. Changes (location, emotes for example) in each numbered character would be tracked and transmitted but the whole character doesn't have to be resent.
If programmers were simple people you could be so simple talking about what they can do. Amazing complexity results from relationships between very simple elements.
|
Goblin Squad Member
|
It would be lovely to be able, for once, to have an avatar that resembles my character concept. If I want a hulking tub o'lard cleric, or a female fighter that gets mistaken for a guy, or a pretty-boy thief that resembles a teenage girl, I am not going to get one in a MMORPG (well, the last one possibly).
I'm not saying that it is better to play such characters, but I like to have avatars that stand out a little and show that my character hasn't stepped straight from the 1979 AD&D book of class stereotypes. My avatar and name are so important to me when playing a game that I am prepared to put up with a good deal of game-play limitations if I am able to play the avatar I want.
Conversely, an otherwise good game that has poor avatars (DDO, I'm looking at you), makes me lose any emotional investment in my character.
By the same token, I like freedom to dress my avatar in whichever styles I see on the streets. Seeing NPCs in better clothes than I have is immensely frustrating.