1975 年代的電腦圖學

Jul 12 2009

animation

原文:原文:HISTORY OF COMPUTER GRAPHICS – 1970-79

1974/1975,可能是電腦圖學界的轉淚點,主要原因之一是圖學引起更多人的注意,另一方面是圖學重鎮由猶它大學(UU)轉往紐約技術學院(NYIT)。前兩者一起看的話,就是有更多人投入 NYIT 這個新圖學重鎮。

很多人參與 NYIT,我們來小聊一下其中一位先生 – Alvy Ray Smith。Smith 是生平事蹟是少有的比較有趣的。他一開始,對於電腦圖學是沒什麼興趣的,挺多就是在學生時代,使用電腦圖學做了個作品出來(就一張圖)而已。就著,他前往紐約大學任教,因為他過去有圖學的經驗,所以紐約大學的 Herb Freeman 先生非常希望他可以投入這一塊的研究(譯注:跟上時代?!),但他實在不感興趣,不管怎麼樣都無法被說服。隨後,因緣際會下,在他因為溜冰斷了條腿後,有了時間重新思考他的人生方向而決定回去 UC Berkeley 任教時,順道去拜訪了他的好友 Richard Shoup (Shoup 當時在 PARC (Palo Alto Research Center) 開發一個叫 SuperPaint 的系統),本只想順道看看 Shoup 的 SuperPaint 的,就十來分鐘就好。結果一看,極為驚人,當下他下了個決定:電腦圖學是他今後要投入的領域。於是他加入 PARC,最後又輾轉到 NYIT(譯注:因為 PARC 不打算投入彩色營幕相關的研究,而著手寫 SuperPaint 的 Smith 就跑去找 Catmull 了)。

譯注:個人非常推薦喜好電腦圖學的人,常常去看看 Alvy Ray Smith 的個人網頁。裏頭的東西多到讓人會嚇一跳,很容易看到一張不曉得在幹麻的圖片,但整個就是很有趣,除了網站做得非常醜以外。另一方面,Smith 因為在 SuperPaint 與 Compositing 上的貢獻,先後得了兩次美國電影藝術與科學學會(Academy of Motion Picture Arts and Sciences)的技術獎項。又,Smith 也是 HSV color space 的發明人,因為有了他的發明,讓我們可以免除使用 RGB color space 來調整顏色時的痛苦,在這邊特別感謝他的貢獻。另一方面,Smith 其實也是 Pixar 的創立合夥人之一的,只是後來 Pixar 被賣到 Steve Jobs 下時,他們大吵了一架,然後他就退出 Pixar 了。也因此,Pixar 的歷史都沒有提到這個人,但這個人的確是個非常重要的人…所以說…是有心結什麼的是嗎XD 不過 Smith 手邊倒是有留下一些歷史記錄

1975 年有另兩件事,一是 Steve Jobs(當時他在 Atari)與 Steve Worniak (當時他在 HP)合作出來的第一台個人電腦,最後以美金 666.66 元賣出。同時,一位年僅 19 歲即自 Harvard 輟學的 William Gates III(就是 Bill Gates)與他的好友 Paul Allen 合創了公司 – Microsoft,他們靠 BASIC 與最後的(譯注:某種形式上偷來的) MS-DOS 成軍。最後,這兩家公司都影響了整個世界。

NYIT in itself was a significant event in the history of computer graphics. Here we had this wealthy man, having plenty of money and getting us whatever we needed, we didn’t have a budget, we had no goals, we just stretched the envelope. It was such an incredible opportunity, every day someone was creating something new. None of us slept, it was common to work 22 hour days. Everything you saw was something new. We blasted computer graphics into the world. It was like exploring a new continent.

Source: Alvy Ray Smith .

上頭是 Smith 表達了對於 NYIT 這個環境的看法,說明了一個好環境對於一個新穎技術的發展是多麼的重要。隨後,Catmull 等人發現到一些問題,然後加上 Lucas 當時想組自己的一個電腦圖學實驗室(因為他恰巧看到了一些電腦圖學的結果,他發覺電腦圖學的確是有機會拿來拍電影的),於是 Catmull 被挖角過去,連帶地一些研究人員也跟著過去了~(譯注:下頭是 Catmull 的心聲,說明了他為什麼離開 NYIT 前往 Lucasfilm)

Alex Schure funded five years of great research work, and he deserves credit for that. We published a lot of papers, and were very open about our research, allowing people to come on tours and see our work. However now there are a lot of lawsuits going on, mainly because we didn’t patent very much. People then subsequently acquired patents on that work and now we are called in frequently to show that we had done the work prior to other people.

We really had a major group of talented people in the lab, and the whole purpose was to do research and development for animation. We were actually quite stable for a long time, that first five years until I left. However, the primary issue was to make a feature film, and to do that you have to gather a lot of different kinds of skills; Artistic, Editorial, etc.. Unfortunately, the managers of the school did not understand this. They appreciated the technical capabilities. So as a group we where well taken care of, but we all recognized that in order to produce a feature film we had to have another kind of person there, movie people, and basically those people weren’t brought into the school. We were doing the R & D but we just could not achieve our goals there. So when Lucas came along, and proved that he did have those kind of capabilities and said I want additional development in this area (of computer graphics), we jumped at it.

Source: Ed Catmull .

這段期間的論文發展非常的驚人,有 Jim Blinn 的 Bump Mapping 與 Environment Mapping。在 Cornell 的 Robert Cook 發現到大家在用的 Phong illumination model 的結果太像塑膠了,於是他們開始研究一個更好的 model 來呈現世上任何東西的質感,這造就了後來的 Torrance-Sparrow model 與 Cook-Torrance model 的出現。(譯注:感謝他們敏銳的觀察力,讓我們得以有更多的工具可以使用)

至此,世界上的圖學發展中心不再侷限於一處。UU, NYIT, Lucasfilm, Triple I, NASA JPL(譯注:因為當時 Jim Blinn 在這邊! 好巧不巧,太極影音今年要發佈的 Quantum Quest 也與 NASA JPL 有關), 與後來由 Don Greenberg 於 Cornell 成軍的電腦圖學研究室,共同主導了整個圖學的發展。同時 1976 年的 SIGGRAPH 也有了展場,吸引了更多人前往,奠定了後來世上最大學術會議這個頭銜。

comments powered by Disqus