Power by Lenshuang~2+3D Game Vfx Base since 2003~2014
  • 火影忍者鸣人 疾风传  终级风暴2

    以超越动画作为目标的游戏视觉效果。
    去年,被看作达到动画般的画面品质的[火影忍者鸣人终级风暴],与其说是震惊了游戏界,不如说是震动了动画界,很快便以第[2]部的姿态登场了。开发组是如何刻画[疾风传]和成长了的鸣人他们的呢?
    作者 久代忠史

    毫不妥协的踏实努力所获得的结晶
    相对于标榜[超越动画]并将其实现的前作,本作中宣扬的主题是[游戏和动画的超融合]。担任本作项目领导的西川裕贵先生说,要让玩家能够体验到[火影忍者鸣人 疾风传]精彩的剧情。举个例子,在被要求输入按键的Quick Time Event[interactive,Action]中,分析可以作出[神作画]的动画巨匠们的工艺 ,最终完美得在实时CG上再现。

    还有,把动画的舞台做成玩家能够冒险的[Ultimate adventure Mode ]中,实际在制作[火影忍者鸣人 疾风传]的动画时,就从Studio Pierrot定制了背景。通过使用了动画般的美术背景,在游戏中构筑了被称为[Artistic Field]的高品质的动画背景效果。

    并且,在其他部分也和前作同样的进行了不知妥协的彻底制作。当中最重要的就是角色动画了,用了大量的手工才做成。这些即使花费大量的时间和工作量,也要做成符合艺术家意图的姿势和动作,正是这样才判断手工是最有效的。但是,一些被认为不必要的动作部分,使用了公司自己开发的动力学工具,设法减轻了工作。

    模型也是同样,关于基本的制作手法和前作基本没有变化,但在本作中,对于建模的轮廓或骨骼,Skin Weight,shader等,吸取前作的教训把全部的项目重新修改,加入了微调整。

    虽然本作是这样持续推进着开发,但实际上最大的挑战还是工作量。为了让人看到剧情,准备了比前作多几十倍的分镜场景,能够制作完那些很大的原因,就是在开发初期进行的预见性的优化。艺术总管的石桥洋平先生说,[这些在当制作到达后期量产化时就显出功效来了啊]。

    本作作为把动画[火影忍者鸣人]的风格毫无折扣的做成了游戏再现出来。如果玩一下本作,不管是粉丝还是不知道本作的人都会更喜欢[火影忍者鸣人]吧。

     

    前排左起
    项目主管,西川裕贵先生。交互动作负责人,竹下勋先生。
    背景负责人,四所铁矢先生。模型,友池润先生。

    后排左起
    艺术总管,石桥洋平先生。特效负责人,鹰野典久先生。
    动作负责人,石原学先生。(以上都是Cyber Connect 2 公司的)

    九尾化鸣人的表现手法
    本作中除了Cel shading以外还使用了其他的Shading。其中之一就是九尾化的鸣人(图A)使用的Fall of Shading。使用拥有Alpha的Fall of 用贴图(图B),平滑的控制了外围部分的颜色(图B),流淌在身上的花纹用贴图的UV动画(图D)表现,还有脸的以部分想做成黑色的区域,根据顶点颜色(图C)更容易实现把颜色做到Mask上。

    A Fallof Shading 

     

    适用在鸣人身上的Falloff Shading

    B  Falloff 用贴图

    C根据顶点颜色作出的有颜色的Mask

    脸上一部分想做成黑色的地方,根据顶点颜色更容易实现把颜色做到Mask上

    D 贴图的UV动画

    流淌在身上的花纹用贴面纹理的UV动画来表现

    NextStep的模型
    本作的角色们是吸取了前作开发时的教训做成的。
    做了很多的改善来支撑本作的品质。

    微调(fine tuning)过的角色模型
    前作中角色们是少年,在本作中要符合原作的进展成长为青年。因为那样也要重新制作模型,针对前作中进行了高中低3个等级的LOD,本作中变更成了高(图A)和中(图B)两个等级。负责模型的友池润先生说,这是因为在前作中即使替换低模,关系到GPU的负荷也没什么优点,正是基于这个事实,在本作中为了中模和高模而没有制作低模的那部分资源,把脸等经常活动的部分作为中心做出更平滑的模型,由此进行了品质的提高。顺便说多边形数,前作的高模16000,中模8000,本做中大约是19000和11000多边形。

    高模

     

    中模

    对应多种多样角色的动画

    本作中拥有异想天开般形态或技法的新角色大量登场。
    把那些角色做了怎样的工作才使他们活动起来呢?

    有效利用Asset的绑定
    关于绑定,基于前作进行了骨架适应化的工作。在前作的故事时间点,有些角色看起来没有成年人的外貌,因为出于灵活运用面部动画等Asset的观点,基本上那些都没有很大的变化。但是,角色身上增加了摇动物体或附属品,只有那些部分追加了自由骨骼,就可以对想活动的部分设定动作了。

    鸣人的绑定

     

    特殊角色的Rigging
    本做的开发因为使用了3ds Max,绑定使用的biped,所以复杂的构造或活动的角色调整中都可以灵活运用。还有,动画负责人石原学先生说,动画基本上是用手工做成的,重吾(图A)的自由骨骼的多关节手臂或另一角色的四足山椒鱼(图B)的攻击动画,都是把每一个关节单独手工进行活动才完成的。但是,对于毫无意义的动作,也使用了动力学工具,八尾(图C)下半身的章鱼部分,根部分用手工分成了3个自由骨架,从那上面用动力学工具自动的变动着。还有,蝎的线(图D)的情况下,用自由骨架作成,沿着Spline装配缩放动画。

    手工制作的动画

    衣服的活动
    本作中也和前作同样设定有可以控制衣服活动的骨架,把那些移动,旋转,扩大,就可以对应到各种各样情况的服装活动。还有,为了实现比起前作更贴近图像的衣服轮廓(silhouette),skin的weight要进行调整,就可以做出膨胀时越来越涨大那样印象深刻的服装活动。

     

    公司自己开发的物理模拟工具
    对于摇动的物体或碰撞的判定使用了公司自己开发的物理模拟工具。这些是同一公司从开发的PS2同系列游戏时就采用的动力学程序,经过了反复的升级。机能方面,在这个工具上因为可以生成碰撞用的球体,而使用其他工具做成碰撞用的object就没有必要了,在这个工具内就可以作完整的调整了。

     

    由公司自己的工具进行面部动画的管理
    面部动画在前作中虽然被做成了库来管理,但在本作中将那些作进一步完善,技术美术(Technical Artist)用3ds max的脚本工具做成了工具,用这个进行表情数据的管理。由于这样,登录好的表情不管在哪个工具里都可以使用了。还有,能够登录的表情数据,被分成了整个脸,嘴的范围,嘴以外三个种类。画像就是案例。

     

    剧本的交互和行动
    本作复制出的,是把[进一步进化的游戏和动画的超融合]进行具体化,就是这个交互和行动。请看这个惊人的研究。

    有”神作画”意识的动画
    担任交互和动作负责人的竹下勋先生说,本作中的目标是再现出被称作[神作画]的超绝动作的动画,所以也参考了[火影忍者鸣人疾风传]以外的所有动画。其中之一被称为[板野杂技],导弹等飞行物体毫无拘束的交错乱飞,图像中沙土的移动被注入了那种精髓。这个沙土是很长的带状的对象,要让那些沿着路径活动,因为只有那些还不够充足,所以要以帧(Frame)为单位中让路径点蜿蜒曲折,由此让人看到板野杂技型的活动。

    剧话风格的强调线
    本作中让人吃惊的是彻底的再现出漫画和动画的表现手法。图片是高潮时的鸣人的表情,加入了剧画风格很强的线条。这些是在交互行动的情况,利用了决定好的摄像机,在摄像机前大量的配置贴上了强调线纹理的板型多边形,由此实现了符合表演的摆动效果。

    为了平滑连接的动画的Motion blur
    “神作画”不只是关系到动作,印象上也要保持神韵。比如说,两个角色在高速的做动作时,只有2和3帧(Frame)里表现出了极端的残像,图片虽然再现了那些,但在一连串的动画之间,还是要硬放入如图AB那样不可能的残像,这就是要有意识的做出在高速中要让眼睛跟的上的平滑的动画。还有,残像是在板型多边形上贴上贴图来再现的。

    电影级别的动画背景
    本做中是在Studio Pierrot的协作下,搭载了拥有新要素的Ultimate Adventure Mode。那个美丽画面的制作过程是什么呢?

    背景的基本

    图片是鸣人VS卡卡西进行战斗的生存演习场的背景。多边形数是106630多边形,贴图的种类有,diffuse和lightmap两种,从128乘64像素到2048乘2048像素的贴图,两种合计使用了69张。还有,在实际主机上显示的情况时,在画面内显示的多边形数要调整到10万以下。

    实机画像

     

    顶点颜色

    本帖最后由 trace 于 2012-1-17 12:43 编辑

    电影级别的动画背景
    本做中是在Studio Pierrot的协作下,搭载了拥有新要素的Ultimate Adventure Mode。那个美丽画面的制作过程是什么呢?

    背景的基本

    图片是鸣人VS卡卡西进行战斗的生存演习场的背景。多边形数是106630多边形,贴图的种类有,diffuse和lightmap两种,从128乘64像素到2048乘2048像素的贴图,两种合计使用了69张。还有,在实际主机上显示的情况时,在画面内显示的多边形数要调整到10万以下。

    实机画像


    顶点颜色


    线框 

    由CG做成动画风格的背景
    一般的背景制作方法是在贴图中画入细节,但那样细节就过多了,图A就是那样因为信息被整理过,就不宜做作为动画的表现。所以在本作中使用了顶点颜色(图B),在diffuse的地方画入阴影和顶点颜色,并保存,做了使画面颜色尽量流动的工作。还有,使用了光lightmap(图C)进行阴影部分的表现。

    A 背景的完成画面

     

    B 顶点颜色

    C lightmap

    Artistic Field的制作流程
    本作不是只有战斗,也包含冒险的要素。那就是[Ultimate Adventure],这个的背景就被称为[Artistic Field],担任[火影忍者鸣人]系列的动画制作的Studio Pierrot做出的动画型的背景的美术世界是可以在当中冒险的(图F)。这个制作手法,首先要做出布局Layout 图A,再向Studio Pierrot订货,然后是交货(图B)。这个交货除了构成要素外,还要被分层(动画用语中称为book,图C),对于把这些立在原本的3D模型中,再进行Billboard贴图(图D)就完成了。而且,也可以对树木等物体设定骨骼(图E)使其可动,就可以把游戏做的更加深奥了。

    1 布局
    图A

     

    2 动画背景美术的做成
    图B

     

    3 Camera Map
    图D

     

    图E

     

    4 完成
    图F

     

    我爱罗VS迪达拉的空中战背景
    本作中最惊人的就是我爱罗和迪达拉的空中战了。高速移动的动画背景风格的建筑群一看好像是预渲染的一样,但无论如何这些都是用实时多边形表现的。负责背景的四所铁矢先生说,当初这个部分考虑是用map chip组合来作成,不过在游戏风格的视觉上会过于缺乏趣味,所以尝试挑战用实时的多边形来表现。但是,因为资源方面非常严格,所以以帧(Frame)为单位计算显示的多边形数量进行优化。还有,因为当初用实际尺寸不能完成绘制,所以全部缩小到1/10,精度提高了。

    完成模型

     

    Flat Shading 模型

     

    Wire Frame

     

    高品质的影子表现
    作为一个从前作而来的大变更点就是影子的表现。在本作中的生成方法也和前作同样采用了lightspace perspective shadowmap的方法,但生成的对象在前作中是影子专用的固体(solid)模型,相对于前作,本作是使用了中模。这个变更,是因为在本作中因为不得不做出庞大的角色队伍,目的就是消减制作固体模型耗费的工时,作为次要的效果,影子的精度提高了,就连到手指的间隙都能清晰的表现(图A)。还有,在水面上,影子因为受到了水面shader的法线贴图的摇动参数的影响,也实现了摇动的表现(图B)。

     

    向新的特效表现挑战
    在前作中也博得好评的忍术特效在本作中依然健在。
    这次让大家看看是如何实现漫画或动画的风味的。

    沙缚牢的侵蚀表现
    我爱罗的沙缚牢那样的慢慢消失被侵蚀的表现,使用参考Alpha,在一定值以下的部分不进行绘制的原创shader。具体的对于沙球,分配Fractal类型的alpha贴图,再进一步做渐变。还有,最初是设定可以全部表示,所以可以让alpha值进行变化来实现。

     

    忍术[沙缚牢]中捕捉到敌人的沙珠,和不规则图案的Alpha 贴图

    须佐能乎的表现
    鼬的最终奥义须佐能乎中也加入了原创的构思。半透明物体的表现,直到现在还是使用调整模型上被分配的shader透明度的方法,但是aplah交点处的外观要好看就很难。所以在本作中,为了做出原作或动画的图像,使用在VideoBuffer上半透明混合的手法。由于这个方法实现了非常美丽的半透明表现。

     

    式纸之舞的表现
    本作的忍术中存在着大量的物体爆发或活动。但是,按照看到那些的物体数量做出来,立刻就会把资源耗尽了。所以,担任特效的鹰野典久说,为了在保证效果的同时消减数据而绞尽了脑汁。图片是以小南的[式纸之舞]为例,把贴上纸片贴图的圆筒做成多层,再进一步设置各自的UV动画的时间(图A)和碎屑射出(图B),这样就正好让人看起来是大量的纸在飞舞C。

    A贴了纸的贴图的圆柱体

     

    B粒子的输出

     

    C 实机上的图像

  • 日本独有的游戏影像的进化体系。
    [火影忍者鸣人----究极风暴]

    本作是把超人气漫画按着原作的动画片系列进行的游戏化。
    把PS3作为平台的[超越电影的次时代动作游戏],在发售前集中对各种话题进行着讨论。
    这次来开发组打探的任务,是为了让大家了解如何实现动画和游戏的融合。

    为了让动画片和游戏没有界限
            本作是在06年初时开发的,当时Xbox360刚在日本发售,PS3的发售还要有些时候,这时不论是谁都在思考着今后的次时代游戏的开发战略。在这样的状况中,负责开发Cyber Connect 2公司的董事社长松山洋先生,决心把[伴随着次世代机的图形性能飞跃的提高,各种开发经费也高涨]这个问题,通过以全世界的市场作为目标进行开发而得到突破。

            为了实现这个坚固的理念,深思熟虑过我们自己对于世界的优势,导出的方针就是[把日本独有的动画和漫画的表现贯穿始终],那就成为了所谓的[把动画和游戏的界限消除]的具体目标。为了这个目标,首先编成了10人的研究开发组作为先头部队,这期间有7名设计师进入开发组内,从事制作被称为[超动画]的[贯穿始终]的视觉表现的预想影像,经过了半年时间终于在3dsMax上做了出来。

            为了达成目标如动画般的画面表现,本作的动画全部用手工完成,一方面是松山先生回想起那些庞大的工作量就说[总之是动作组(Motion Team)一直很努力才顺利前进],另一方面动作指导的石桥洋平先生也回答说[确实太辛苦了,但是有研究开发时适当策略可供参考,所以可以冷静的应对]。确实是可靠的言词。

            还有,本作是有动画版权的游戏,所以要由原版权公司检查。具体上,和动画设定的出入,对于游戏中存在的IF(假象剧情),原创忍术的[奥义]的表现进行检查。这些都给予了细心的注意,动画师不时地一边参照动画影像一边进行工作,还要根据每个负责人的文字描述做出概念,之后由导演进行核对,整个工程全部的阶段都要在内部核对,要做到尽可能不返工的程度,用这样精炼的制作来接受原版权公司的监督。职员们这样的努力,已经出色的成为本作的一个理念[兼顾画面的观赏性和作为游戏的游戏性这样的两个目标]。事实上在本作的预告片公开时,也能看到用户和游戏业界对于过高的游戏品质比起动画业界发出了更多的赞赏声。

            最后,通过采访最有印象的是,开发职员要面对把原作动画作为初始的动画本身,从言词的细微传来了敬意和热爱。从下页开始要介绍这种丝毫不妥协的热爱,平时挑战各种各样限制本作构思和技术平衡的真髓。

    前排从左开始,代表董事:松山洋先生;特别指导:下田星儿先生;项目指挥:西川裕贵先生。
    后排从左开始,四所铁矢先生;石桥洋平先生;梅田公一先生。
    (以上都是Cyber Connect 2公司的人)

    让人感觉不到CG噪音的简单角色表现
    动画作画的关键就是对信息量的控制。用少量的信息表现出强烈的个性,怎么才能做出活灵活现的角色呢?

     

    角色模型的LOD
            角色的建模使用3dsMax。图里是主人公漩涡鸣人的身体和头的建模(从左开始是高、中、低模)。多边形数量各自为高模16000(头部2400)、中模8000(头部1400)、低模4000(头部700),基本上由从镜头的距离算起:5米以内、5到7米、8米以上来区分,以此进行LOD处理。梅田公一先生表示,这是为了在减轻描画处理负荷的同时,表现出动画特有的[远处的东西被画出线条是经过整理的形态]。还有,有时有必要让人清楚的看到表演性质的脸,还会发生由于大量的角色登场引起的负荷上升那样局面的情况,要根据情况强制性的切换高模或是低模。还有,脸上使用到了很多的多边形,本作中使用法术时结印的手指也分配到了相当精细的多边形。

     

    重视画面效果的生动配色阴影
            公司内部以前使用过的乘法Cel Shader,绘制出皮肤的阴影部分的褪色效果是个难点,见上图A。在本作中,为了解决那个问题,模仿了Photoshop的soft light图层效果开发出了Cel Shader,见上图B。用这个方法虽然确实在皮肤的影子上能得到良好的结果,但同时衣服等东西原本的色彩很高的部分影子也变的明亮,结果作为影子就会出现破绽。于是开发出了新的彩度高的像素上使用乘算,彩度低的像素上用soft light的,把影子颜色在直线上混合描绘的shader,实现了取得平衡的阴影,见上图C。另外,项目主管的西川裕贵先生表示,因为这个shader保持相当的的彩度,根据动画上经常用的色调也可以适用在阴影的表现上,例如说把夜里的阴影做成青色时也很有用。

    整理出动画风格的阴影
            虽然图A把高模做出了Glow Shading,但是如果使用这个光源设定进行Cel Shading,光照会像图B那样忠实的表现出阴影来,导致表情变的非常僵硬。这样就达不到动画片的效果,所以需要像图C和D那样让模型膨胀再整理法线,作成调整过的模型,取得了法线信息后,把那些适用在原模型上,就可以再现出图E那样动画风格的柔软的表情。还有,在实际主机上的再现是开发了专用的插件对应完成的。这个Shading技术可以说是初期研究开发阶段很大的一个成果。

     

    必要的轮廓线(Outline)抽出法
            作为轮廓线的生成法,从PS2的时候就是被经常使用的方法,把模型按法线的方向推出很大的一圈,再把多边形的法线反向作为轮廓的绘制手法,图A。用这个方法,因为对于全部的多边形都适用一样的处理,头发等的轮廓也忠实的绘制出来,表现出期望中的动画风格的质感。还有一种生成法,一次性把模型渲染到2D buffer中进行post处理,是一种抽出模型的轮廓(silhouette)的绘制手法,图B。用这个手法,虽然使用silhouette生成了简单的的轮廓线(Outline),但不能在silhouette的内侧生成。所以本作的Cel shading上附加了轮廓线(Outline)选择的选项,由于设计师会对应状况选择绘制方法,实现了这两个手法的并用。

    动画中加入了其所特有的夸张
    动画的深奥妙趣,就是充分利用省略和强调张驰。
    那么如何通过CG来表现这些并灵活的运用呢?
    这里来披露一些吧。

    索具(Rigging)的准备配置(Setup)
            角色的索具装配使用了Biped。骨骼的构成,被称为全角色共通的基础的躯干部分有52个,各种调整用的自由骨骼是22个,面部是67个,服装和头发等摇动的物体用的延伸骨骼是30到80个,全部大概每个角色可以算出是170到220个。还有,做成的动画数,漩涡鸣人的情况,在自由战斗时是125个,极限任务模式里是90个。

    动画拖影(Anime Blur)的实现
            作为动画片固有的速度表现,速度过快的会有轮廓被拖长的情况。在本作中把三角锤的模型重叠 (图A),提前加入角色的末端部分(图B),把骨骼拉出来(图C),实现了这个效果(图D)。在骨架上加入了弹簧控制,根据动作的速度把拉出的长度自动的变化,虽然做出的动作也有些和表演意图不同,但为了取得平衡,全部的人都加入了。

     

    面部配置(Setup)的基础
            面部正如前述那样由67个骨骼控制(图A)。作为本作特有的设定,下唇的下骨骼占到脸的三分只一。这是为了可以表现出即使是动画独特的张嘴,下巴的位置也不会改变,或是表现出图B那样,脸有很大的扭曲。还有一看就明白的,真实系的面部配置中,大多被设定的脸颊或鼻根的骨骼极少。这样着色在平面上难以反映变化,所以并不是要设定最低限度的骨骼,而代替那些皱纹表现的,要用专用的3D模型(图C)来实现,图D。顺便说一句,图C的两个白色圆板是眼睛的高光,这样可以把皱纹以及高光从shader上分离,容易按意图做出表情。

     

    动画独特的不合常态的关节
            图A作为人的动作当然相当的正确,但是按动画来看就太过死板缺乏趣感。正是像图B那样柔软有弹性的形状变化才符合动画片。为了实现那些,要用基础的骨骼(图A)制作动作后,再把子阶层模型顶点网格通过自由骨骼对权重进行控制(图C的红色部分),这样就可以制作动画独特的柔软性。

     

    大胆夸张的动画风格构图
            在实际的镜头中,一旦让被摄影对象的人完全的进入画面,怎样努力也如图A那样有透视的限制。所以为了得到如图B那样动画型的夸张的构图,要通过Biped的骨架的辅助动画中加算上Scale,让其在物理性上巨大化。

     

    面部的作业环境
            使用了以前的目标变形(Morph Target)是因为要在本作中变更骨骼,如图那样,准备了选择器,工作中可以迅速的对应新的手法(图A)。还有,由于这个变更,表情的数据库管理化就成了问题,但是可以把表情的骨骼和标准的姿态的骨骼的差值做成文本档案化,解决了输出脚本的制作问题。由于这样,通过目标变形能够混合两个表情的同时,还可以转移到拥有同一个骨架构造的其他角色上做出类似的表情,对工作的效率化,还有风格的统一等有很大的作用。

    再现手绘触感的背景和特效
    不管从哪里看都可以作为一张画而成立的拥有统一感的背景,还有艳丽有活力和动感的特效。在制作这些中不断的加入了独自的创意方法。

     

    背景的概要
            上图是战斗舞台的其中一个,角色的移动可能范围按实际尺寸是直径50米,最远景是500米左右,大约用了5万多边形做成。担当背景制作的四所铁矢先生表示,开发当初好像有40万到50万多边形,一边要维持品质一边要把多边形消减到能够实时绘制处理的水平,很辛苦呢。

     

    丰富的河流表现
            让水面上反射出美丽阳光的河流(图A),如何能用一张平面的多边形作出来(图B)。使用的贴图有,color(图C),height(图D),normal(图E),shadow(图F),环境(图G) 5个种类,水的颜色在shader参数里设定,在颜色上也可以增加高度的信息。环境贴图拥有alpha通道 ,加入了模拟的HDR信息(图H)绘制太阳一类的效果。像素描绘的结果,会显出很强的Glare Filter效果,花了很多工夫。

     

    背景的贴图
            在贴图上使用了Albedo贴图(图A)和阴影贴图(图B)。根据适用的地方,在2048X2048到256X256像素的范围内选择尺寸,使用DXTC格式总计30MB的程度。还有,用Photoshop的插件做成Mipmap来使用(图C)。虽然由于这些贴图的容量会增加33%,但和没使用时比起来,GPU的处理负荷减轻了接近一半,那个效果是非常大的。

     

    背景的作业环境
            公司内部在构筑开发环境上也倾注了全力,在背景制作中使用了30台以上的wacom的21寸液晶绘图板。更有应该注意的是在负责人的头上,有个防止面对监视器时映入灯光的特别定作的圆盘形的遮光罩。这是日产汽车和slik共同开发的产品,由于这个抑制了照明,实现了没有那么明亮的作业环境。

     

    重视画面效果的DOF
            图A是应用DOF前的状态。在这个基础上由于适用了从ZBuffer中抽出的(unfocus mask),虽然再现出了DOF,但是因为在角色上有轮廓线(Outline),模糊的部分和不模糊的部分的差距鲜明的显出了,轮廓若隐若现,图C。因此像图D那样在unfocus mask本体上加上Blur来抑制上述的不合适,图E。实际仔细看来,本来的焦点应该符合背景而显得模糊,但考虑图像的融合还是采用了这种方法。顺便说一句,Blur使用了高斯filter来处理,可以很轻松的完成。

     

    火焰的特效
            图A里是佐助的[豪火球忍术],图B是实际的模型构成。这个特效可以由中心的球体模型(图C),平面的正面化模型(图DE),对应移动的生成平面多边形的轨迹模型(图F),这三个部分构成,把这些合成后,可以实现拥有2D的样子并在3D空间中没有不协调感的特效。还有,适用于Flare或轨迹的动画用贴图如图所示各自八张的循环,每个一种,但是担任特效的下田星儿先生表示,最初是做成了平均性形状的多边形,在实际组合时让多边形变形,即使用一个种类的循环动画也能够表现出形态丰富的火焰。顺便说一句,这个特效为了再现动画版的时间,虽然是一秒内8到15个画面,但游戏本身是按30fps活动。试验性的尝试过60fps和15fps,但在60fps中不能流畅的做出活动的动画,一旦用15fps做出动画,作为游戏的反应就变差了。

  •    In Sucker Punch Productions’ PS4 open world game inFAMOUS Second Son, players control character Delsin Rowe, a Conduit who is able to summon superhuman powers from his environment and other Conduits. In order to realize the necessary particle effects for the game’s signature smoke and neon powers, Sucker Punch relied on a powerful expression-based particle editor and advancements in lighting and real-time curl noise generation. We talk to senior visual effects artist Matt Vainio about the artistry, tools and techniques behind the work.

       声名狼藉:私生子是由Sucker Punch公司出品的PS4开放世界观的一款游戏。玩家控制角色Delsin Rowe拥有吸收环境来召唤超人力量的能力。使用粒子特效来实现实现游戏标志性的烟雾和霓虹力量,Sucker Punch依靠一款强大的特效编辑器以及先进的即时灯光螺旋噪点。资深游戏特效师Matt Vainio将介绍游戏中使用的艺术效果和工具技术。

    Neon powers concept.

    fxg: There’s a distinctive look to the powers in this game – how did you approach the art direction of them?

         游戏中有独特的特效力量表现。你们是如何达到这样的美术效果的?

    Vainio: From the very beginning, one of the pieces of art direction we had was to make the powers feel like another character in the game. They had to have a personality all of their own and feel like something the hero was manipulating. They were not just summoning powers and running wild. The director wanted to make sure the player was in full control of the powers they were manipulating.

         从游戏的前期,游戏美术的一方向就是我们想让力量效果表现看起来是另一个游戏角色。英雄操控的力量需要看起来个性十足。不仅仅是召唤力量和狂野释放。制作人想要确保玩家能完全操控这种力量。

    Smoke: dash concept.

    fxg: How did that manifest itself into actual concepts?

         如何把概念设计实现到实际游戏效果中的?

    Vainio: There were a lot of fluid effects and curl noise that we would constrain to a shape like a sphere or a cylinder. So we actually set limits on some of the noise to help constrain it to a shape – and that really helped feel like it was being controlled by the hero.

         我们把许多的流体特效和扰动效果约束在一个像球体或则圆柱的形态。为了让特效感觉是被英雄在操控,我们设置扰动限制来形成具体的特效造型。


    Curl noise example.

    Even for smoke, we didn’t want it to waft away continually all the time. The hero would have links to his smoke via this wispy tendrils. We had to make sure they were part of him. When he dashes, it’s one smooth motion and the smoke isn’t left behind in the world. It’s part of him and carries with him and re-forms into him.

         即使是烟,我们不希望它是飘荡持续很长时间。我们需要确认英雄拖尾的烟就像纤细的卷须是他身体的一部分。当英雄飞奔身后的烟在画面中有一个平顺的动态。它们是英雄身体的一部分,随着英雄带来,又重新形成了英雄。

    fxg: What was some of the research you looked into in terms of turbulence, fluids and particles?

         是什么给了你启发来实现这些扰动,流体和粒子?

    Vainio: We looked at a lot of movie-grade tools when we were developing our own toolset. So we were using 3ds Max and Krakatoa. We used a bunch of Houdini and Maya Fluids too. A lot of it was trying to get a sense of how other tools worked in creating turbulence, and bringing that back into our own toolset.

       当我们开发自己的制作工具的时候我们参考了很多的电影级软件。我们使用的是3ds Max和Krakatoa。我们也使用过一些Houdini和Maya的流体工具。在借鉴了很多软件如何实现扰动的方法后我们让它应用在自己的工具集里。


    Neon: dash example.

    You can get a lot of real reference for smoke, but you can’t get a lot of real reference in neon – in the form that we needed it. So we had to get some next best representations in terms of fluid and aerosol gas that would stand in for plasma gas. And we researched a lot of stuff written by Ken Perlin who created Perlin Noise. We used a variant of Perlin called Simplex to create a lot of our patterns in real-time. A fair bit of effort went into making it possible to run that in real-time on hardware.

        你可以找到很多真实烟的参考,但是你不能找到很多我们需要的霓虹形态的真实参考。所以我们不得不在流体和喷雾气体形成的等离子气体中寻找替代效果的表现。我们研究了许多Ken Perlin写的关于噪点的资料。我们使用其中称作Simplex的一个变种来创造自己的霓虹即时效果。随后又做了一些努力来让它能在游戏硬件环境中运行起来。

    fxg: How did you orchestrate the relationship between the hero and the powers he is conjuring?

        你们是如何协调英雄与他所创造的特效之间的关系?

    Vainio: Everything is procedural in our game, nothing was really animated. It’s done with expressions and math. The character animates with a bunch of blend animation, so we can’t really be assured of any given animation – it’s actually a weighted percentage of up to five or six anims. So what we really needed to do was access all the joints of the character so we could create a spline out of different joints and use that as reference to emit from or to.

       所有一切都是实现程序实现的在游戏中,没有真实的动画设置。都是由数学表达式来完成。角色动画使用一系列的动作融合,所以我们无法保证添加动画,因为它的权重百分比由五到六根骨骼构成。我们真正需要做的是连接角色所有的骨骼来创造一条曲线在不同的骨骼之间,使用它来作为粒子发射器的参照。


     

    Drawing neon powers.

    We actually created a low-res version of the hero’s mesh and we were able to use that as a positional locator for corresponding from and going to. It’s called a particle mesh – it’s basically emitting from the surface of the hero and that’s picking up all the different colors of the hero and the normal directions that the triangles respond from.

        我们实际上创造了英雄的低网格版本来让我们能使用它作为位置定位来匹配特效出现和消失。它被称作为特效网格,从英雄表面拾取不同区域的颜色和基于模型法线来发射粒子。

    Particle mesh emitter.

    fxg: Aside from the particles, there’s also the ‘ribbon’ look – can you talk about those?

        除了粒子,还有一些拖尾丝带的表现,你能谈谈那些么?

    Vainio: Some of the powers had a lot more of wispy look, and that’s what we used ribbons for. It was a different way to triangulate the billboards to have a more continuous look. The UVs stream across that entire set of ribbon as it comes in. We were able to do a lot of interesting things with the faces of that ribbon coming through too. If you can envisage a chopstick in the air, you can rotate or twist that chopstick and actually create that plane.

        一些力量的效果看起来有许多纤细的形态,那些是我们使用了拖尾丝带制作。那是用不同的方式用公告板(永远朝着摄像机)三角面来演算出连续的效果。UV分部于整个拖尾丝带当它产生的时候。在丝带拖尾的表面生成的过程我们能做很多有趣的事情。如果你能设想空中有一个筷子,你能旋转和扭曲这跟筷子并且能创造平面。


    Smoke: dash ribbons example.

    fxg: How did you specifically approach the smoke effects?

        你是如何实现烟雾特效的?

    Vainio: For the powers, all them tended to contain a billowing volumetric kind of smoke, some ribbons and ash and embers. We tended to use those in different portions depending on what the effect is supposed to communicate. Early on we had a lot of design challenges with the smoke feeling like it had weight, impact and force. Smoke, by itself, you don’t really have a lot of sense of those things. So we would add a lot of ash and embers at the head of this stuff.

        这个力量特效包含体积翻滚的烟尘,一些丝带拖尾和灰尘余烬。我们倾向于使用在不同的部分给予不同的需要表达的效果。早期我们有很多设计上的挑战对于烟感觉它需要有重量,影响和力量。烟本身你是没有太多的感觉。所以我们增加了灰尘余烬等东西在里面。


    Smoke: dash example.

    I came up with the analogy that if someone blows cigarette smoke in your face, it’s kind of annoying, but if somebody puts out a cigarette on your face, that’s going to hurt really bad. So that drove the idea with smoke, that the smoke needed to be the contrail to the embers. So the embers were the lead in the powers, and the smoke was behind.

        我想起一个比喻,如果有人朝你脸上吹香烟灰,这会让人生气。但如果有人在你脸上熄灭了香烟,那样伤害会很糟糕。所以基于这个想法烟需要在轨迹上有灰烬。灰烬成这个特下的主体而烟在后面。

    fxg: What was the particle system engine that you used?

        你使用的是什么粒子系统?

    Vainio: It’s a real-time editor in the engine. It’s an expression based system and we compile these expressions into shader language. This shaders are then run on the PS4’s GPU. And this allows us to have a ton of particles. These are known as compute shaders on the GPU.

        它是引擎的即时编辑器。它基于表达式系统我们编译这些表达式成为材质语言。这些材质可以在PS4的GPU下运行。可以允许加载海量的粒子。被演算为GPU材质。


    Particle System Editor in action.

    Part of the reason we went with an expression system was because at the outset of the project we had a few ‘knowns’. One of those knowns was that we were going to have a lot of different powers and we weren’t really sure what the designers were going to ask us to do. So we tried to build a powerful and flexible system with the goal of knowing we’d have to create a lot of crazy stuff.

        部分原因我们开发基于表达式是因为项目前期我们有一些共识。其中的一个就是我们会设定很多不同的力量并且我们不是很确定设计师需要我们做什么。所以我们尝试去构建一个强大而灵活的系统来满足将来要创造很多疯狂东西的目标。

    Early on we tried to come up with particle editors that were artist friendly, but we found that we were a little bit too constrained by the user interface for this. One of the downsides of one of the node-based editors – what you’re restricted to is what the UI can let you do. You’re restricted in what the engineer has exposed to you. But with an expression based system, everything’s exposed and if you want really complex motion but the math is too complex, we can hand it off to one of our high level graphics engineers who can create part of the sim for us and hand it back. So it proved to be a really powerful tool.

        早期我们想开发粒子编辑器来方便美术,但是我们发现用户界面有过多限制。基于节点的编辑器一个不利的地方,局限在于UI能让你做什么。你被限制在引擎提供给你的有限。但是一个基于表达式的系统,所有的一切都公开。如果你想要真正复杂的动作但是数学又太复杂,你可以交给我们的高水平图形工程师,他能为我们创造部分模拟效果并交回给我们。所以这被证明是一个强大的工具。

    Keyframe editor.

    fxg: How did the neon-looking powers differ from say the smoke powers?

        霓虹效果力量与烟的力量有什么区别?

    Vainio: When we using things like the curl noise, for smoke we would use multiple octaves of curl noise that gave you large and small waveforms that gave you really crunchy noise. But with neon we tended to use more single waveforms so they were very smooth. Also, with the containment I mentioned earlier where we were constraining that shape.

        当我们对烟使用像卷曲噪点,我们会使用蜷曲噪点多个八度可以提供大小的波形产生真实松脆的噪点。但是霓虹我们倾向于使用单波形这样可以更加平顺。此外,我前面提到的约束控制可以约束形态。


    Neon: dash example.

    fxg: Can you talk about how the neon powers interacted with the world of the game?

        你能谈谈霓虹力量是如何与游戏世界产生互动的么?

    Vainio: One of the art direction goals was that everything had a lingering lighting feel. So most of our stuff used a lot more emissive qualities. We’re casting a ton more light with the neon power, so light became a lot more important and central than it did with smoke. So we had to create a bunch of more tools to be able to emit light. We had to emit lights from pretty much any of the particles that were spawned. Not every particle does cast light because it doesn’t really give us the right look. We tended to use a lower amount of bigger lights that gave us an aggregated look. If we used thousands and thousands of tiny lights, not only is it performance expensive but it doesn’t really give you the same look.

        美术效果的一个目标就是所有的事物都有一个残存的光感。所以大多数的东西都使用了发光的物质。我们释放大量的光源在霓虹特效中,让光比起烟来更加核心重要。我们不得不创造一堆的工具能够发射光。我们不得不从发射的任何粒子中产生光。不是每个粒子都主动产生光因为这样不会让产生有正确的效果。我们倾向于使用降低一个大数量的光来产生一个聚集的效果。如果我们使用成千上万的微光,这样会耗费效率反而不会有同样真实的效果。


    Neon: draining example.

    Technically, what we would do, there’s actually an emissive texture that’s mapped to a mesh that’s on that sign. Our game looks at the emissive texture and looks for texture color values that are above a certain threshold. If they are, it will actually spawn a particle there. So that way it can source the color and the facing of that sign.

        从技术上讲,我们所做的,在标志网格上贴一个真实发射的贴图。我们游戏看到产生的贴图图片颜色数值高于一定的阈值。如果是这样,这会实际在那里产生一个粒子。这样就可以设置光源的颜色和记号的朝向。

    When the particles are spawned, we kill all the emission values on the neon signs. We also turn off any reflections and any lights the sign is casting, so there’s an immediate change-over. The particles also spawn their own lights in, and they swirl down to his hands.

        当粒子生成的时候,我们结束所有霓虹灯的发射值。我们也关闭任何发射和灯光产生的记号。有一个立即切换。粒子也会产生自己的灯光,向下旋转到主角的手中。

    Neon: drain particle mesh.

    fxg: They also cast shadows – how was that achieved?

        它们也产生阴影—是如何实现的?

    Vainio: Yes, one of our big goals was to make things feel like they sat in the scene correctly. We didn’t really want things to feel like they were ‘composited’ on. That was a big focus for us, so it was important that all the particles cast shadows – they receive shadows and cast light and they blend through all the techniques. A lot of games out there don’t necessarily blend their particles with the haze and so the black smoke will look like black really far away – we wanted to make sure we didn’t have any of those kinds of things in there.

        是的,我们一个大的目标就是让事物感觉它们正确生成在一个的场景里。我们不想让事物感觉是被合成的。那是我们很关注的地方,所以所有的粒子产生阴影这是很重要的。它们接受阴影也产生光同时与所有的技术融为一体。许多的游戏里不一定叠加粒子同雾霾,所以黑烟将看起来黑且远。我们想要确定这些问题不会在我们游戏里面。


    Particle system editor showing casting of lights.

    We also used a physically based rendering approach – it’s an energy conserving model so that the reflection off of a shiny surface doesn’t send back more energy than actually went into it. Along with that comes exposure from the camera. So we had to make sure the particles were reflecting the right exposure for all the different times of day.

     我们同样使用一个物理的渲染方式,它是一个耗费节约的模式,它使一个没有反射闪亮的表面不发回更多的能量比真实走进。连同来自摄像机的曝光。所以我们必须确保粒子能反射正确的曝光在一天中不同的时候。

     

  • 2012年09月02日

    色彩原理【转载】 - [特效理论]

    http://tut.aboutcg.com/2d-tutorials/colour.html


    colour_summon_finished

    Learn how to control your colours, and take advantage
    of different colour theory principles to make your art pop!

    Hands down, on the ranking of the most important art principles, you find colour theory on a solid second place, only beaten by the most important theory; values. From all of the different aspects of art studies, colour studies are probably the most studied of them all, and it’s probably also the hardest one to master. Staying on top and handling the most important principles of art will give your advantages in the art scene for sure and shouldn’t be overlooked.

    学习怎么控制色彩,在不同色彩理论中吸取精华,让你的艺术作品变得更受欢迎!

    我们按照艺术理论的重要性来排列,你会发现色彩理论排在第二位,仅次于最重要的:明度(这里指的是用素描关系绘制体积)。在艺术研究的不同方面里,色彩研 究也许是最深奥的。而且也许是最难精通的。我们能肯定的说,掌握和处理最重要的艺术原理能够让你的艺术作品增添魅力,而且这也是不容忽视的。

    The main characteristics of colour

    The principle of colour might seem simple enough, but taking the four main characteristics of colour into consideration; hue, value, saturation, and temperature, there are really no limit of what you can do with this if you know a little about how it all works. There’s a lot of theory and studies on colour principles, and you don’t need to know them all to be able to create nice compositions of colour, but knowing a few of them is a good start for any artist.

    色彩的主要特性

    色彩原理看起来好像很简单,但是我们来考虑色彩的四个主要特性:色相,明度,饱和度(也就是纯度)和色温(色温简单说就是平常大家说的冷暖),如果你对这 些色彩特性是怎么工作的了解一点的话,那么它们对你该怎么去画是毫无限制的。色彩原理的理论学科有很多种,不是需要了解所有理论才能创作出好的色彩作品, 但掌握些理论对任何一位艺术家来说都是个好的开始。

    Simultaneous contrast
    同步对比

    2

    All colours and even greyscale values are influenced by their surrounding colour, and can be used in several ways to create the expression you’re aiming for. This is important to keep in mind when choosing your palette, and if you’re planning on having a strongly saturated and dominant background colour, make sure to put the influence of this colour into account early in the process when blocking in your composition.

    Surrounding colours also influence grey-scale surfaces. Look at the centre example and notice that the grey circles seem to absorb some of their surrounding colour. This can be used to create some very special and interesting effects in your compositions. The grey circles at the bottom all have the same greyscale value, but are influenced by their surrounding value and therefore appears to be different from each other.
    所有色彩甚至包括灰色,都会受到周围颜色的影响,色彩的这种特性可以通过各种方式应用到你的作品当中。当你选择色板的时候考虑到这一特性是非常重要的。而且当你打算用饱和度很高和面积很大的背景颜色时,确定在色彩草图之前的阶段就把颜色相互影响的特性考虑进去。

    环境颜色同样会影响到灰色,看例图的中间一栏你就会发现灰色的圆点好像吸收了周围的颜色。这可以在你的作品中创作出一些特别而且有趣的效果。最下面一栏的例图中,圆点的明度都是一样的,但他们周围的颜色对圆点产生了影响,使它们看起来有了不用的明度。

    The power of complimentary colours
    补色的力量

    3

    Now, let’s take a look at that slightly intimidating device known as the colour wheel; what would an article about colour be without one, right? When working digitally, the importance of the three primary colours (the triangle in the centre) aren’t that important as you mix colours differently on the computer, but you can still use this wheel to figure out your complimentary colours. If you look at the outer ring, the colours positioned perfectly opposite each other are complimentary to the other; red versus green, blue versus yellow, and so on. These colours have the strongest contrast, colour-wise and are a good choice to use in areas of interest to bring out attention to specific part of your painting. Mix the two complimentary colours, and you end up with a neutral colour.

    现在我们来看下这个有点诡异的东东,大家都管它叫色环~色环对于色彩作品来说是多么的不可或缺啊,是吧?当我们进行数字绘画时,三原色(中间的三角形就是 三原色了)的重要性比你自己在电脑上调出来的各种颜色差远了,(其实传统绘画不也这样嘛。。。都用原色画会崩溃的。。= = 老外废话多?) 但这个色环会指导你补色的运用,外环上颜色的位置很好的显示了补色位置,红对应绿,蓝对应黄,等等。(色环上相对180度的颜色互为补色)这些颜色有着很 强对比度和色性,用这些颜色来突出画面的视觉中心,增加画面趣味性是个不错的选择。把互为补色的颜色混合得到的是偏中性的颜色。

    4

    Simon Dominic is using complimentary colours on this painting to bring out the purple birds towards the green main element in the composition.

    Simon Dominic(上图作者名)在这张画中,通过运用补色使在大面积绿色元素上的紫色小鸟凸显了出来。

    5

    Colours have weight, and it’s quite simple; the darker the colour, the more solid and heavy the painted object will appear. The brighter colour and values, the lighter and more delicate the painted object will seem. You should use colour weight to balance your composition, and for creating striking moods. Philip Straub is using colour weight to bring out the main element in focus and balancing the composition in this painting. This is a great way to empathize areas when you’re working within a limited colour range and don’t want to use complimentary contrast as an attention seeker.

    颜色有重量感,非常简单,深色看起来更重,存在感比较强,亮色和明度高的颜色看起来更轻更精致,亮色画出来的东西比较虚无飘渺。画面需要色彩的量感来平 衡,并以此取得吸引人的气氛。Philip Straub(上图作者名)用颜色的量感来突出主题元素使画面得到平衡。这是当你只能用有限的颜色和不想使用补色的时候,用颜色的量感突出视觉中心的很好 方法。

    Stick with it
    确定配色表

    Once you’re in the process of deciding on a colour scheme, no matter if you’re starting off with colours from of the first rough, or if you already have a grey-scale version ready of your composition; create your colour scheme and stick with it. Altering too much throughout the process might mess things up more than you think. Minor tweaks are naturally a necessity if needed.

    一旦你开始决定配色方案,无论是从最初的色彩草图还是已经完成的黑白稿版本,确定你的配色并按照配色表的颜色去画,在绘画的过程中不停的修改颜色会使画面比你想象的还要糟糕。如果需要的话做些小的调整时可以的。

    Add depth with colour 
    用颜色制造景深

    6

    An efficient way of supplementing depth to a painting with already good values is taking advantage of the effect of warm and cool colours. This can be done quite simply by using warm colours on elements close to the viewer, and keeping the background elements and backdrop with cool colour. This is because warm colours seem to come forward and cool colours seem to recede away. You can also use the same effect to also create striking contrast in mood, as shown on this beautiful landscape by Gary Tonge.

    有效的给一幅作品增加景深的方法是利用色彩冷暖的优势,这么做很简单,只要在近景部分用暖色调,然后保持背景元素的冷色调就可以了。这是因为暖色有向前的 感觉,冷色有后退的感觉。你可以用同样的方法创造出令人着迷的气氛。我们展示下Gary Tonge(上图作者名)这张漂亮的风景画。

    Light, shadow and colour
    光影和颜色

    7

    When painting any subject, light and shadows are what will build the shapes in the piece. Values are extremely important to be settled early in the process, and when the final thrust of the creative effort is set by adding colour to the light and shadows, your piece will really spring to life.
    Keep what was mentioned earlier in mind about adding depth with colour and settle for a colour for your main light source, and a colour for your ambient light. The ambient light is the space that your character or scene is set inside. If you have a clear blue sky, your ambient will be blue because of the massive blue sky-dome covering the scenery, and this colour will influence all of your objects from all angles, and especially the shadowed areas. This is why we often see that blue or purple shadows or backlights/rim-lights always seem to work out nicely in paintings.

    Here, Izzy Medrano is using a main light source from the left, and using the surrounding, blue, ambient light make up the shapes of the shadowed side of the character. 

    画面中,光影塑造形态,体积和素描关系在初始阶段中确定下来非常重要。当最终的创造性的效果通过颜色和光影加到画面中的时候,你的作品就会鲜活起来。
    保持对作品的第一感觉,包括颜色塑造的景深,对主光源和环境光制定的色调。环境光取决于你在画面中设定的角色或场景的空间,如果画面中有蔚蓝的天空,你的 环境光应该是蓝色,因为宽广的蓝色天空覆盖了整个场景。这种蓝色会从各个角度影响到场景中的所有物体,特别是阴影部分受环境光影响最明显,这就是为什么我 们经常能看到画面中很漂亮的蓝色或紫色的阴影或逆光下轮廓的蓝紫色高光。

    这幅作品zzy Medrano(上图作者名)在画面左边设置主光源,然后利用周围环境,蓝色环境光塑造角色阴影中的形体。 

    Environmental influences
    环境的相互影响

    9

    If you want to take the previous principle about light and ambient light and complicate it a bit further, you can start thinking about what material the different surfaces in your scene is made out of and how they all will reflect and bounce off light. The light will then cast shadows which will have to be sorted out as well. Looking at this alley painted by Levente Peterffy, you will see that he has taken advantage of ambient lighting to establish depth, and shadows for grounding the lit parts of the composition. Softening out the lit and saturated parts makes it appear as if the air around appear lit, which makes for the delicate and soft expression in this piece.

    The same goes for semi transparent colourful surfaces, or solid yet strongly saturated elements where the colour of the surface will bounce off and tint the surrounding areas, both the air and surfaces, as Daniel Kvasznicza is nicely showing on his Chinatown conceptual sketch.


    如果你想用之前关于光线和环境光的理论去继续深入你的作品,你可以考虑下作品中各个物体不同的材质表面,这有助于你想象出各物体间相互影响的反光。光投下 阴影,场景场景中的物体也就区分出来了。我们看Levente Peterffy(上图作者名)画的这幅小巷,作者用环境光来表现景深、地上的阴影和画面中的光感。柔化光线,饱和度高的部分会体现出光感,就像空气弥漫 在光线周围,使整张画面色调柔和精致。

    这同样适用于半透明物体的表面,或者固体但饱和度高色彩浓烈的物体,物体表面的颜色会反射影响到周围区域,包括周围的空气和物体表面。就像Daniel Kvasznicza(下图作者名)漂亮的展示出了他的唐人街概念草图。

    10

    Areas of interest
    趣味中心

    11

    The same way as using strong contrast when working with values, adding attention to specific areas in your composition can and should be done with colours as well. Therefore; hold your horses, and don’t use your colours on full strength all over your canvas, but reserve the strongest hues for areas of interest. Using complimentary contrast or simply stronger hues of a specific colour will work as an automatic attention seeker, and can be used cleverly if you are aware of how to take advantage of this effect. Remember that a painting with dull hues all over will be hard to read. You need strong colours to function as eye-catchers, but only in the right places.

    In this painting, Simon Dominic has taken great advantage of colour saturation to empathize the unique features of the main character, and left the rest of the composition less saturated, yet within the same colour range. Add more saturation to the background and the painting would be hard to read.


    我们应该而且必须同样的方法加强明暗对比和色彩对比,使画面中特殊区域引人注意,因此,请再三考虑,不要使整个画面处理的太平均,要让最漂亮的颜色在画面 的趣味中心,对比强烈的补色或是饱和度高的颜色会自动抓住人的视线。你如果意识到这种效果带来的好处,就会聪明的运用它。记住用暗淡无趣的颜色画出的作品 会让人很难接受,你需要用浓烈的色彩抓住人的视线,但一定要用在正确的地方。

    在这张图里,Simon Dominic(上图作者名)利用饱和度高的色彩的优势,使视线集中到了这个独特的角色上,让画面其它部分的饱和度降低但不能差别太大。如果使背景饱和度变高整个画面会很难看。 

    Single-coloured surfaces
    单色表面

    Unless it’s the expression you’re aiming for, having clean, single coloured surfaces can make your painting seem a lot duller than it could be, no matter how strong that colour is. Texture is key; try messing up that surface a bit with some textured brushes, or a simple noise and you’ll see that simple measures like this can create interesting surfaces and make your painting come to life, even if the noise is within the same colour range as the clean base colour.

    除非有特殊需要,干净单一的表面会使你的画面单调无聊,无论这颜色多纯都一样,纹理是关键,试着用些笔刷让表面增加一些纹理或是简单的杂色。你会发现这些简单的方法可以创造出有趣的表面,并使你的画面鲜活起来,甚至这些杂色下是相同的单一基本颜色。

    Using complimentary colours to bring out areas of interest
    用补色突出趣味中心

    12

    An expressive and efficient way of adding depth and settling where to put the areas of interest is by using complimentary colours. On this painting by Philip Straub, he is using the complimentary contrast of the red mushrooms and the green backdrop to make his composition easy to read. He is also taking advantage of colour saturation to bring out the foreground by keeping the background duller both in values and colour saturation.

    用补色来增加景深和刻画趣味中心是一个有表现力而且有效率的方式,在Philip Straub(上图作者名)这幅画中,他运用红蘑菇和绿背景的补色对比使作品变得好看,同样他利用高饱和度的优势使前景突出,背景的颜色饱和度稍稍降低使背景退到远处。

    Using warm and cool colours to create depth
    利用冷暖制造景深

    13

    Gary Tonge is showing a great example of how good composition and colour handling can make a painting really pop. By using warm colours close to the viewer, and cooler colours further away, the combination of colours will add depth to the piece because warm tones seem to come out towards you and cool colours recedes. The nice perspective of this specific piece and the placement of the point of interest are also helping.

    Gary Tonge(上图作者名)给我们展示了很好的例子,通过颜色处理使作品突出出来。作品中近景使用暖色,越远的物体颜色越冷,这种颜色组合会使增加景深,因为暖色感觉向前,冷色有后退的感觉。整张画漂亮的透视和趣味中心的布置安排都使画面增色不少。

    Colours influenced by environment
    环境对颜色的影响

    14

    The atmosphere in this painting by Gary Tonge is bathing in a beautiful, golden light, which is influencing the entire scenery. Any other stronger colour or complimentary contrasts would break the wonderful mood he has created here. He is also relying on colour weight to make the structure closest to us appear more solid and heavy. 

    Gary Tonge(上图作者名)这张作品的大气中弥漫着漂亮的金色光芒,光线影响着场景中所有物体,任何其它的颜色和补色对比都会对这漂亮的气氛造成破坏。他这里也利用颜色的重量感使前景的建筑看起来结实厚重。

    Step 1, getting the base values up
    步骤1,画出基本的明暗光影

    15

    As already mentioned; having good values is crucial for your painting to work. No matter how well the colours are composed, bad values can still break the piece and it will be hard to read. If it works well in grey-scale as well as in colour, then you’ve succeeded. 

    就像之前提到的:好的明暗关系对整张画来说至关重要。不论色彩多漂亮,差劲的明暗关系仍然会破坏效果,使整张画很难看。如果画的明暗和色彩一样好,那么你就成功了。

    Step 2, applying basic colour
    步骤2,铺大体色调

    16

    I wanted to create depth to this painting, even though it’s a very simple composition. I decided to use a cool colour for the background because cool colours appear to recede back into the canvas. I wanted the woman’s skin and dress to consist of a warm palette, along with the closest tentacles to drag them closer towards the viewer and away from the cool background.

    我想在这幅作品中加入景深,即使这是一张很简单的画。我决定用冷色来画背景因为冷色给人向后退的感觉,我想让图中的女人皮肤和服饰偏暖,和最近的触手一起拉到近景,远离背景。

    Step 3, adding the final pass
    步骤3,最终完成

    17

    Adding more contrast around the woman’s head will work as an attention seeker, and I also applied a blue backlight-rim around her and on the closest tentacles to link them to the blue background. If you look at the top of her dress, you can see how the blue light is bouncing off her dress, and also how her raised arm reflects some warm shades of light as well. 

    在女人头部周围增加对比,使之更吸引眼球,加入蓝色背光提亮女人和触手的轮廓,使得在颜色上和背景联系起来。在她的衣服上部你会看到蓝色反光是怎么反射到衣服上的,还有她抬起的胳膊上反射的光的暖色阴影。

  • http://xuelang-bian.blogbus.com/logs/106832916.html

    如下图,我们就以这个火焰效果为例吧:)

    我们首先要介绍制作这个特效所使用的软件

    3DS MAX

    FMUE FX插件

    KRAKATOA插件

    插件的下载地址:http://www.verycd.com/topics/2747331/

    呵呵,就别再找我要了

    首先如上图,我们要创建这样一些物件:plane, donut, fume fx, fx object src

    创建完了么,我们按照图片上的去拾取关联吧,fx object src去拾取donut

    然后选择fume fx的橘黄色框,在修改面板中激活他的选项框,在objectsrc栏中拾取fx object和plane

    我们找张贴图付给donut,贴图我随便找的哈,你们也就随便吧,呵呵。

    我们再次打开FUME FX的属性栏,把贴图同样给smoke和Temperature.

    我们来设置一下donut的UV动画,记得先在修改中添加一个UVW Mapping,选择贴图plane的平铺方式,然后点开子层级 选中,做Z方向的旋转动画。

    UV的缩放动画

    DONUT自身的半径1缩放动画

    DONUT自身的半径2缩放动画

    我们再来设置一下FUME FX属性面板的参数

    OK,开始解算,看,能看到大致的样子了,但这不是最终效果。

    我们按下快捷键“6”,打开PF面板,连接方式如上图,参数如下图:

     

     

     

     

    现在你能像我这样看到场景中的粒子了么?

    我们现在要用牛B 的KRAKATOA粒子加速器了,注意从渲染设置面板(快捷键F10)中调出。

    设置的参数如下:

     

    都设置好了么,迫不及待要看效果了吧,赶紧按RENDER吧,然后耐心等待一会.....

    哇,神奇的效果出来了!还不赶快去试试。

  • ‍Apophysis是一款免费的炫光特效制作软件,通过计算分形图案可以产生丰富多变的炫光或者奇幻图案,非常适用于绘画和动画特效。使用上也很简单,这里我给大家简单的介绍下用法。

    1.首先打开软件,我们可以在左边的分形库里面找一个预设好的特效。

    2.点击mutation按钮可以打开一个演化面板,在这里可以选择一种随机演化结果,选择后视图的图案就会变化,speed滑块控制演化速度。

    3.打开Adjust可以调节图案的亮度,伽马值,远近等等,每调一个参数预览图都会更新,很直观

    4.打开Gradient面板,这里控制分形图案的色彩,下拉菜单可以选择很多渐变预设

    5.如果你希望自定义颜色,点击这里去拾取任意一张图片,软件会自动分析色彩将其变为渐变色

    6.点击Edit进入分形变化面板,这里最直观的就是这些三角形坐标,分别代表着分形结构的不同结构与分形位置,随意拉动各三角形的圆形顶点就能改变分形图像的结构,小窗口会有预览图像,很快。

    7.右边的参数栏控制着分形图案的结构坐标,其含义是

    Variation变化

    Vars2变化等级2

    Colors随机颜色

    Triangle参数方式控制三角形坐标

    Transform整体形变

    每调接任何一个参数你都能得到预览的反馈,很容易就能知道这些参数的作用与结果

    8.进一步的复杂变形特效我们需要使用脚本来完成,在这里可以打开脚本库选择一种变形方式

    挑选一个特效处理脚本

    9.然后点击“运行脚本”,这样就能看到处理后的结果。

    10.点击文字菜单File》export flame渲染图像,注意最好保存到英文路径!

    11.最终图像

    12.最后通过Variation菜单里面的各种分形变化属性还能快速改变图案结构,接下来大家可以自己试试咯!


    软件下载地址:http://www.apophysis.org/downloads.html

  • How many of you loved that bullet trail in The Matrix?! Yesterday, it was only a fantasy to make such thing in some 3d program, but today, that fantasy became real! BryanA. leads you throught his amazing tutorial which will teach you how to make Matrix Bullet-Time Bullet Trail Effect in 3ds MAX...

     

    Matrix Bullet-Time Bullet Trail Effect
    by BryanA.

    UPDATE: Bullet Time Bullet Trail v.2 ONLINE (November 9th, 2003)!


    Do you like math? Neither do I. But that's too bad because the Matrix Bullet-Time Trail Effect works best when you know what you're doing mathematically. Don't worry though, it's nothing tough.

    For the time of it all, let's figure that we want a five second scene at 30 frames per second. Set MAX to 150 frames at 30 fps.

    Our rooftop is 200 units (now we'll scale that to feet) - so it's 20 feet across and across, making a perfect square.

    The Matrix Bullet-Time Trails fade out into invisibility. They also change shape a little after just being fired. There are three noticable stages of a trail as you see in this picture:

    1 is a more recently fired bullet (you can still see the bullet),
    2 is what we'll call a Standard Trail, and
    3 is an Expiring Trail (it's fading away and the distortion of the geometry is "leaning" to one side).

    Here's what we need:

    A 200x200 roof that will scale to 20feet by 20feet. (Super-Easy)
    A bullet (Easy)
    A trail reference shape. (Harder: Splines, Copying, Grouping, Modifying)
    A cup of coffee? :) (Lot's of sugar please)

    I used a Plane built in the TOP viewport at 200x200. Then I converted it to Editable Mesh and selected the outside polygons and Extruded them with a value of 3.0.

    The bullet. Let's first build it larger than we need it and then we'll just scale it down with the Xform Modifier. Let's use a Capsule at this size (Radius of 2 Hieght of 8 and 12 sides) in the justify viewport (where we build it will be important later). After achieving these dimensions, convert it to Editable Mesh. Use Vertex as the sub-object and grab what will be the front tip. Activate the soft selection and set it to 4 and pull it "forward" about 1.0 unit, and then let go, and then pull it again about 1.0 unit.

    Now tuck in the backside by taking off Soft Selection and starting with the last vertex. Push it in to meet the next column of vertices, then collect them in the selection and keep doing this until you have a reasonably flat backside.

    The last thing to do is apply the Modifier, Xform and Scale the Gizmo to 18%.

     

     

    Now for the Spline modeling. This isn't the most advanced spline you ever made - but it was my first time ever believe it or not. So if you know a more efficient way to do it, feel free to do so! For first-timers like me, follow this part carefully and take no shortcuts.

    We need to "trick" MAX into making one object out of several 2D shapes. For accuracy, we're also going to use Keyboard Entry for this. Here are the best directions I can give for this:

    Create one 2D Circle with the justify viewport active. *Ahem* The justify viewport active. Keyboard Entry is as follows: x=0, y=0, z=2 radius is 2.0 "Create." Magical, huh?

    Now you have to do this little step: Look up top above "Circle" and "Line" and "Rectangle" where it says "Start New Shape." Click the check mark out of there so you DO NOT start a new shape. All additional shapes you make now will apply to the same object.

    Now we continue with the Keyboard. z=0 Radius is 6 "Create."
    The thrid entry is: z=-2 r=2 "Create."

    Now click the Modifier Tab, look under the General rollout and choose Adaptive instead of Optimize.

    The following modifiers require that you build a spline in the correct, logical order as we have done. We will treat these Circles as being cut-through cross-sections of a final shape. Telling MAX to treat it that way is easy. Look under the Modifier "More" list for "Cross Section." Apply it (choose Smooth) and then look for "Surface" under the list by clicking "More" a second time. Surface will complete the procedure and create a renderable object. You can control how complex the surface is with Patch Tolology which is one of the few Parameters of "Surface."

    To prepare for the Invisiblity of the Trail, set the Animate button ON. Goto frame 10. Right-Click the Oject's Properties. Halfway down, on the justify column, see "Visibility: 1.0." Change that to 0.0. Turn the Animate button OFF. With the Move Tool, move a COPY some units away along the X-axis. Then look at the COPY'S Keys on the time slider. There should be two: one at 0 and one at 10. Select them both and move them ONE frame to rest at 1 and 11.

    With the Move Tool, move ANOTHER COPY some units away along the X-axis again. Now change the new keys from 1 and 11 to the new frames of 2 and 12. See a pattern? Make seven copies or eight if you please. Just keep COPYING in the same direction everytime, and always adjust the Keys by one frame for every new copied Object. Get this:

     

    !VERIFY! If you select ALL eight of your shapes now, do your Keys look exactly like this? If not, DELETE and do it again.

    With all eight Objects selected, make a Group of them. Name it TrailMother. Apply FFD 4x4x4 to TrailMother. Do your best to make this shape:

    What we're doing is making the Standard Trail from which all others come. Now Right-Click the Group and "Convert Selected to Editable Mesh." Here's the easy part of this; we will copy this Group and adjust the Keys again so Invisibility "runs up" the trail we're making. Move a COPY in the X axis direction to where it looks right next to the original. Select all of the Copy's Keys and move them 8 frames later (0 and 10 become 8 and 18, etc.). Make six more copies and continue moving the Keys for the Visibility. When you're done, set Perspective Viewport so that you can see the whole trail. Hitting F3 will toggle a Shaded View. With a Shaded View, hit Play and make sure that the Invisibility runs correctly. Did it? I hope so; it's confusing to reconfigure.

    Optimizing the Refractive material is the first thing we need to do before rendering too much. Following these tips will cut down the time it takes to render multiple trails in the future. Hit "M" on the keyboard. The material browswer is up and you should be in the first slot. Glossines, Specular to 0. Soften to 1.0. Extended Parameters: IOR to 1.2. Material Maps: Refraction Slot - Raytrace. Raytrace Options button (match the picture below).

    Congratulations! You just saved yourself minutes of waiting for still shots and maybe an hour for a future animation. Faster would be Anti-Aliasing UNCHECKED totally, but it looks bad that way so we have it checked to allow it. Finally, apply the Material (TrailMat) to the new, eight objects.

    With all eight big objects selected, Group them and name it Trail. Right-Click them and UNCHECK Cast Shadows and Receive Shadows. OK. Then, apply Modifier FFD (box) to the Group, Trail. Set the points to 3,2,3. It should look like this:

    For ease in the future, you will also need to adjust the Pivot Point. Move the Pivot Point to one end of the Group, "Trail." This will aid in predicting where the Trail moves, rotates and scales later on.

    Now there are several choices you get to make on your own since we have run out of time. Obviously, the Trail is way too big. It should be scaled down uniformly on every axis. Here was my decision:
    I scaled it all to 25% for the Standard phase.
    For the fired bullet phase: 28% NUScale for the radius; 65% NUScale for the length. For still pictures, you can Scale the rear Control Points to 175%.
    Just as it's Expiring: NUScale 50% for length but grew the radius with NUScale 150%; I moved it forward to meet the bullet and scaled the rear of the FFD bigger to 150%; finally, I scaled all of it to 65%.

    A word about animating. With luck, I can include one, small MAX file with one Bullet animated to cross the floor, followed by a Trail that goes through all three phases. To make another Bullet, you would simply select the Bullet and the Trail and MOVE a Copy of both being sure to alter the Keys of the copies so that two Bullets fire at two different times and each Bullet is followed by a correct Trail.

    To animate your own Bullet and Trail, follow these guide lines. You first need to set the position and timing of the Bullet. Put it WHERE you want it; Animate/Move it where you want it shot to. Adjust the two keys so that it moves as fast or slowly as you want. Now you can mess with the Trail (first Bullet, then Trail). See all the Keys to the Trail (they should only be Invisibility Keys). Adjust them so that Invisibility begins AFTER the Bullet leaves the scene. Place the Trail so that the Pivot Point is where the Bullet was fired from.

    This assumes that your first Bullet will travel in a straight, uncomplicated path.
    1. Goto the first Key of your firing Bullet and Move the Trail behind it.
    2. NUScale the length of the Trail so that it sits entirely behind the Bullet.
    3. NUScale the radius of the Trail accordingly so that it is small.
    4. Goto the end Key of your firing Bullet (Animate ON now) and NUScale the Trail's length so that it meets the Bullet.
    5. With Animate ON, choose a frame where your Trail should be "Standard." Remember it - and look 16 frames BEFORE this occurs.
    6. Under the Motion Tab, click the button "Scale" to create a Key here for locking in the Fired Bullet phase.
    7. Go BEFORE this another 16 frames and tug the center two Control Point for FFD (box) forward (towards the Bullet) a little bit.
    8. With Animate still ON, go forward those 32 frames. Now it's safe to NUScale the radius of the Trail to look "Standard."
    9. Absolutely, those Invisibility Keys should be AFTER what we've done so far. Goto the end of those Keys, and Move the Trail slightly forward and expand the backside of that FFD (box).
    10. To finish, select the center Control Point of the FFD (box) and Move them backward (towards the gun).
    11. Animate OFF.

    This tutorial was originaly written by BryanA

  •  

    ParticleGen 是一套製作多面即時 3D 粒子效果的靈活而高效的工具。擁有互動式

    即時預覽等眾多專業特色,還有曲線圖參數的精確設置,本程式完美的適用於DTP、視訊合成、後期加工、軟體開放和各種視覺化處理(爆炸、燃燒、水波、雲霧)。是該軟體的第一版本,但由引起 CG 業界的廣泛關注,對於製作特效、短片、廣告、電視包裝、遊戲界等特別應用

    particlegen

    的可實現幾乎任何效果,從爆炸到霧水。除了利用自己的粒子可以創造任何形狀的常規圖片數位節目,該軟體還允許你輸入到 3DS MAX檔案,然後你可以用粒子或動畫。

     

    What is ParticleGen?

    ParticleGen is an API that implements a two-dimensional particle generator in Flash Actionscript 3.0.  ParticleGen allows for the real-time modulation and trajectory tracking of graphical particles within an arbitrary force field. The classes can be used to create a variety of particle effects that mimic man-made and natural phenomena.

    There are several concepts related to using ParticleGen that must be understood to best utilize the package. The following is a schematic of how the various components of ParticleGen fit together.

    ParticleGen:

    This is where the main logic of the package resides.  This class contains all the methods to generate the PARTICLE STREAM and track all the particle trajectories with a given set of FORCES and MODIFIERS applied.

    The Particle Stream is a virtual construct which refers to the stream of particles ejected from the ParticleGen EMITTER. The Emitter can be configured for various patterns of emission and randomly spawns new particles from the PARTICLE TYPES in the generator system whenever existing particles die. The Emitter renders particles in the stream onto two different layers, depending on the type of particle in question. Varticles are rendered on the VARTICLE STAGE, and Blarticles are rendered onto the BLARTICLE STAGE.  The reasons for this are discussed below.

    Particles:

    PROJECTILES are graphical classes that model the state of a particle including its age, velocity, position, lifetime, mass, frictional coefficient, and visual representation as a function of its state variables. There are two major classes of Projectiles:

    Flash’s internal rendering is done using vectors, and a Varticle can dynamically create its graphical representation at run-time as a function of any number of its state variables. The disadvantage of using Varticles is that vector rendering is painfully slow.  ParticleGen will bog down the CPU while tracking only a few hundred Varticles.

    By contrast, Blarticles are bitmap-based, and are not dynamically created at run-time.  Instead, they are initialized once and are kept resident in memory until all instances of ParticleGen are cleared. Because of this, Blarticles are limited to graphically evolving as a function of a single state variable only (for example age, or velocity). Blarticles are blindingly fast to render, however, allowing real-time tracking of 10,000 plus particles.

    The Flash IDE can be used to create the graphical representations of  both Blarticles and Varticles. Each frame in the particle’s corresponding source MovieClip graphically represents a single state of the particle.

    Forces:

    FORCES are classes which act upon Projectiles to alter their velocity, analogous to physical forces in nature. At the current time, inverse square, vector,  and point-spring forces have been modeled in ParticleGen.

    Modifiers:

    MODIFIERS are classes that directly alter the state of a particle. For example, a boundary can be modeled as a Modifier.  Any class that directly alters particles independent of the Forces in the system is referred to as a Modifier.

    The best way to understand the functionality of ParticleGen is to play with the interface program, PGInterface,  that has been written for it. Although the ParticleGen implementation is more robust  than the options in PGInterface would imply, PGInterface is an effective way to tour the features of ParticleGen.

    网页版游戏特效编辑器:http://creativeprogrammer.com/PGInterface.html

    游戏特效编辑器下载地址:http://www.verycd.com/topics/119952

    官方地址:http://www.crystalpictures.eu/particlegen/en/