Just a quick note about something I figured might be relevant to your interests:
After we released the Aviary book last year, Martin’s been getting a lot of inquiries about his In The Rocks slingshot. So he decided to hand-make a small run of them – along with a custom-designed glass that perfectly fits the slingshot – and make them available for sale. Included with the slingshot is a modified recipe for making the ice spheres at home, circumventing the need for a super-expensive superchiller.
In Martin’s words, “we’re making a small number of Slingshots for the home user. We produced our own custom glass and are offering the slingshot in two gift sets, one with a single glass and another with two glasses. Beware that this process is intended for the ambitious cocktail enthusiast. It will require some trial and error to refine the sphere-making process based on one’s specific freezer. But we can promise once you’ve dialed in the process, the resulting crack will be ever so gratifying.”
Despite waking up with what my kid calls “a cold in my beak” (thanks, Daniel Tiger), I nevertheless hopped out of bed with a bit of a spring in my step this morning. Today is a little bit exciting for us here in Chicago.
This time last year, we were just working on sharing our years-long work on The Aviary Cocktail Book with you and the rest of the world. The response to the book has been amazingly overwhelming. In the time since we last checked in with you, around 70,000 other folks have joined you in ownership of the book, and nearly all of the feedback we’ve gotten about it has been positive.
This reception has allowed us the ability to build new projects, and build we have. This year, we have been working on – not one – but three new publishing projects for The Alinea Group. We’ve been producing these projects simultaneously, sort of like how Peter Jackson filmed the Lord of the Rings trilogy (our production involved considerably less armor), and we’re sharing the first of these today.
Sarah and I are collectors of art we find inspiring, and over the years have amassed a sizable amount of what might best be called “slow-read magazines”. You’ve probably seen projects like this in your local hipster coffee shop or other places that deal in higher-quality magazines. They sort of sit somewhere in between a magazine and a softcover book: the paper and general print quality is high, and they feel notably less-disposable than a typical circular.
We’ve often thought it would be fun to make something in a format like this, but the notion has just sat on the back burner for several years.
When we were building The Aviary book, a subject that popped up in our conversations over and over again was about how we approached writing recipes. We went back and forth several times about whether we should be refactoring them all to be easier to make at home, or keeping them faithful to the way we make them here. We ultimately chose the latter direction obviously, but it wasn’t a decision we arrived at lightly. And it’s a decision that continued to itch us well after we’d sent our book off to the printer.
Why is this? Both myself and the entire culinary team here have a deep-seated interest in sharing knowledge, and despite wanting to proudly share the experiments and discoveries we’ve made over the years in our kitchens at The Aviary, we also dislike the notion that our use of esoteric equipment and ingredients makes the experience of recreating some of these recipes in accessible to some.
So it was only a matter of time before these two ideas slammed into one another. We decided to try making an inexpensive booklet filled with recipes that have been designed from the ground up to be easy and fun to make – not in a high-end restaurant – but at home.
For our first of these projects, we chose to center our work around one of our favorite times of year: the holiday season. The period of the year we’re currently entering always seems to be highly-focused around parties, family get-togethers, and other opportunities to share food or drinks with loved ones. We built around 30 recipes that all make use of seasonal, easy-to-find ingredients. Some of these are single-portion cocktails, but a good deal of them are large-format “punches”, meant to be ladled generously from large bowls into the glasses of each of your guests.
This project included one critical difference from the way we built The Aviary Book: Sarah and I stepped into the position of being test chefs. Because we needed to ensure everything could be made at home – without any specialized kitchen equipment – we worked for months carefully making and testing everything. We would periodically pack all our stuff into coolers and bring them into work to let the chefs taste and verify we’d gotten things right.
Our first shipment of these booklets has arrived in the U.S., and is being transported to our fulfillment warehouses as we speak. They’ll soon be unpacked and inventoried, and will begin shipping mid-October if everything goes to plan. If you’d like to reserve a copy for yourself today, we’ve opened up preorders on our site.
This update is presented in two parts; let’s get to them.
Part 1: The Book
After the pages of our book were printed last month, they were folded and sewn together using the binding techniques we’ve mentioned in past updates. The printer is currently in the process of fastening these sewn “book blocks” into their hardcovers, which ultimately yields complete, finished books.
After the book assembly itself, a bit more work needs to be done to create the custom case we’ve designed to house the Reserve Edition. These cases are individually handmade, and take a fair bit of extra time to produce.
Typically, a printer would bind all of the books and make all of the cases, then pack everything up, load everything into large shipping containers, and send the entire order to the various fulfillment warehouses to which they are destined. This would be the tidiest and simplest way to manage our shipment.
But our printer – knowing that we have an audience of supporters who have been following our progress with this project for quite a long time now – has offered an alternative shipping idea. They suggested that as they complete each book, they could begin filling two shipping containers – a 40-foot one and a 20-foot one – with enough copies of each edition to cover our early-supporter orders. These containers would then be sent to us as soon as they are filled, while the printer continues working on the rest of our order.
This obviously seemed like a pretty great idea to us, and so we’ve pressed forward with it. As of this writing, our two “early” containers have been filled.
The larger container holds our Standard Editions. This container will soon be loaded onto a cargo ship named Ever Eagle. We can collectively follow the ship’s progress as it makes its way from China over to LA, where our books will pass through customs into the U.S.
In the smaller of the two containers are our Reserve Editions, which need to be individually signed by Chef Achatz, Nick, Micah, Sarah, and myself. For this, the five of us need to physically exist in the same space as the books in order to sign them. This leaves us with an interesting choice: we could pay the cost of bringing the container directly here to Chicago, or we could collectively fly to the warehouse that will house the books to do this signing work.
It’s cheaper for us to go to the books, but we were advised that our fulfillment warehouse wasn’t set up for a bunch of people to spread out and sign books for several days. Instead, we need to bring the shipment to us. Because the route of this container differs from that of the larger, it resides on a different ship, the OOCL GENOA, which – as of this writing – left Hong Kong a few days ago and is on its way across the Atlantic.
We currently anticipate the arrival of both shipments sometime in mid- to late-October. Once they arrive, the containers will be unloaded, the books unpacked, and each will be individually-labeled and sent on its way to you.
Part 2: The Game Theories
While I recognize that the topic of marketing leads us pretty far afield from cocktails or bookmaking, I’ve found the process (which we’ve detailed in our recent updates) so interesting that I feel compelled to share a bit more about it. When we last spoke, I’d mentioned that Sarah and I had begun work on a handful of ads to try running on social media platforms. For our first experimental run at this, we created a handful of ads designed to test a few different ideas.
Here’s one of the first I made. The image itself is a closeup photograph of some raspberry ice marbles, which I felt were pretty and unusual:
And here’s one of Sarah’s first, featuring some illustrations she made of some of The Aviary’s cocktails:
Nick often suggests the use of some sort of motion, which he suspects can strongly add to the appeal of an ad. Sarah, taking note of some of the ads being shown to her in her own feed on Instagram (Allbirds being one example that strongly appeals to her sensibilities), decided to test Nick’s theory by adding some subtle animation to her illustrations:
Here’s another example Sarah built, again with a small amount of animation:
I wanted to try creating something with some motion as well, so I leaned on my own visual effects experience to make this:
Sarah suggested we might try one example featuring a very classic cocktail, so she made this one using an image of The Office’s Old Fashioned:
We also made one other ad, but explaining it requires a little bit of backstory.
While walking around one morning, Sarah and I happened upon a paper store. “Let’s check this place out,” Sarah interrupted our conversation to pull me into the shop. Once inside, as we were strolling up and down the aisles, Sarah stopped me and asked “would it have occurred to you to stop in this place on your own?”
“Hm. Not really, unless you’d told me to come here specifically to pick something up for you.”
“Why do you think that is?” she asked.
“I don’t really know. Why?” I shrugged, unsure where she was going with this.
“Well, what’s this place sell?”
I looked around. The shelves were filled with dozens upon dozens of pretty paper products: large, handmade sheets embedded with flowers, or stationery kits emblazoned with pastel filigree. The whole place was colorful and swirly. “It seems like the perfect place to gift shop for you” I smiled.
“Right. But you probably wouldn’t have come in here to buy something nice for yourself, even though I could point out lots of things in here that I know you’d like. And look around in here,” she gestured at the dozens of customers in the store. “It’s mostly women. Something about the vibe of this place suggested to you when you walked in here that it’s meant for women.”
“I wonder if most of the ads we’ve made so far skew masculine,” Sarah continued. “There are a lot of delicious recipes in this book that my sister would love, or that your sisters would enjoy. But they have sort of a different design sensibility than you, and the ads we’ve made probably wouldn’t be interesting to them.”
I found this observation forehead-slappingly insightful, and immediately begin wondering how I might craft an ad that might skew more feminine. I began thinking of my sisters, both of whom devote a lot of attention towards curating their homes to feel comfortable and beautiful, and so I wondered about creating an ad that leaned less-heavily on the flashy cocktail theatrics and more towards an imagined lifestyle for the owner of our book. Where might my sisters, for example, keep this book in their homes?:
Sarah wanted more time to develop her own ideas for this, but with the above ads we had enough of a variety that we felt comfortable beginning some tests.
But before we get to the spoilers, I’ll invite the reader to scroll back through these again. Which ads are your favorites (if any)? Which would cause you to pause and read more carefully about what you’re seeing?
So, the nitty gritty. We chose to run these ads on both Facebook and Instagram, using Facebook’s advertising tools. Facebook allows us to suggest some basic descriptions of our desired audience. Because we wanted to completely avoid bias, we instructed Facebook to show these ads to people who had expressed an interest in cocktails in some way…but we also wanted to specifically exclude anyone who had “liked” any of the social media pages for The Aviary or any of the other Alinea Group restaurants. We wanted, effectively, to be speaking to complete strangers, without any inherent knowledge of us.
After specifying this audience, we allotted an equal budget for each ad, and ran them all for one week. By the end of this period, we had some interesting data to sift through. There are several metrics by which we can measure the relative success of each ad, but for simplicity we’ll consider “engagement”, which is jargon for some expression of interest on behalf of the viewer (so, in the case of Facebook, “liking” the ad, commenting on it, sharing it, etc.).
What we found surprised us. We felt that our strongest ad would be the fireball animation, followed possibly by the “lifestyle” bar scene and maybe the animated illustration. But all three of these ads were dramatically eclipsed by the “Playfully Delicious” one above. In fact, “Playfully Delicious” outperformed ALL of our ads by a large margin.
The second-strongest performer was the Old Fashioned, which was another surprise for us. Our choice to create this ad was almost as an afterthought, and it was intended to represent sort of our “control” case, because it differs so very little from most other cocktail branding we’ve seen elsewhere. We worried it would be too predictable for The Aviary.
And while the animated illustration did well, we were surprised to find that it performed almost 10 times better than its non-illustrated counterpart.
This left – as our poorest-performing ads – the ice marbles, the fireball, and (in dead last place) the “lifestyle” bar scene ad.
Facebook of course offers no suggestions or indications why our ads performed as they did. But trying to understand this is quite a bit of fun. While I suspected, for example, that the flashy theatrics of the fireball ad would make it a clear winner, we can now theorize that it performed poorly because – if someone has zero context of The Aviary or that this is a book about cocktails – it’s not entirely obvious what’s going on.
Conversely, we can posit that the “Playfully Delicious” ad performed well because it combined Nick’s theory about including a little bit of fun-to-watch motion with something that’s understandably edible. Similarly, the Old Fashioned likely performed well because of its recognizability. While Sarah and I feel that The Aviary itself differs quite a bit from other cocktail bars (and our book differs drastically from most cocktail books we’ve seen), highlighting these differences might be less effective to viewers who may be wholly unaware of The Aviary than establishing some fundamental similarities first.
Sarah and I have been surprised to see similarities in this creative process and the creative process of the Aviary (and even Alinea itself): regardless of how disorienting or surprising a given dish might be, the chefs are careful to always include some sort of “access point”: a toehold of familiarity that invites you to explore further. Sometimes this can be a vaguely familiar flavor note, other times it might be using a visual presentation that seems familiar. It seems that both on the plate and in these ads, a note of understandability or recognition can help bolster the connection we’re trying to make with people.
The fun part of all of this experimenting is that we can easily test our own theories by making, say, a few more ads in the style of “Playfully Delicious” and see if they consistently perform as well as the original. And I can gut-check my intuition about the fireball ad by swapping out the pages for something that’s more understandable as a cocktail (I could, for example, just try using the Old Fashioned image instead). This entire process is – as I mentioned in our last update – a completely fascinating exercise in empathy that I find to be a lot of fun.
In the time since our last update, our book has been printed. A few days ago, the final moment of approval arrived at our studio in the form of what the industry colloquially refers to as “F&G’s” – short for “folded and gathered signatures”. A sample of each printed sheet in our book is folded and trimmed as exemplified in this short video. The resulting signatures are then gathered together in the order they will appear in the book.
If we examine the rear spines of each signature, we can see a small bar printed along the edges; these allow the binder to confirm at a glance that all signatures are in the correct order. These marks will eventually be covered by the outer spine of the book:
These F&Gs are the first time Sarah and I are seeing our book in its final, “for-realsies” form: printed on the paper stock we’ve chosen, with full-resolution images, under the tight color guidelines we’ve sought to adhere to throughout the process, with all varnish and metallic ink passes applied. The point of this final approval is to ensure that nothing has gone amiss, that the pages are ready to be sewn together and the hardcover (referred to as the “case”) applied.
Ideally, we should find zero errors in our F&Gs, but unfortunately, there are a few things that weren’t quite right. Here are some examples:
The foam on top of this cocktail contains a tricky transition of beige tones that sit right at the edge of the Magenta and Yellow color gamuts. We can see here that the press struggled to reproduce these subtle tones faithfully, producing instead a harsh edge with magenta “fringing” that looks obviously broken and incorrect.
Here’s another example: the shadows in the above image (seen most predominantly on the forefinger and in the frozen ginger snow mixture on the left) are clipping in a harsh way. We’re unsure what’s causing this – the anomaly is not present in any of our proof files, so we’ll need to track down the source of this problem with our printer.
There are some other errors we’ve discovered as well, but the matter of fixing these is more complex than one might assume, and here’s why. The printing press we use can generate 18,000 printed sheets per hour. At 30,000 books for our order, that’s less than 2 hours to print each set of sheets. The big cost is the setup and tuning of the press, not the actual time/paper/ink spent printing. Because of this, the printing plant actually prints all of our books before sending us these F&Gs to approve. I would have originally supposed that they would print one sheet of each set of pages, have us check them, and then print the full remainder. But this is horribly inefficient for the press; it ends up being more efficient for them to recycle and reprint sheets with errors. (!)
Because of this, Sarah and I are thoughtful about things we might like to fix. We need to be mindful of our paper supply, for example: fixing a single error requires another 30,000 sheets of paper to be sourced and printed. The two images above do not reside in the same signature, so that’s 60,000 sheets of paper that will be consumed to fix both of them. There is also the setup and re-print time involved – each fix we might like to impose will push out our delivery date further.
Balancing our desire for perfect quality, our abhorrence for waste, and our interest in getting books on their way to you makes this particular phase of error-spotting a rather emotional process for Sarah and me, and leads us into the slightly grey area of needing to prioritize what we feel is worth redoing vs. what we will correct for subsequent print runs in the future.
In any case, once this final stage of corrections has been completed, the books will be bound, loaded onto a large cargo ship, and sent on their way to us. We discussed the matter of receiving and storing these books in our last update, an underlying consequence of which is that we will soon begin incurring costs to house them all.
The timing of this is both fortuitous and difficult, for one major (and perhaps obvious) reason: the Christmas holidays. On the positive side, the climate will hopefully be favorable for us to sell our books expeditiously. On the other hand, the opportunity provided by the selling season is not lost on warehousing services. Amazon, for example, capitalizes on their sellers by raising their warehousing costs to three times their normal amount during the holiday months. If we choose to make our book available on Amazon for the Christmas season, it’s strongly in our interest to minimize the time our books languish in their warehousing system.
And so it is here that we intersect with something that I’ve grown to find completely fascinating over the past few months: advertising.
First, let me provide some context: both Sarah and I completely loathe advertising. I have no fewer than three ad-blockers installed in each web browser I use. We jumped ship on cable TV over a decade ago because we hate commercials. I have a complete distaste for the seemingly-inescapable ad-driven online ecosystem bequeathed to us by Facebook, Instagram, Google, and the like – I find the thought of having my online movements constantly monitored an unwelcome and uninvited intrusion on my privacy. I generally find most advertising crass, obnoxious, and tasteless, and am irritated at the insistence on appealing to the lowest common denominator that most advertisers seem intent on doing.
All of this is to say: making ads initially struck me as the most repugnant way I could possibly spend my time. But coming to an understanding of the various costs and schedules involved in dealing with this large order of books made clear to me the need to engage with the marketing aspect of this project. Sarah and I have to learn to dance with this devil; we have to learn to be good at something we initially despised.
Our toehold for coming to terms with this came in the form of advice from Nick. “We’ll just make a game of this,” he suggested cheerily. “Just because you think most advertising out there is terrible doesn’t mean you have to make advertising you think is terrible. What do you want to see?”
Turning adversity into a tailwind is something I’m totally fascinated with, and Nick’s comments fueled me with curiosity. This thing I completely hated suddenly became interesting to me. I started viewing ads with a ton of scrutiny. Why do I hate a certain ad that goes by on Facebook? What do I wish it was? What would I find compelling to see?
My first thought was “I like learning. I would want to see something that presents me with interesting information”. So I started rifling through the photos I’d taken for this book. This was the first ad I came up with:
I didn’t hate it. I like this image, and I liked the thought of offering a little “behind the scenes” information. I tried making another one:
Again, I liked it. It seemed like fun little trivia to offer, and hopefully provoked enough interest to cause someone to want to learn more (maybe?)
I decided to show these to Nick and ask his thoughts.
“So, here’s the thing with these,” he began after I explained what I was trying. “These ads assume that the viewer 1) knows what the Aviary is, 2) knows that we are making a book about cocktails, 3) knows that these images are from our book, 4) is interested in some behind-the-scenes information about how we made the book, but 5) hasn’t yet decided to buy this book.”
“I think it’s fine to try using this ad,” he continued. “We should, in fact, just to see what happens. But I suspect that your audience for these is relatively small. The thing to keep in mind here is that there are way, way more people in the world who have neverheard of the Aviary than people who have. So part of this game is figuring out how to be clear and concise about what this is.”
It has been this last bit that I’ve found the most fascinating and insightful. It’s caused me to be very mindful of how I interact with ads on services like Facebook or Instagram. I notice that if I scroll by something and realize it’s an ad, I usually just flick right by it in less than a second. I suspect this isn’t uncommon, which helps me realize how short of a story we have to tell in that split-second. This stands in stark contrast to the world from which Sarah and I come, in which we usually have a few leisurely hours to tell a story.
Sarah and I went back to the drawing board with all of this in mind. The distaste we had for advertising started giving way to conversations in which we tried to imagine conveying our work in a split second to someone who had no context for any of this. Having no real experience creating work like this, we fell back to the things we knew best as a starting point.
Both of us have a strong fondness for punchy, vivid colors; my taste for this is reflected heavily in the photography of this book. I wondered about photographing the book itself using this same style, with a brightly-colored background. But we don’t actually have a physical copy of the book yet, so I tried making a computer-generated one instead.
In addition to a bit of extra visual interest, this suddenly adds some clarity: this image we’re looking at is one of many in a book. But what kind of book? It’s not clear that this is a cookbook, and it’s definitely not clear it’s about cocktails. As interesting as the image is, it’s working against the story I need to quickly tell about what we’re doing.
So I tried again:
The inclusion of some text here seeks to add more information: in that split-second of scrolling, I’m trying to convey to the viewer that what they’re looking at is a recipe, probably for some sort of beverage, and that the beverage is a bit unusual. Will they want to stop and learn a bit more?
Sarah is running in a different direction, one that plays to her strengths as an illustrator and animator. She originally had the thought that she wanted to try creating an animated illustration that highlighted The Office:
The above was her first version of this animation. We both love it, but after scrutinizing it for a while, we have some notes for ourselves about how this might be made more effective. We find The Office a lush, comfortable, and inviting environment to be in, so trying to tell that story was an obvious thought: the animation above spends much of its time building and then lingering on the Office environment itself. But dwelling on this aspect of things presupposes that someone else knows and likes this environment as well. To someone for whom this is wholly unfamiliar, there’s no establishing information to help give context to what this animation is about. Is it about a bar interior? Is it about a cocktail? We don’t discover until the very end that it’s apparently about a book (and even then, the “book” aspect is sort of lost…the word “book” is presented in the smallest typeface featured in the animation).
After talking about it, Sarah worked on a second version:
Here she’s modified her animation so that it both begins and ends with the most important information, and she conveys this information concisely (“This is about a book. Specifically, it’s about a recipe in a book, which implies it’s a cookbook”). She speeds through the buildout of the Office a bit more quickly, and ends her shot with the cocktail larger in frame. Beginning and ending with the same text and composition allows the animation to loop seamlessly, which makes it more fun to watch and also follows the visual parlance of social media platforms like Instagram or Twitter.
While we can continue honing these ideas and conjecturing about what might make a better ad, the only real way to know for sure is to try. Our mechanism for that is using the very online advertising machinery Sarah and I have tried so hard to eschew. The ecosystem provided by Google, Facebook/Instagram and others let us test our theories with an exactness that is simultaneously unsettling and fascinating. These tools allow an advertiser to understand which ads tend to attract the most interest and whether a given ad eventually translates into a sale. This differs dramatically from something like a billboard, a radio ad, or even a television commercial, for which there’s no real way to tightly tie the cost involved in creating the ad to income generated from it.
The whole experience of learning about this has indeed begun to feel like a game to Sarah and I. We find ourselves transitioning away from asking “What kind of ad would I like?” and towards questions like “What kind of ad would my mom find interesting? Or my uncle? Or that guy who lived up the hall from me in college?” These questions in turn lead us to ask “What bits of our book do we think would be most interesting to that person, and how can we make an ad to highlight those bits?” It’s become a fun exercise in empathy.
By this time next month, if all goes well, books should be en route to us from China. Once we have confirmation of this, we’ll begin the process of collecting shipping info from each of you and coordinating the delivery of your rewards.
Having one’s book appear on the New York Times Bestseller list is not strictly a matter of popular opinion. Rather, inclusion on that list depends entirely on selling a predetermined amount of books within the first week of a book’s release. For the Times, this magic number is 9,000. Inclusion on the Wall Street Journal’s bestseller list is a bit more forgiving: one needs to sell only 3000 books within the first week to qualify.
Achieving these numbers is rarely accidental; a book publishing company has a number of tools at its disposal to ensure being listed on these lists. Presale orders, for example, count towards those first-week sales (so, when you see a book being “pre-sold” on Amazon, the publisher is often front-loading sales for the purposes of getting on a bestseller list). Another tool is including a “buyback” clause in book contracts. The way this works: a publisher offers an author a deal to publish a book, and in this offer is a clause requiring the author to buy a set number of their own books from the publisher. The way this buyback is usually framed is “here’s a stock of books you can sell yourself”, but spoiler alert: the buyback amount is often coincidentally the same number as a bestseller list threshold.
The astute reader will surmise that preorders count as “presales” for our Aviary book, and would therefore allow us to get within shooting distance of some of these bestseller lists. And this should be correct.
But it probably won’t be.
To give context as to why something like a bestseller list is interesting to Sarah and I at this moment, it might help to paint a clearer picture of the scale of our book order. When she and I created The Alinea Project, we printed 750 copies. In the world of offset printing, this is a tiny order. Yet when 750 books showed up at our small apartment in Oakland, it easily filled a spare bedroom (and most of the hallway). Our apartment was a loft, and I had promised to sign and number each book, which meant carrying each one up a flight of steps to our dining table, unpacking it, signing it, re-packing it, and carrying it back down to our car to drive to the post office. The book weighed about 8 lbs. per copy, which translates to nearly 3 tons of books we moved by ourselves. Packing and shipping that amount of books took us about a month of solid work.
In contrast, we’re currently printing 30,000 copies of the Aviary book, which is a number large enough to threaten being meaningless. Of these 30,000 books, about 3,500 copies are spoken for by early supporters such as you. This leaves 26,500 books unspoken for. That’s 26,500 books (weighing over 100 tons, measuring over 4,000 cubic feet in volume) that need to be transported and stored somewhere.
It probably goes without saying that this amount of weight and space far exceeds the bounds of a spare bedroom or an unoccupied storage closet in a restaurant.
There are, of course, warehousing, logistics, and fulfillment companies who can help us manage this. These services come at a cost, however, and comparing one service to another can be a bit complicated. The longer our books hog up space in someone’s warehouse or the more they need to be shuffled around to a more advantageous geographic spot, the more cost we incur.
At the risk of sounding absurd, there is nothing “automatic” about figuring out how to deal with all this. Once we’ve done the work of producing a book, it will not magically just appear on the shelves of bookstores or in libraries. These tasks are usually handled by a book publishing company, which might have an extensive network of infrastructure designed to accomplish them efficiently. But in this case, it will be Sarah, myself, Nick, Chef, and our team who work to figure out how to navigate these waters.
As an example: an obvious matter for us to consider is selling our book through Amazon. If we choose to warehouse and ship books ourselves, the cut Amazon takes for providing us a bit of digital real estate on their website is relatively low. But if we handle the matter of shipping a book to customers, the book won’t be eligible for “Prime” shipping on Amazon. To be eligible for “Prime”, Amazon wants to be in control of shipping, which means our books need to be sent to one of Amazon’s warehouses. This incurs storage fees and an overall higher cut that Amazon takes of the sale price.
To add further complexity: when our books are loaded onto a boat in China, the boat needs to know where these books are going. Amazon has warehouses all over the world. We could choose to allocate the books to these warehouses ourselves (at a lesser cost, but without really knowing where the books are most likely to be sold), or we can allow Amazon to distribute the books as needed throughout its own warehouse network…at a higher cost to us. Learning how to weigh the variables of these equations is something we are rapidly coming up to speed on.
This may help explain why learning about inclusion on a bestseller list is suddenly an interesting matter to us. We’ve learned, however, that bestseller lists like that of the New York Times depend on numerous sources to report numbers during the first week of sales, including brick-and-mortar bookstores and the publishing company itself. And here’s where we fall into a black hole: we’re a restaurant group. Because we hold no formal status as a publishing company or a bookstore, we aren’t allowed to present sales numbers on behalf of our book. For this reason, it is unlikely you will see this book on a newspaper’s bestseller list, regardless of its quality or how well it is received.
Rather than feeling daunted by the scale of these new problems, there seems to be a collective heightening of excitement as we work with the rest of the team here on these issues. Working in unfamiliar waters is something that’s woven tightly into the fabric of this entire group, spearheaded by Chef and Nick’s insistence on innovation. “I get kind of uncomfortable if I know exactly what I’m doing,” Nick is known to say from time to time. When Sarah and I – creatures of habit, planning, and lots of consideration – get hung up with questions that start with “But what about…”, Nick and Chef will often shrug and just say “I dunno. Try it and see what happens. We’ll probably screw some stuff up, and we’ll fix it, and we’ll learn and get better over time.”
Thinking like this helps Sarah and I worry less about getting things exactly right and allows us to focus more on what we want to try first. One of the things we’re currently working on, for example, are marketing campaigns for the book. This is another area in which Sarah and I are amateurs, but we have a sizable runway to come up with ideas for advertising, and it’s pretty interesting trying to empathize with the different reasons this book may appeal to someone. When we present Nick with ideas and ask which he thinks is better, he often says “I don’t know, we should try them all and see what works.” This “no umbrellas in a brainstorm” attitude exemplifies Nick’s rather amazing ability to enable a broad range of creative thinking, and once again we’re reminded of how fun it is to get to watch people do things they’re really, really good at.
In continuing our discussion about the prepress process, let’s consider the following:
We can see here an image that Sarah created as a design element for one of the essays in our book. If we look closely, we can see that the color of this image appears notably different on each item on which it’s displayed, including the printed page seen just to the left of the laptop keyboard.
To understand what’s going on here (and to identify which image is “correct”), we need to push further into our exploration of color. Doing so will involve quite a bit of technicality, so now would be a good time to put on your nerd goggles.
The blue color Sarah intended for the header image above is one that she originally created in the RGB color space. Specifically, the red, green, and blue values Sarah chose are:
Red: 50 Green: 80 Blue: 144
Adding these values together, we arrive at the following color:
Now, math is math, so these numbers should all add up the same every time, for everyone, anywhere…which is to say that this blue should look the same to everyone everywhere, correct?
Unfortunately, it’s not quite so simple as this – things start to get sticky when we need to get image information into or out of the digital realm. Let’s first start with the moment that Sarah chose her color. She chose it looking at a monitor that’s sitting in our office. Sarah’s monitor is the screen of an iMac. Our office has windows. The windows have sheer curtains. The floor in our office is carpeted. The carpet is blue. The wall behind Sarah’s monitor is green, and the wall adjacent it is brick. The brick is yellow.
Why is all of this important?
Light comes into our office through our window, through the curtains, and bounces around our office – off our blue floor, off the green wall, off the yellow bricks. All of this ambient light creates a color cast on Sarah’s monitor. To add further complexity, this color cast can change minute-to-minute as the light coming through the window changes (as a result of changing weather, time of day, etc.)
Sarah’s monitor itself consists of electronic components that are – despite being manufactured to within very tight tolerances – still physical components that can experience fluctuations based on manufacturing differences, environmental conditions, and age. Consequently, there is likely to be some amount of drift or difference between her monitor and the monitor of my computer (which is a Dell display), or the monitor you might be reading this text on right now (maybe that of a mobile device, or a laptop, or a tablet).
We ultimately need to take the RGB values that looked nice to Sarah – in that environment, on that day – and render them on paper. Printing introduces another whole host of imperfect conditions we need to be aware of. Paper color, ink density, how firmly the printing plate is pressed against the paper, the resolution of our stochastic screen, temperature, humidity, variations in ink formulation – there are dozens of imperfect, ever-changing factors that influence how a color is reproduced when printed.
The question of which version of the image above is “correct”, then, is a two-part issue: there is the original, intended color that Sarah chose, and then there is the final, printed result. We need to ultimately ensure that what ends up printed on the page matches what we’ve been looking at on our screens for the past year or so as closely as possible.
Accomplishing this requires a concerted effort on the part of ourselves, iocolor, and our printing partner in China to constantly measure and re-measure our colors as we work. This begins with Sarah and I. When working with color-critical images that we know are ultimately destined for print, she and I use colorimeters attached to our computers:
These little devices accomplish a few things. First, we can use them to calibrate our monitors to a known standard. This ensures that our monitors match each others’, as well as those of iocolor’s and the printing press. Having everyone’s monitors calibrated helps ensure that when we share an image with our printing partners, they see it the same way in their environment as we see it in ours.
The second feature offered by our colorimeters: they can periodically read the ambient light in our room and adjust the monitor slightly to accommodate for varying viewing conditions. This, technically, is actually not terribly desirable – in an ideal scenario, Sarah and I would work in an environment with tightly-controlled and unchanging viewing conditions. But we don’t actually have such a working space here, so these devices help us keep ourselves in step with each other, so that colors and images look the same on our screens minute-to-minute and day-to-day.
As a potentially-interesting side note: the importance of monitor calibration is something that’s been impressed upon Sarah and I during our time working in visual effects. It’s almost always the case that shots in a film that are meant to be viewed sequentially are distributed among many artists within a vfx studio, and so the need for consistent color and lighting reproduction on every artist’s monitor is of paramount importance. The artists in these studios frequently work in light-controlled environments, and their monitors are periodically calibrated to ensure uniformity throughout the studio.
*see hyper-technical footnote below if you really want to geek out harder about this.
Once Sarah and I have completely finalized the content of our book and are happy with the visuals of it, we ship all of our digital files to iocolor. They then print a “proof” of each page of the book. The proofs are generated on a large inkjet printer that tries to emulate the paper color and ink behavior that will be used on the printing press in China as closely as possible.
Just as with our monitors, the printer on which these proofs are generated is also calibrated, and has an onboard device similar to the colorimeters Sarah and I use (this device is properly referred to as a spectrophotometer). Alongside every page generated by the printer is a small calibration target consisting of swatches of known color:
The spectrophotometer scans these color bars as the pages come off the printer, measuring them to ensure color consistency page to page. If any drift in color is measured, the printer is stopped and the cause of the color drift is identified and corrected.
Once the proofs are generated, they are shipped to Sarah and I, where we visually inspect every page to ensure it looks like what we see on our monitors. This proofing process is what we’ve been working on for the past few weeks.
Once we’ve approved the proofs and said “Yes, these images look like what we expect,” they are shipped to the printing press in China. When it comes time for our book to be printed in earnest, the sheets of our book coming off this press will also have color target bars printed on their edges. The press itself has onboard spectrophotometers that measures each sheet for color drift. The press is a closed loop system, which means that it can rebalance ink density as needed to maintain color consistency over the duration of the print run. This ultimately ensures that all 30,000 books we will create match each other within a very small tolerance, and ideally also match what Sarah and I see on our monitors here in Chicago.
Right now, we’re at the point of this process where we’ve approved digital color proofs from iocolor, and these proofs have been shipped to our printer in China. The printer has begun work setting up the press run for our book. The first step in their process is taking our individual pages and imposing them on the large sheets of paper we learned about several updates ago. To verify that this page imposition has happened correctly, the press prints another proof to send us to review. Because the thing that we’re checking here is simply whether the layout of the book remains as we designed it (i.e. that all the pages are in the right order, are oriented properly, are not being cropped oddly, etc.), the proofs for this step are often created at a very low quality. This version of the book is sometimes referred to as a “plotter proof” or an “ozalid proof”, depending on the printing technique used.
This proof represents the final opportunity that Sarah and I have to effect any changes we may want to make (we’re still trying to comb through the book at every turn to find and correct things like typos or other mistakes). We will spend the upcoming days doing this and passing any changes back to the printing press. Once we’ve put our pencils down on this, the press will begin the full print run of the book.
Until next time;
* Oh boy, you really want to go deep on color calibration! Here we go:
The astute reader will note that there are thousands of different devices in the world that are used to display images: monitors, projectors, televisions, VR headsets, mobile devices, etc. Each of these devices is constructed differently, with different mechanisms for producing an image that you can view. As such, the “correct” way to reproduce Sarah’s blue color on her iMac monitor necessarily differs from the way, say, my 13-year-old Sony TV reproduces the same color (which is to say: the actual electronics in the iMac are operating in a different way than are the electronics in my TV). How, then, can all of these devices aligned such that they can display the same blue swatch identically?
The answer lies in the concept of a device profile. A device profile is sort of like a translator that says “This is how this particular device works, and so here are some instructions – specific to this device – for producing a particular color on it”. The international standard for expressing these instructions are ICC Profiles. If you’ve done any amount of work creating digital imagery, you’ve likely run into this slightly-inscrutable term at some point in your tenure. ICC profiles provide a translation for representing a color from a known color space on a particular input or output device.
What is a known color space? The international standard for this is the CIE 1931 Color System (sometimes you’ll see this synonymously referred to as CIE LAB or CIE XYZ). This concept has for years remained opaque to me until very recently, when my good friend Erik (who is an extremely talented professional image inspector in New Zealand) shared this very interesting and understandable article explaining CIE 1931 with me. Paraphrased simply, CIE 1931 can be thought of as a centralized, internationally-recognized master language of the perceivable color spectrum, and ICC Profiles can convert color values to and from this language on the behalf of whatever devices they are describing.
Note that ICC Profiles are the mechanism by which color values are translated not only from CIE, but also to it. While Sarah chose her blue based on viewing her monitor (which has its own ICC Display Profile, to describe how to translate her blue to be properly displayed on her calibrated screen), our book is also filled with photography. The color my camera captures also needs to be converted into this master CIE space, and so my camera similarly has a device profile for this purpose. The same is true for any color acquisition device, such as film or flatbed scanners, webcams, etc.
Learning about this helps me understand why some photographers create profiles specifically matched to each of their cameras (and, in fact, to each lens for each of their cameras), to ensure accuracy in their own color pipeline. Because I’m learning this as I go, I have not taken such measures, and so there will undoubtedly be a (hopefully small) amount of inaccuracy in the color pipeline Sarah and I have built with iocolor as a result. This helps underscore the value of proofing our book repeatedly as we go: just as with the recipes, we have to accept some amount of tolerance for the fact that the images coming off the press may not match our monitors exactly, but we want to ensure they nevertheless look good.
This past month has been – as we anticipated – quite busy and challenging. We’ve begun the prepress process with our printer, iocolor, in earnest, which has involved some finicky details which we intend to cover in depth in this and the next few updates.
To understand what “the prepress process” means, let’s build ourselves some foundational understanding of the printing process as a whole, starting with the idea of a prism.
You’re probably familiar with the underlying principle of a prism, a triangular shape that can split white light into its constituent colors. Light itself exemplifies what’s called an additive color model; colors are added together to yield white. We saw this principle demonstrated a few posts ago when discussing visual effects. The primary colors in an additive color model are red, green, and blue. These three colors can be mixed together in varying proportions to yield many other colors.
Computer monitors embody a real-world example of this additive color model: computer monitors work by projecting light onto a screen. A given pixel on a computer monitor starts off “black” (i.e. “no color”). To make a pixel “white”, a monitor combines red, green, and blue light. Digital photography works on the same principle; the pixels in a digital photograph all consist of varying levels of red, green, and blue to yield a final image.
A printed image, however, works differently. When we print an image, we start with white paper. Producing color on the paper involves depositing pigments onto the paper. Light has to travel through this pigment to hit the paper; when this light reflects back to our eyes, part of its color has been absorbed (subtracted) by the pigment. Printing, then, is an example of a subtractive color model.
A subtractive color model is probably what most of us learned about back in our earliest days of school. We learned that red, yellow and blue were the primary colors, and that if we mixed these in various proportions, we could produce other colors from them with our finger paints.
Our desire when printing, then, is to take an image represented by an additive color model and convert it to something that works in a subtractive color model. How is this done? Let’s press further into color theory:
The complementary color of red is cyan, which means that cyan pigment serves as a filter that absorbs red. The amount of cyan applied to a white sheet of paper controls how much of the red in white light will be reflected back from the paper. Likewise, magenta is the complement of green, and yellow the complement of blue. By setting cyan, magenta, and yellow as our primary subtractive colors, we’re effectively letting light’s primary additive colors of red, green, and blue reflect off a piece of paper back to our eyes. This model works nicely to effectively “invert” the additive color model computer monitors exhibit.
This seems great; we’ve learned that we can combine cyan, magenta, and yellow pigments in varying amounts to produce more or less any color we want on paper. If we were painters, we could start with a palette of these primary colors, and could dab around with our brushes to combine these pigments to create just the right color we wanted before applying it to our canvas. But pre-mixing each and every color we want to appear on a page is time-consuming; if we want to print 30,000 copies of a 468-page book and get it to you as soon as possible, pre-mixing every single color you’d see would be completely unreasonable.
So, printers use a clever technique to make this process more efficient: they separate an image into its cyan, magenta, and yellow primary components. To help explain this more clearly, let’s introduce some more visuals.
Consider this idyllic barn set against a backdrop of the Tetons, an image which I’m sure will make you feel peaceful and relaxed.
For each color seen in the image above, we can ask “How much cyan is in this color? How much yellow? And how much magenta?” We can separate each of these subtractive primary colors like this:
Mixing full amounts of each of these three primaries in an attempt to create black doesn’t quite all the way work; what one ends up with is more sort of a muddy brown color. It’s also highly inefficient: imagine printing a book full of black text using three different inks. It would be expensive, and there would also be all sorts of registration problems getting the three inks aligned exactly right for small text.
To remedy this, printers add black pigment as a fourth “primary” color. Here’s the same barn image, now split into cyan, magenta, yellow, and black components. Note how it makes much more efficient use of the colored inks.
These four colors together comprise the CMYK color space.
So, we have our four primary colors, which mix in varying amounts to yield many other colors. How is this mixing accomplished? Offset printers leverage a technique called “halftone screening” to determine how much of a given color pigment is applied to paper. Halftoning is an optical illusion that seeks to represent continuous tones using dots. Viewed from a sufficient distance, these dots appear to meld together to form continuous gradients of color. The “amount of color” in a halftone is controlled (in most cases) by varying the size of the halftone dot: larger dots = more color, smaller dots = less color.
The relative density of these dot patterns are measured as “dots per inch”, or “DPI”. Generally speaking, the more dots per inch, the finer the achievable resolution of an image.
Here’s a closeup of an image of my niece, printed using Blurb’s print-on-demand service.
Note the ‘spotty’ quality of the image; this isn’t an artifact of my camera. It’s the halftone screening process used, and is a combination of several layers of halftones. Each layer is made using one of our primary subtractive colors: cyan, magenta, yellow, or black. I can separate out each of these layers like this, and we can clearly see the halftone patterns:
You might notice that the ‘grids’ of dots all seem to be oriented slightly differently. This is a technique that aims to obfuscate the halftone optical illusion and make the dot pattern less obvious.
Below is a closeup of one of my favorite cookbooks. Note that we can again see the halftone pattern at work here. Sometimes the various degrees of rotation of each of the halftone screens produce secondary, circular artifacts called “rosettes”. Note also here how the dots are more densely-packed. As a result, the image is a bit more crisp, and finer details are able to be resolved.
Let’s compare these images with one made on a photographic inkjet printer. Inkjet printing works a little differently than halftone screening; instead of a fixed grid, dots are randomly scattered by a nozzle in an inkjet printer. We can see the dots in the image below are smaller, and much more densely-packed. The overall print quality is higher than that of the typical grid-arranged halftone dots used above. The downside with inkjet printing, however, is that it’s slow and almost prohibitively-expensive to produce a book using this technique.
Let’s now consider this closeup, which is from one of the Modernist Cuisine cookbooks. Rather than modulating the size of the dots (“amplitude modulation”, or AM screening) in the halftone screen, these books were printed using a technique called “Stochastic screening”, or “frequency modulation” (FM) screening. Rather than adjusting the size of the dots, the color gradations in this image were made by adjusting the frequency of the dots. The dots themselves are the smallest and most densely-packed in all of these examples, offering crisp, sharp edges and no rosette artifacts. As a result, the images produced by stochastic screening have a great amount of fidelity and sharpness.
It is this stochastic screening technique that we will be using for our book. Having established that and finalized all our content, our attention now turns to making sure that the images that we see on our screens here in Chicago exactly match those that will come off the printing presses in China. We’ll discuss how this color pipeline works in our next update.
I’d like to begin this update with some dry, “status update” info before delving into an interesting design problem Sarah and I have been working on recently.
As has probably been evident over the last year, we’ve done a fair bit of wandering around and experimenting with what we collectively want this book to be. During the past month, however, we’ve shifted our efforts away from this experimentation and towards finalizing the book’s content. This means taking (or retaking) any remaining photography, completing all recipes and essays, and in general trying to get to a spot where we can confidently step back and say “Ok, this is our book”.
Being able to stabilize the bulk of our content allows us to work on some ancillary things that need to be addressed. A table of contents or an index, for example, can’t really be created until we know what page numbers need to be referred to. We have a few places where we refer to other pages or sections of the book, which up until this point have just contained placeholder text. There are lots of these kinds of little details, and we are working carefully to ensure the accuracy of each.
We’ve mentioned this before, but just to be clear: because we ended up adding quite a bit more content than we originally budgeted for, our delivery date has been affected. We’re currently aiming to deliver sometime around September of this year. This is obviously later than we’d originally estimated, but the scope of our book is richer than we’d originally planned.
This delivery date remains just an estimate, however, and there are still a good many unforeseeable factors that stand to influence it. When Sarah and I printed our own book, The Alinea Project, we happened to deliver its contents to our printer around July, which put us right at the height of “busy season” for book printing.
“Busy season?!” I asked iocolor, flabbergasted that such a thing even exists.
“Yeah. All the big book publishers want their books available before the holiday season, which means they all want to print stuff around July or August, to get books around October or November,” explained iocolor.
Because our print run for that book was so small (750 copies total), we were placed into the printer’s queue later, as a lower-priority project (and, incidentally, received our books after the Christmas rush had ended). Our book was further delayed by a dockworker strike in Los Angeles at the port to which our books were meant to be delivered. This strike lasted for several weeks, leaving our books languishing in a storage container on a cargo ship while Sarah and I helplessly stressed out about late deliveries to our Kickstarter backers for that project.
So, lots of unexpected things can go wrong; part of the next phase of this project involves factors that are out of our control, which will require us to shift energy from creative matters to managing these as best we can. Our updates will shift sharply to focus on the process of physically producing this book in the coming months.
To highlight one of the more interesting design challenges we’ve been working on over the past few weeks, I’d like to share a snazzy little video:
Anyone familiar with the English language (and, perhaps more specifically, with American language) could be forgiven for finding the lyrics of this song to be frustratingly familiar but not quite comprehensible. They are, in fact, complete gibberish. The singer, Adriano Celentano, is an Italian musician/comedian/actor who designed the song to sound uniquely like American music, despite the actual words having no meaning at all.
I find this completely fascinating. The act of decoupling words from their meanings frees a person to evaluate the lower-level qualities of language itself, which in turn can reveal interesting biases and assumptions.
Sarah and I are currently exploring a design challenge that swims in these same waters. The Aviary’s recipe for A Year In Kyoto is actually several recipes in one, comprising a tasting flight of four small cocktails that The Aviary created during Next’s Kaiseki menu several years ago. Each cocktail in the flight is meant to evoke one of the seasons of the year. When discussing the design of the flight, Chef Micah pointed with interest at our small flavor signifiers we’ve mentioned in previous updates:
“What if we include some text in Japanese that evokes something unique about each cocktail?” he asked. We thought this was an interesting idea, and after a bit of discussion we decided on a single Japanese word for each of the cocktails that we intended to place as a decorative element within the recipe.
Now, I should point out that Sarah and I are not Japanese. We neither speak it nor read it, and have zero Japanese ancestry in either of our families. So the issue of placing these words we’d chosen into the design of this book poses two major problems.
The first problem is obviously one of translation. This seems relatively easy to solve by using any number of online language translators. We hopped over to Google Translate and popped in one of our chosen words: “barbecue”. Here’s what that looks like:
Sweet, totally done, right? All we need to do now is copy-paste this into our design document, and move onto problem two, which is picking a typeface. We’re working on Macs, which come with a few pre-installed Japanese typefaces. Surely we can just pick one of these, dust our hands of the matter, and call it a day, right?
Let’s pause for a moment and consider an example of what this recipe looks like without our Japanese text added to it yet. What I’d like to look at, however, is not the recipe, but rather the text itself. To help force myself into a space where I can focus on this, I’m going to write a little bit of code to randomize the letters and words in this recipe. I want it to feel like an English-written recipe, but I ultimately want to render it meaningless gibberish:
With meaning removed from the words, we can focus more clearly on what Sarah is trying to communicate with the design decisions she’s making. What is the overall feel that this type conveys? To help offer a counterpoint, what if I re-set this text using a different typeface:
While the meaning (or lack thereof) of this text hasn’t changed, the page itself now feels very different. The above text is set in perhaps one of the most ubiquitous typefaces of all, the venerable Times New Roman. Times has been a default system font since the early days of desktop publishing, which makes it one of the most commonly-seen typefaces today. Before its appearance on computers, Times enjoyed an extensive career as newspaper type, having been created in the 1930’s for use in the eponymous British newspaper.
Times New Roman’s wide deployment renders it simultaneously everything and nothing, eminently readable yet totally unremarkable. Because it is so strongly associated with being a “default” typeface (undesigned web pages, e.g., with no other type opinion expressed, fall back to Times), the use of it subtly implies our page is undesigned, as if we simply offered zero consideration to the typeface choice at all.
What if we try a different typeface, this time Courier:
Courier is another widely-recognizable typeface, designed to mimic the output of a strike-on typewriter. This inherently puts our text in a particular time and place in history, and – because the typeface choice is likely to be widely-recognizable – also begins to draw on some associations a reader may have with other uses of this typeface. For example, are we reading a screenplay here? Is this some sort of top-secret dossier from a spy film? The Courier typeface has such a strong personality that it becomes dangerous to use because of the associations it inherently provokes.
We can continue this exercise with any number of other type choices:
In each case, there is no change in the text itself. But we can see how each typeface can provoke associations that may or may not be entirely intended, and can raise assumptions or biases about what the intended meaning of our unintelligible text might mean. It’s possible you might look at the above three examples and wonder if these recipes are meant to be part of a kid’s book, or maybe some sort of kitschy boutique.
This, of course, assumes that you have some exposure to the English language (and, again perhaps even more so, American culture and use of type). But what if you don’t? If you’ve never seen the typefaces in the above examples, you have no preconceptions about them. And if you don’t even know what the words mean, the matter is reduced to a purely aesthetic one of lines and shapes.
The fact that we can deliberately do this exercise with our English(-ish)-language text means we can also accidentally accomplish the same provocation of (potentially undesirable) bias by not approaching our Japanese translation and typesetting choices with a high degree of consideration. Worded more simply: simply slamming some words into Google Translate and applying a system-installed Japanese font, dusting our hands and calling this design task done stands to be grossly insensitive to anyone with any familiarity with Japanese language and design. It could possibly be the equivalent of using a typeface like this:
(The typeface above is an example of what some designers refer to as a “chop suey” font – sometimes also referred to as a “wonton font” or a “chopstick font” – a genre styled to mimic the brush strokes found in some Asian calligraphic writing. These typefaces are meant to convey some sense of “Orientalism”, but are in fact a completely Westernized fabrication that bear no history in or connection to actual Asian typography.)
So the decision to set four seemingly-simple words in Japanese has involved Sarah and I spending the last few weeks learning about the Japanese language and consulting with a few experts on our typeface choices. Type that looks interesting and appropriate to us has, in many cases, turned out to provoke strong associations in directions we did not intend. “This looks like an ad for cheap kimono,” was one bit of feedback we managed to provoke. “This is a typeface used by a large Japanese pharmaceutical company,” was another. It’s probably obvious that such a drastic juxtaposition in style might feel jarring and confusing to readers who have fluency with both Japanese and English language and design, and we want to be sensitive to that. Ultimately, we would like the two languages to sit together on the page harmoniously, both visually and in meaning. We recognize that the effort involved in honing this is likely to go unnoticed by many, but we still feel it is work worth doing.
This sensitivity is not incongruous with what we see practiced by those with whom we work. In fact, the staff of the Aviary (along with those at Next, Roister, and Alinea) all go to quite great lengths to educate themselves on matters such as these so as be prepared for moments in which they may be able to offer an extra modicum of care to a guest. There’s always the chance that the time may come where a small gesture of sensitivity can contribute immeasurably to someone else’s experience…the cocktails are just a small part of the equation.
Before we launched this Kickstarter, Chef Micah and I had a “practice run” photography session. The exercise was meant to get me thinking about how we might want to handle the photography of different kinds of drinks. Micah chose a range of drinks to test, including Meet Me In Tennessee, which he worried could look boring when photographed:
Because the drink had been presented to me in this way (“This one’s pretty boring, let’s see how we can make it look more interesting”), I’d already been primed to try some unusual lighting. Then, to add more visual interest, I generated some weird CG flourishes and composited them in:
Another drink that presented me with something to think about was Bring Another Smurf. This incredible drink is actually two cocktails in one glass, separated by a disc of ice. Upon being presented with the drink, a guest is invited to blow into a straw protruding from the glass. This disrupts the ice raft, turning it into a paddle that stirs the drink, which in turn triggers an interesting color change. For our test shoot, we affixed a tube attached to an aquarium pump to the straw to blow bubbles into the drink:
Looking at the photos, I realized that any single one of them sort of failed to properly capture the experience of seeing this color change. We could obviously present a spread of multiple images, but that felt a bit predictable. And in any case I was on this jag to think critically about how vfx could be used as a tool, so I started wondering about the idea of using computer graphics to try collapsing a span of time into a single photograph. I set up a fluid simulation of two liquids and wrote a small bit of code to diffuse their colors together over time. I then rendered out a few frames of this fluid simulation, and composited these into a single photograph of the Smurf.
After finishing these tests, I took a step back to consider them both.
To me, the Tennessee experiment wasn’t particularly successful. Sure, the weird little light curves do indeed add some extra visual interest…but to what end? Including some random computer-generated element that calls a lot of attention to itself for no apparent reason would be, I think, a clumsy application of this technology. It’s not really different than Chef Achatz’ general disdain for garnishes which serve no purpose on a plate, in fact.
The Smurf test, however, is much more interesting to me. It makes me wonder: can I selectively employ computer graphics in a way that supports the “story” of this cocktail? Does doing so provide more information in an effective or interesting way? Does it add anything, or is it just functionless digital garnish?
Even if I’m not directly augmenting an image with computer-generated elements, I find opportunity to leverage techniques I learned working as an artist in VFX studios. Let’s consider one example. Here’s a simple image of a computer-generated sphere:
We can see the sphere is illuminated by a single, blue-colored light coming from screen right
Here’s the same sphere lit by another light, this time a warm-colored light coming from screen left:
And finally, here’s the same sphere lit by some ambient light data I collected from The Office in NYC:
(This ambient light data looks sort of like a panorama):
A bit of physics knowledge teaches us that light has the peculiar property of being “additive”. This is a fancy way of simply saying that if you have two lights, and you shine them onto a surface, you end up casting “Light 1” + “Light 2” amount of light onto the surface. We can see this visually here:
In the image above, we can see my ambient-lit sphere, my left-lit sphere, and my right-lit sphere. If we add each corresponding pixel in each of the first three images, we end up with the result on the right.
Now, imagine you’re an artist showing a director the above, rightmost image. It’s possible the director might say, “I love this, except I wish the light coming from screen right was less blue”. It would normally be pretty tricky to isolate the contribution of the blue light alone in the rightmost image. But, because we actually have each light split out separately, we can do a neat trick where we can adjust just one of the images, re-add them up, and change our result:
In the above image, I’ve re-colored the screen-left and screen-right lights to be more neutral in color, and re-added everything up, resulting in a more neutral overall image.
I can even do interesting things where I adjust the intensity of the lights independently, as if they were each on a dimmer switch:
Here, I’ve boosted the intensity of my ambient light, and lowered the intensity of my screen-left light, which yields the result we see here.
Working this way allows a vfx artist to respond quickly to a director’s comments without having to re-do a lot of extra work.
When photographing cocktails in The Office, I knew I wanted to include the Office environment itself in my photos. The lighting in the office consists of lots of differently-colored lights: the bar itself is sort of purple, and is tinted by all the bottles sitting on it. The ceiling lights are all a very yellow tungsten color. And my flashes are “daylight-balanced”, which means they’re more blue-white in color. I suppose a “for realsies” photographer would know how to use colored gel filters to tint the color of each of these light sources to bring them all into balance. But, instead, my instinct has been to try using the above light-splitting idea when shooting in situations like this. Here’s an example of a very, very old bottle of bourbon, lit from above:
Here’s the same bottle, now light from screen-right:
And again from screen-left:
I can add these images similarly:
The above example is a relatively simple one, but other photos are a bit more complex:
Working this way has bought me a lot of flexibility: I don’t have to ask for reshoots nearly as often, and I can effectively change my lighting after the fact. This latter bit is useful if we find that this image might need to be placed on the left page in a spread vs. the right (which might encourage me to ‘relight’ it a bit to make the whole spread feel more balanced).
Another example of leveraging computer graphics: Sarah and I talked a bit about designing a bird flock motif to use as a design element throughout the book. Rather than hand-illustrating several flocks of birds, I offered to help Sarah with this work. I started with a simulation of particles:
Then I created an animated cycle of a bird flapping its wings:
I can then “stick” a copy of this flapping bird onto each particle. I can offset the animation forward or backward in time (so that the birds aren’t all flapping in unison).
Sarah then chooses a frame or two she likes, and we render them out for her to use in her design.
There are lots of other places Sarah and I are finding ways to use our collective vfx knowledge (Sarah also has an extensive career in the industry) to apply towards this book. The above examples are meant to highlight how we’re approaching this, attempting to apply our skills in ways that enhance rather than detract from things. We’re hoping the final result feels harmonious, balanced, and well-integrated.
Of all updates we’ve shared on this project, this is the one I’m admittedly the most excited for.
I’ve made plenty of noise about all the things that I have little to no experience with on this project. But, I’ve had the privilege of working as a visual effects (“vfx”) artist on feature films for about 20 years, which makes it a field I do know a little bit about. In our very first meeting about this book with Chef and Nick, Sarah and I asked, “how do you guys feel about the use of visual effects in this book?”
I felt sure I already knew the answer to this. The “branding” assumptions I mentioned in our Photography updates led me to assume that the images in this book should be spartan and minimalistic, focusing wholly on the drinks themselves. Surely computer-generated embellishments had no place in this project. But I was genuinely surprised when Chef and Nick both exclaimed “Awesome! Go for it!”
I was ecstatic! The idea of finding ways to apply at least some of my vfx experience to this drastically different context was super appealing to me, but I also wanted to be sensitive about how it was done. So I found myself in the unusual and very interesting position of thinking very critically about the role visual effects might play (if any) in a book about cocktails.
A bit of context to explain how I’ve approached thinking about this:
The thing that attracted me to visual effects in my younger years was that the whole thing seemed like a magic trick: I could type some stuff on my computer, and a picture would appear on my screen. It was so cool! The more time I spent typing, the cooler I could make the image.
When I entered the vfx industry, I quickly learned that visual effects in films can be a pretty time-consuming process: I once spent about 8 months working on a single shot in a film, something that flicked by on screen in less than a few seconds. That amount of time and emotional investment tended to mean I really wanted my work to shine – I wanted what I was making to call a lot of attention to itself. And for many years as a vfx artist, that’s kind of the attitude I carried with me.
This kind of myopia, however, can cause a “can’t see the forest for the trees” kind of problem. I’d often labor on details that became very important to me, but that were largely irrelevant to the film’s director or the audience itself. When showing a director a computer-generated splash of water that I’d labored lovingly over weeks, he might say “Can we make it smaller, and make it move to the right instead of to the left?”
“But that’s not physically accurate,” I’d argue. The director often wouldn’t care, but rather than trying to understand why, I just tended to feel frustrated and pouty.
Then I had a very memorable experience that really stuck firmly in my head, and that drastically reshaped my understanding of this art form. It came during my time working as an artist at Weta Digital in New Zealand. While developing the technology needed to make Avatar, Weta worked exhaustively to develop a cutting-edge way of recording facial animation from actors, then using this animation data to drive the facial animation of our computer-generated characters. The system saved thousands of hours that would have otherwise been spent painstakingly hand-animating characters’ faces to get similarly-believable results.
This system was subsequently used on Rise of the Planet of the Apes for the animation of the completely computer-generated ape, Caesar.
There was one notable exception to the use of this system, however. A climax of the film involves a moment in which Caesar’s trust is violated. Distraught, Caesar presses his face pleadingly against a thick plane of glass behind which he’s held captive. The decision was made, for this critical moment, to forego the use of the facial motion-capture system.
This decision scuttled the efforts of the facial tracking software team, and foisted a large workload on animators. It would have been far easier at this point to just use the tools that had been developed for this purpose. So why did Weta and the filmmakers do this?
The facial mocap rig was an intrusive, bulky device that sat right in front of an actor’s face, straight in their field of view. Trying to act like it’s not there is difficult; some psychological effort on the part of the actor is required to overcome this intrusiveness in order to deliver a believable performance. This moment in the film needed to involve the actor acting frantic and terrified, while pressing his face desperately against a pane of glass – something that would have been impossible with the facial rig on. In this case, the technology would have interfered with the authenticity needed to properly convey the story of this shot. So, rather than sacrificing this authenticity for the sake of convenience, the filmmakers chose to do things the hard way to create a more emotional moment in the film.
Understanding why this decision was made gave me an understanding of visual effects as a supportive storytelling tool, rather than something that clamors for attention in and of itself. I began developing a sensitivity to knowing when visual effects should shout, and when they should whisper. When considering how vfx might be applied in the context of a cookbook, it’s been this idea that I’ve tried to keep in mind as I explore.
The use of visual effects in a cookbook context also involves another peculiar issue. In the case of films, the objective is often to deliberately distort or fabricate reality. People tend to go see films to be entertained, and vfx allow a director to realize a vision that’s often impractical or impossible to create in the real world. But the images in a cookbook, in addition to being eye candy, also need to serve the functional purpose of illustrating how a final dish or drink should look. Falsifying reality in a cookbook threatens its credibility, and I worry it could ultimately undermine the efforts of our chefs and bartenders.
So my Golden Rule for approaching this has been “Do not mislead or distract”. It feels ethically irresponsible to manipulate images for the goal of misrepresenting the cocktails or suggesting an impossible standard for the reader. Instead, I’ve sought to leverage my experience with this stuff to enhance the book in ways that add beauty or clarity without misrepresenting anything.
We’ll see what that’s meant for the book in the next update.