西藏之旅

我本不是一个爱旅行的人,但不知为何,西藏却是一直向往的地方。然而到了西藏才发现,这里确实是一片神奇的土地。

由于这段旅途过于传奇,我也就先不多写太多,只是贴贴照片,写写流水帐罢。

瑰丽青藏线

北京出发,乘火车经过青藏线进入西藏。由于西藏太过广袤,而很多景点都很遥远,在火车沿路观景是个相当不错的选择。如果不考虑自驾的话,火车入藏能看到不少路途美景。(感觉有相当一部分的美景是在火车上看的)

首先是翻越唐古拉山,高山冰原上行车别有一番风味,而周围的美景,绝对是此行中的一个亮点。火车穿行在永久冻土的冰原上,周围的景色,绝非其他地方能比。

过了唐古拉山,经过可可西里无人区,而后就是高原上的荒漠,偶尔会有一些绿洲。

高原湖是西藏的一个奇景。由于很多湖泊都是当地人信奉的圣湖,不能污染,没有过度开发,湖泊纯净天然。蓝天,白云加上平静的湖泊构成一副震撼人心的景象。火车会途径海拔最高的淡水湖——圣湖错那。

一路还会路过无数的雪山,对于雪山控而言,就是无尽的毒药啊。

青藏铁路一路美景无数,如果不愿坐长途的旅客,也可里考虑从西宁直接上车,仅走青藏线就是。

佛教净地

藏传佛教也是西藏神秘且独一无二的景致之一,由于佛缘不够,只是草草的参观了大昭寺,扎什伦布寺和布达拉宫。

大昭寺
扎什伦布寺
布达拉宫
布达拉宫黄昏夜景

最后附上一些经幡阵

比日神山经幡阵
南迦巴瓦前的经幡

醉美林芝

林芝位于西藏东部,海拔只有不到2900米,不会有太强烈的高山反应。同时,由于水系丰富,空气潮湿,植被覆盖率会高许多,素有“西藏的小江南”之称。同时318国道从林芝往四川段,可谓美景无数。既有美丽的林海和漫山遍野的杜鹃花,又有雅鲁藏布江奇景,更少不了最美雪山南迦巴瓦,林芝成了西藏最佳的旅游地点之一。同时由于海拔低,气候适宜,更成为西藏旅游的首选地。

鲁朗林海
雅鲁藏布江

高山圣湖

高山圣湖的美景前面已经说到,纳木错,羊卓雍错这些名字也早就令人熟悉,不多说,就上图好了。

纳木错
纳木错
羊卓雍错
羊卓雍错

飞回北京

从拉萨机场返回北京,一路也是欣赏了不少奇景,从高空看白云映衬下无尽的雪山,绝对是难忘的体验。

西藏确实是个太神奇的地方,而更神奇的则是这次旅行本身,至于在旅行之外所收获的,知情的看客自然就已经了解。而不知情的,就等后面的分解吧。

 

 

SIGGRAPH Asia 香港行

各种忙项目的时候,还是去香港开了一趟SIGGRAPH Asia,说实话 SigAsia这时间选的还真是要命…..现场准备presentation,糊糊涂涂讲完,总算抽空出去玩了一天半。整天早起,弄得回来都没力气ps了…..总之,姑且上几张旅游图随手拍算了….也算有个交代。

T3出发,很早的飞机,空荡荡的T3候机厅

登机口在T3的一端尽头,窗外刚好能看到01-19跑道起飞

往返都是乘坐这架,国航/澳门航空的亚速尔号

 

香港最好的还是夜景

只可惜镜头擦不干净就是各种悲剧

另外附送一张同样内容的白天景色

当然,对于俺这种最喜欢重口味的,当然拉丝是必不可少的最爱

总之,最爱夜晚光亮的丝丝就是了….

另外,住在了著名红灯区骆克道,表示压力很大

 

剩下一些旅游图

迷路了?看路牌?我们在哪?最后还是minmin的强大离线地图拯救了组织

 

去旺角路上,随手拍

 

第一次到湾仔道

香港是个新旧交融的城市,世界顶级奢侈品和杂货摊铺可以在一里之内和谐共存

甚至还有幽暗闲逸的宝云道

 

最后是一点扫街照,希望没有侵犯肖像权

 

Lytro is out, an amazing product

The long waited, revolutionary Lytro camera is finally out. Personally, I think this product is excellent and even beyond my imagination,  it is kind of magical.

The reason I regard Lytro is a successful product, comes from threefold: the cool application, the simple design and the right price.

The cool application

Actually, Lytro is not the only one focus-free camera that developed, go through each year’s SIGGRAPH, you can always find some similar paper related that results the similar effects. For example, coded aperture can do the similar trick, MIT also have some all-round-in computational  photography cameras. However, Lytro, the light filed camera stands out for a reason.

Although there are many classification methods for computational-photography cameras, there is always one secret while classical classification: whether they fully utilize the light come into the camera. Anything related to a specially designed aperture, or natural dense filter or similar filter involved will block some lights go through the lens. That results much longer exposure time, or lower quality with noise as trade-offs. Light efficiency is very important in the industry, that is why several years later, we still find a HDR camera paper published in SIGGRAPH 2010 [ref]. And by using micro-lens rather than light blockers, Ng’s light field camera falls into the glory “light efficient” category.

More over, the demo application is so cool. Even several  years after his paper published, out-of-focus is still a problematic issue for some cases, especially when you want a shallow depth of field. The online demo is very intuitive and directly express the advantage, or at least the fun of this new technology.

The simple design

The key job for Lytro camera is to make photography simple, with out additional focus adjustment. There even no adjustable aperture, since all those things can be done after the photo has been taken. Like the fun application, taking picture of Lytro is very easy and fun, just snap it, that is it.

The more important thing behind the simple design is that, they find the right market target. This is a commercial product, just light another point-and-shot camera. Shooting with it is only a fun hobby, not for serious photographers. Why this market? They may say that: causal photographers tends to free themselves from focusing and they like cool stuffs, like the re-focus demo. That is partially right. Pro-photographers can handle the focusing quite efficiently and they often don’t need the re-focusing.

But, when you think in deep, if the refocuse feature comes for free, why not using it. Actually, light field camera have very limited spatial resolutions for the final generated image. Although it has a 1.1Megapixel sensor, because the light field has much more data need to be captured, the actual image generated should be much smaller, say 1500×1500 for re-focused results (this is simply based on my guess, things my vary). It is far away form enough for pro-photographer compare to monsters such as Canon 1DX (even my poor 60D got a 5184×3456 resolution) However, for consumer, the resolution is not that a big deal, the application is cool, the design is sexy, it done right.

The right price

The right market leads the right price, no one wants to pay over a thousand bucks on a cool high-tech toy, the 399$/499$ price range just fits everyone line. For just that price you can buy a “next-generation” camera, with cool features and the design looks really good, it really a good deal. For gadget fans, show-off guys this is right for you.  This price really make computational photography into everybody’s life.

I didn’t mention too much down side of it, because it is cool, for cool stuffs people can always forgive minor drawbacks such as lacking windows supports….(windows guys are not cool, lets meet Apple fans only 🙂 )

My deep respect to Ng Ren

I really admire and respect Ng, not only for his great research in light field camera and PRT as well. He really has a big vision for computational photography and he managed to made it. I believe the Lytro camera can really spread the power of latest graphics technologies and ignite a new revolution. Good luck for Ng and his all-star team.

One more thing

Let’s wait the video Lytro rock our life. I’m not sure whether they will release a video version of this (I heard from a interview of Ng that they are planning on this). Never the less the video re-focusing will definitely shock the industry. Focusing for video is much much harder than for stilled photo, not even talk about the smooth focus transitions. It really frustrating and experience-intensive job for handling video focusing. On the other hands, video often come 1080p, that is enough, and I think this is not impossible for Nytro to achieve that.

So when the Lytro Video comes out, we can expect another huge succcess for them.

 

 

 

 

追忆Steve Jobs

坚持自己的原则,有自己的特点不一定都是好事,Jobs的伟大之处并不在于他始终固执地坚持自己的原则,而是他的绝大多数原则都是正确的。

同样的,有smart idea的人其实很多。图形化UI不是苹果第一个做的,鼠标也不是苹果发明的,早在iPhone之前,微软早就做过多少代失败的Windows Mobile,早在iPad, MacBook Air之前,就有失败的Tablet PC。在NeXT之前,BSD早早的就在那里无人问津,在Pixar之前,也一样有computer graphics和realistic rendering。很多技术从研究开发到实际的产生巨大市场,都是需要一个临界点,智能手机上如此,Tablet是如此,Computer Graphics以及Computer Vision一样是如此。Jobs的伟大之处在于,他总能准确地把握这个临界点,创造出这个临界点上的产品。

当然,最狭隘的说,Jobs创造了自己独特的价值观,让全世界的Apple users接受的一种价值观。通过简单的UI,真正实现了让每一个人都能使用,并最好的享用计算机(如果广义上这些设备还算计算机的话)。

后Jobs时代,Apple的转折点

我们刚刚迎来了所谓的Post-PC时代,苹果又被迫进入了后Jobs时代。失去了Jobs这个精神领袖之后,苹果将何去何从?现在是Apple的一个转折点,iCloud还在襁褓之中,苹果的云战略显得如此的暧昧不清。虽然投资巨大,但比起先前的巨头,苹果依然是市场的后入者。尤其是面对Amazon来自内容/技术/基础架构的全面压力,苹果在云端需要有一场硬仗(啊?你说还有一家M开头的公司?好像是有云端餐厅来着)。后Jobs时代,Apple如何赢得云端,我们只能拭目以待。

最后是贴图,发帖要贴图,能吸引更多的看客,就算图和内容完全没关系也无所谓

张家界秋游图

第一次跟公司offsite,去潘多拉,呃不,是张家界,确实风景不错。随手贴几张秋游图做纪念…..

简直就是各种西游记的外景地。另外,十里画廊的那个小火车坐的真不值,还不如走着,能驻足观景。

另外张家界雾气较大,又是随手乱拍,只好借助各种不靠谱的ps了…..但求不真实…..

哈里路亚山好多人都拍过了,确实没啥爱,就不放照片了

然后是一些天门山的照片,一样雾气重重

峰顶云雾

吓人的盘山道

强大的缆车上缆机构

 

回程的A320

空旷的机场只有一家孤单的单通道飞机

经典的CFM-56,曾经的头像,A320-NEO要有别的引擎了,CFM要hold住啊….

飞机上拍到的奇怪的云

回程的最后,还看到了不错的北京夜景,很赞

诺贝尔奖与IT-2009年诺贝尔奖颁给光纤和CCD成像

2009年的诺贝尔物理奖,并没有颁发给深邃的传统理论物理,而是颁发给了两项真正改变我们生活的应用物理技术——光纤通信和CCD成像技术。

光纤通信使得信号无衰减的穿越大洋,没有光线通信就没有今天的Internet。同样的CCD技术也使得数字摄影成为了现实,使得人们可以快速的获取大量的图像数据,可以说也是现代计算机视觉和计算机图形学的实际基础。

然而,这样的技术获得诺贝尔物理奖,确实有点出乎我的意料倒是。

6 October 2009

The Royal Swedish Academy of Sciences has decided to award the Nobel Prize in Physics for 2009 with one half to

nobel-prize-for-physics-2009-awarded-$7036322$300

Charles K. Kao
Standard Telecommunication Laboratories, Harlow, UK, and Chinese University of Hong Kong

“for groundbreaking achievements concerning the transmission of light in fibers for optical communication”

and the other half jointly to

boylesmith-big

Willard S. Boyle and George E. Smith
Bell Laboratories, Murray Hill, NJ, USA

“for the invention of an imaging semiconductor circuit – the CCD sensor”

Canon unveils EOS 7D

7DIcon

As anticipated Canon has today announced the EOS 7D digital SLR. It boasts a new 18MP CMOS sensor, Dual DIGIC 4 processors, ISO range expandable to 12800, continuous shooting at 8 fps and full HD video recording. It also incorporates a new viewfinder with 1.0x magnification and 100% coverage and a 3 inch LCD with 920k dot resolution.

广大佳能用户期待已久的EOS 7D终于浮出水面。坐拥一千八百万像素CMOS,双DIGIC 4数字核心,最高扩展到12800ISO。EOS 7D的核心功能确实满足了我们的期望。凭借双DIGIC4高速核心,实现8 fps连续拍摄,以及30fps的1080p HDVido使得APS-C级别的DSLR终于拥有了全副级别的高端待遇。凭借全新设计的反光晶体,7D拥有了佳能EOS系列第一款100%可视范围的取景器,相信7D的取景器会让所有摄影师耳目一新。最后值得一提的是,7D内置了Integrated Speedlite Transmitter,也就是内置佳能Speedlite闪光灯控制器,还记得之前即使机顶不闪光也需要挂一台Server Mode的580EX的窘态么,7D将结束这种问题。

7D

总之EOS 7D绝对是弥补了佳能一直以来全副和APS-C产品线之间的一个空白,将APS-C DSLR真正带入一个高端领域。尤其适合我这种中低端用户,EOS 7D绝对是值得期待和拥有的一台相机。