It's also a great place to admire what I've dubbed the Great Spots of Foreshot—aka, the spots of water-looking water amidst a sea of Papa Smurf's aging, leathery skin.

這裡也是欣賞我所稱的 "Foreshot 大斑點 "的絕佳地點--也就是在藍爸爸衰老的皮包骨的海洋中，那些看起來像水的斑點。

I was actually halfway to opening up a Great Spots gift shop when someone pointed out that the spots aren't real—they're just on Google Maps because of how the fine folk at

實際上，我正準備開一家 "偉大景點 "禮品店時，有人指出，這些景點並不真實--它們只是谷歌地圖上的景點，因為 "偉大景點 "公司的好心人在地圖上標出了這些景點。

Google choose to represent water.

谷歌選擇代表水。

Because yes, all the water in the Gulf of Mexico, IRL, does look like water—well, when it doesn't look like oil.

因為是的，墨西哥灣的所有水，在國際上，看起來確實像水，但看起來不像油。

So, why does an app that shows us satellite images of land so detailed you can count these horses fake its water?

那麼，為什麼一款向我們展示衛星影像的應用程序會偽造水源？

What are they hiding beneath that hypnotic blue expanse?

在那片令人迷醉的蔚藍之下，究竟隱藏著什麼？

Let's start by looking at how the Google Maps map is actually made, and, more specifically, how and why you can zoom in on it.

讓我們先來看看谷歌地圖究竟是如何製作的，更具體地說，是如何以及為什麼可以放大地圖的。

Depending who you believe, the Earth is round, and because of that, this fully zoomed-out

取決於你相信誰，地球是圓的，正因為如此，這個完全放大的

Google Map—which is a flat square—is distorted in places.

谷歌地圖是一個扁平的正方形，在某些地方會失真。

On maps like this one, places on the equator are the least distorted, while Greenland looks enormous even though it's smaller than a new country I made called Fiji, Arabia.

在像這樣的地圖上，赤道上的地方失真最少，而格陵蘭島看起來很大，儘管它比我創建的新國家 "斐濟"（阿拉伯）還要小。

Or, put another way, all these circles are about the same size, and if you're thinking, no they're not, well, exactly.

或者，換一種說法，所有這些圓的大小都差不多，如果你在想，不，它們不是，那麼，正是如此。

But if you zoom in on any given point on Google Maps, it gets less and less distorted as you go.

但是，如果你放大谷歌地圖上的任何一個點，它就會隨著你的放大而變得越來越不真實。

This is because the fully zoomed-out Google Map isn't one big picture that you look closer at as you zoom.

這是因為完全放大後的谷歌地圖並不是一幅大圖，你可以隨著放大而仔細觀察。

Zoom isn't a magnifying glass, it's more like a portal to better pictures.

變焦不是一個放大鏡，它更像是一個通向更好照片的入口。

You see, the map is actually a collection of tiles—a bunch of little images mosaiced together to make a big one—and as you zoom in, Google Maps stealthily transitions you from looking at one image, to a closer-up one, to an even closer-up one.

當你放大地圖時，谷歌地圖會自動將你從一張圖片切換到一張近景圖片，再切換到一張更近景的圖片。

Google Maps has about 21 zoom levels in total.

谷歌地圖共有約 21 個縮放級別。

In level 0, you're looking at one tile, 256 pixels by 256 pixels.

在第 0 層，您看到的是一塊 256 像素乘 256 像素的瓷磚。

Every time you zoom in a level, a 256 by 256 pixel tile represents a quarter of the area of the previous tile, so the picture is four times as detailed.

每放大一級，一個 256 x 256 像素的圖塊就代表前一個圖塊面積的四分之一，是以圖片的細節是前一個圖塊的四倍。

So, while the whole world fits on one tile at zoom level 0, at zoom level 21, the world is 2 million tiles wide, and contains over 4 trillion tiles total.

是以，在縮放級別為 0 時，整個世界可以容納在一塊瓷磚上，而在縮放級別為 21 時，世界有 200 萬塊瓷磚寬，總共包含超過 4 萬億塊瓷磚。

At zoom 1, Everglades National Park fits on a pixel.

變焦 1 時，大沼澤地國家公園就在一個像素點上。

At zoom 20, a pixel barely covers two iPhones.

變焦 20 時，一個像素勉強能覆蓋兩部 iPhone。

It's this tiling—this way of piecing the world together—that makes representing the ocean so dang hard.

正是這種傾斜--這種拼湊世界的方式--讓表現海洋變得如此艱難。

Because it means that each individual image snapped of Earth and uploaded to Big Google's map-making machine has to be stitched to a bunch of other ones, so all these tiles look like part of one big continuous whole.

因為這意味著，每一張從地球上拍攝並上傳到谷歌地圖製作機器上的圖片都必須與其他圖片拼接在一起，這樣所有這些圖片看起來就像是一個連續的大整體的一部分。

And how do they do that?

他們是怎麼做到的？

Well, I sent my outside correspondent Amy to go figure it out, and instead of building a satellite like I told her to, she chatted with these two guys from SkyFive—a marketplace that gives people access to real-time satellite images from a bunch of the little buggers whizzing about the sky.

我派我的外線記者艾米（Amy）去搞清楚這個問題，結果她沒有按照我的吩咐製造衛星，而是和 SkyFive 的兩個傢伙聊了起來--SkyFive 是一個市場，可以讓人們從一堆在天空中呼嘯而過的小傢伙那裡獲取實時衛星影像。

So, let's look at some of the satellite images SkyFive sent us.

讓我們來看看 SkyFive 發給我們的一些衛星影像。

Right now, we're looking at land, and we know what land it is, for a few reasons.

現在，我們正在尋找土地，我們知道是什麼土地，原因有幾個。

One, we know where the satellite was and which direction it was pointing at the time the photo was taken.

首先，我們知道拍攝照片時衛星的位置和指向。

But there are also distinguishing features like buildings, roads, and other hard lines that make it really easy for algorithms, such as one called SIFT, or another called Erosix, to match this image with others of the same place and use that info to pin this image on a map.

但也有一些明顯的特徵，如建築物、道路和其他硬線條，這使得算法（如一種名為 SIFT 的算法或另一種名為 Erosix 的算法）很容易將這幅影像與同一地點的其他影像進行匹配，並利用這些資訊將這幅影像固定在地圖上。

And it can do that with just a few patches of distinct, high-contrast pixels.

只需幾塊明顯的高對比度像素，就能做到這一點。

So now it knows we're looking at Port Fourchon, which is great, and even as we venture out onto the water, there are still a few distinct features SIFT can use to tell us exactly where we are.

是以，它現在知道我們在福爾崇港，這很好，即使我們在水上冒險，SIFT 仍然可以利用一些明顯的特徵來告訴我們確切的位置。

But eventually, we've gone too far.

但最終，我們還是走得太遠了。

Look at this picture.

看看這張照片。

What in this picture could possibly pin this location down?

這張照片上有什麼能確定這個地點？

What collection of pixels is unique enough to this part of the Gulf of Mexico that you wouldn't just drop it in the Pacific and clock out?

在墨西哥灣的這一部分，有哪些像素集合是獨一無二的，以至於你不會把它扔進太平洋就打卡走人？

A wave?

波浪？

A glint of sunlight?

一抹陽光？

The shadow of a fish?

魚的影子？

Of course not, all those things move.

當然不是，所有這些東西都會動。

Also, how would you possibly line this up with pictures of adjacent chunks of the Gulf of Mexico?

另外，你怎麼可能把它與墨西哥灣相鄰大塊海域的照片聯繫起來？

Unless they were all taken at the exact same moment, the water would look totally different, and even having distinct features won't save you.

除非它們都是在完全相同的時間拍攝的，否則水看起來會完全不同，即使有明顯的特徵也救不了你。

Like, you could probably geolocate this based on this little island, but the water looks completely different from one day to the next.

比如，你也許可以根據這個小島來確定地理位置，但每天的海水看起來都完全不同。

I asked my editors to stitch together a few pics of water to make a big pic of water, and they did this before resigning en masse.

我要求我的編輯們把幾張水的照片拼接成一張水的大照片，他們在集體辭職前做了這件事。

On land, the same distinct permanent features that you use to pin photos down also get used to stitch photos together.

在陸地上，您用來固定照片的明顯永久特徵同樣也會被用來拼接照片。

Once an image has undergone atmospheric correction, which is to say, had water vapor and other atmospheric conditions edited out, and once it's been orthorectified, which is to say, accounted for the way changes in elevation might distort it, it gets stitched to its neighbors.

一旦影像經過大氣校正，也就是剪掉水汽和其他大氣條件，並且經過正射校正，也就是考慮到海拔高度的變化可能會使影像失真，它就會被拼接到相鄰的影像上。

There are a few ways to mosaic two satellite images together.

有幾種方法可以將兩幅衛星影像鑲嵌在一起。

One is lining them up as best as you can and saying, that's that, but it leaves some harsh transitions.

一種方法是儘量把它們排好，然後說，就這樣吧，但這會留下一些苛刻的過渡。

Another is feathering, or smudging them together and hoping for the best.

另一種方法是羽化，或者把它們塗抹在一起，希望能達到最佳效果。

But the best, most common method these days is to cut photos together using something called seamlines.

不過，現在最好、最常用的方法是使用一種叫做 "接縫線 "的東西將照片剪接在一起。

This way, you don't line pictures up together along your image's borders—instead, you cut your images into funky shapes based on the starkest lines in them, and line them up that way to make the transitions less harsh.

這樣，你就不會沿著圖片的邊界將圖片排在一起，而是根據圖片中最鮮明的線條將圖片剪裁成時髦的形狀，然後將它們排在一起，使過渡不那麼刺眼。

Because it's weird if two halves of a building mismatch, but not so much if separate buildings on opposite sides of a road do.

因為如果一棟建築的兩半不匹配會很奇怪，但如果道路兩側的獨立建築不匹配就不那麼奇怪了。

So, back to the ocean.

那麼，回到大海吧。

What would your seamline look like here?

你的接縫線在這裡是什麼樣的？

Again, I asked my editors to do it, but they just made me do this.

我再次請求我的編輯這樣做，但他們只是讓我這樣做。

But the geolocating and stitching aren't even the only problems here.

但地理定位和縫合並不是唯一的問題。

Water is just straight up hard to photograph from above.

水是很難從上面拍攝的。

For starters, it's what's called a specular reflector, meaning that, sort of like a mirror, almost all the light that hits it bounces off in a dome shape going every which way.

首先，它是一種所謂的鏡面反射器，也就是說，就像一面鏡子一樣，幾乎所有照射到它的光線都會以穹頂狀向各個方向反彈。

The result?

結果呢？

Direct sunlight creates a weird ocean smudge in satellite images that makes the water look brighter than the land.

陽光直射會在衛星影像中產生奇怪的海洋汙點，使海水看起來比陸地更亮。

Do you want that in your Google Maps?

你想在谷歌地圖上看到這些嗎？

Of course not.

當然不是。

Then of course there's the fact that basically nobody with a commercial satellite wants to take pictures of the open ocean because basically nobody wants to buy those pictures, and they're not exactly free for satellite companies to take or transmit down to Earth.

當然，還有一個事實是，基本上沒有商業衛星願意拍攝公海的照片，因為基本上沒有人願意購買這些照片，而且衛星公司拍攝或向地球傳輸這些照片也不是完全免費的。

I mean, we asked Skyfy for this pic of this patch of the Atlantic, but that's just because we're a bunch of sickos, and because I think I lost my AirPod there on one of my yachting trips with Brian Reel Engineering.

我的意思是，我們向 Skyfy 要了這張大西洋的照片，但那只是因為我們是一群變態，而且我想我和布萊恩-里爾工程公司（Brian Reel Engineering）的一次遊艇旅行時，我的 AirPod 就丟在那裡了。

This is one month's imaging plan for one commercial satellite.

這是一顆商業衛星一個月的成像計劃。

Notice anything?

注意到什麼了嗎？

And it makes sense.

這也是有道理的。

Google Maps is showing you whether or not you're near a cheesecake factory, or how to get from apples to orange.

谷歌地圖會告訴你附近是否有芝士蛋糕工廠，或者如何從蘋果到達橙子。

Why would Google waste valuable time and resources creating a zoomable satellite image of the whole ocean when you'd never use it, they can't get businesses to pay to show up in it, and half you losers keep maps in MapView instead of satellite anyway?

谷歌為什麼要浪費寶貴的時間和資源來創建一個可縮放的海洋衛星影像，因為你永遠不會使用它，他們無法讓企業付費在其中顯示，而且你們這些失敗者中的一半都將地圖保留在地圖視圖中，而不是衛星影像中？

Have you no interest in seeing the bean from above?

你就沒興趣從上面看看豆子嗎？

In all, the Google Maps map is made up of billions of images from satellites, airplanes, and street-view cars, backpacks, and so on, that combine to give you two of the major accurate images of where stuff is.

總之，谷歌地圖是由數十億幅來自衛星、飛機、街景汽車、揹包等的影像組成的，這些影像組合在一起，為你提供了兩幅最準確的影像，讓你知道東西在哪裡。

In the ocean, you don't need that, because if you're here, sorry buddy, you're already dead.

在海里，你不需要這些，因為如果你在這裡，對不起，夥計，你已經死了。

So, what does Google do instead?

那麼，谷歌是怎麼做的呢？

Well, they make the Expanse—the big blue mass we all know and love, which uses data from sources like the General Bathymetric Chart of the Oceans, or Jebco, where the Jebs go, to kind of sort of estimate the topography of the ocean floor, you know, for texture—not that we actually know the topography of the ocean floor, it's mostly guesses and vibes based on the bumps and valleys of the surface, but you know, who cares.

他們製作了 "Expanse"--我們都熟悉和喜愛的藍色大塊頭，它使用 "大洋深度圖"（General Bathymetric Chart of the Oceans）或 "傑布科"（Jebco，"傑布 "的去處）等來源的數據來估算海底的地形，你知道，是為了紋理--我們實際上不知道海底的地形，主要是根據海面的凹凸來猜測和判斷，但你知道，誰在乎呢。

It looks good enough in the app, and again, almost nobody looks at it.

它在應用程序中看起來已經足夠好了，但還是那句話，幾乎沒人會看它。

Except us.

除了我們。

In honor of the good and too-often-unsung work of the Google employees behind the blue

為了紀念谷歌員工在藍屏背後默默無聞的出色工作

Expanse, let's admire it for a moment over some smooth jazz.

Expanse, let's admire it for a moment over some smooth jazz.

Ah, lovely.

啊，真可愛。

You know what else is lovely?

你知道還有什麼更可愛嗎？

This video's sponsor, Brilliant.org.

本視頻的贊助商是 Brilliant.org。

It's an online learning platform with thousands of lessons, and it is the best, most fun way to learn math-y, science-y topics, even if they don't come naturally to you.

這是一個在線學習平臺，有成千上萬的課程，是學習數學、科學主題的最好、最有趣的方式，即使這些主題對你來說並不自然。

Remember earlier, when I was talking about algorithms?

還記得我之前談到的算法嗎？

Yeah, I basically only know what those are because of Brilliant.org.

是啊，我基本上是通過 Brilliant.org 才知道這些是什麼的。

Algorithm was kind of just a word to me, until they came along.

在他們出現之前，算法對我來說只是一個詞。

Their lessons are interactive, and make even the most technical concepts easy to grasp—and even better, they're quick.

他們的課程是互動式的，即使是最專業的概念也很容易掌握，更棒的是，他們的課程很快。

You can learn how sorting algorithms work, and then how recommendation engines work, and then how Spotify's recommendation engine turned something like That Me Espresso into a hit—all in 15 minutes at a time.

你可以瞭解排序算法的工作原理，然後瞭解推薦引擎的工作原理，再瞭解 Spotify 的推薦引擎是如何將《That Me Espresso》這樣的歌曲變成熱門歌曲的--所有這些都只需要 15 分鐘。

That's right, in the time it takes for your friend who said he was quote on his way to actually get to wherever you're supposed to meet him, you can learn how GPS works, and maybe even use it to track him down.

沒錯，在你的朋友說他正在路上的這段時間裡，你可以學習 GPS 的工作原理，也許還能用它追蹤到他。

So whether you're in school and trying not to forget too much over the summer, or a real adult just trying to know more stuff, I highly recommend you give Brilliant.org a try.

是以，無論你是在校學生，想在暑假期間不忘記太多東西，還是真正的成年人，只是想了解更多的東西，我都強烈建議你試試 Brilliant.org。

And it's never been easier because you can try everything they have to offer, free for a full 30 days if you visit brilliant.org or click the link in the description.

而且，這從未如此簡單，因為只要您訪問 brilliant.org 或點擊說明中的鏈接，就可以免費試用他們提供的一切服務，為期整整 30 天。

You'll also get 20% off an annual Premium subscription, so what are you waiting for?

您還可以享受年度高級訂閱 8 折優惠，還等什麼？