Morgan Freeman: 285,000 brush strokes later
December 2, 2013 1:58 PM   Subscribe

Morgan Freeman, finger-painted on an iPad. (via)

Painted by Kyle Lambert, using Procreate. More videos here. Some people have been claiming this is a hoax, but apparently people that know Kyle say that it's not. Overlay available here.
posted by phaedon (65 comments total) 8 users marked this as a favorite
 
That the overlay matches so perfectly tends to argue that it is, in fact, a hoax.
posted by fatbird at 2:01 PM on December 2, 2013 [1 favorite]


I can't argue either which way, but Lambert's been digital painting since at least 2009, and on the iPad since 2010.
posted by phaedon at 2:12 PM on December 2, 2013


It's hard to imagine how you'd create that video if you couldn't do the work.
posted by Huck500 at 2:23 PM on December 2, 2013 [6 favorites]


I'm not familiar with the specific software in use, but in virtually every iPad drawing app I've used, it's easy to import an image and use it as a layer in a multi-layer image. It's trivial to start with a photo and draw on top of it on a semi-transparent layer, tracing it, basically. The video isn't continuous, so it would be relatively easy, I think, to work on an opaque surface, hiding the work layer to see the reference image, until you've built up something pretty accurate and reasonably close that lacks fine detail. Then you start using erasers to gradually remove your work, exposing the photograph underneath.
posted by fatbird at 2:32 PM on December 2, 2013


Looks to me like he started with the photo and painted on top of it to obscure and remove detail until he was left with a blank slate, then reversed the playback.
posted by ten pounds of inedita at 2:35 PM on December 2, 2013 [4 favorites]


Looks to me like he started with the photo and painted on top of it to obscure and remove detail until he was left with a blank slate, then reversed the playback.

Yep, that's the way you would do it alright..
posted by phaedon at 2:37 PM on December 2, 2013 [2 favorites]


It looks like he started with a block of marble and removed everything that wasn't Morgan Freeman.
posted by chavenet at 2:37 PM on December 2, 2013 [35 favorites]


I've been playing around with Papers and it's a great way to lose yourself for a couple hours.

And people say you can't create with an iPad! Ha!
posted by Blazecock Pileon at 2:39 PM on December 2, 2013


I have Papers and a few others, but Procreate really is something special -- better brushes, detail, layers. I'm no artist, so I just sort of.... smudge things.

I have no doubt it's possible that he drew this with those tools, but I bet he imported a photo for reference. Even so, a terrific feat.
posted by mochapickle at 2:42 PM on December 2, 2013 [1 favorite]


Hoax or not, I've never understood why you'd want to create pixel-perfect photorealistic art in the first place. I guess it's good training for original work, and there's something amazing about the meticulousness of the process, but if end result is the same picture of Morgan Freeman... well, I guess it's cool because it's an artisinal, locally-sourced picture of Morgan Freeman. Is it just considered a fun hobby?
posted by phooky at 2:42 PM on December 2, 2013 [10 favorites]


When will theae celebrities just provide their actual photo instead of these digital enhancements? Won't people think of the body image issues in baby booming african americans as a result of stunts like this?

And by digital, I mean finger painting with a computer...
posted by Nanukthedog at 2:45 PM on December 2, 2013


Five minutes ago if you'd asked me "Say, is Morgan Freeman a decent-looking guy?" I'd have said "Sure!" But now I realize he's damp and shiny and kinda gross-looking.* So there's that.

* yeah, I probably don't look any better, but you didn't ask about me, now did you?
posted by echo target at 2:50 PM on December 2, 2013 [3 favorites]


After watching that all the way through, my takeaway is that

Mr Freeman thinks I should make more of my life
posted by tigrrrlily at 2:52 PM on December 2, 2013 [1 favorite]


Looks to me like he started with the photo and painted on top of it to obscure and remove detail until he was left with a blank slate, then reversed the playback.

Really? The facial hair and the shading and highlighting work absolutely don't look like that to me.
posted by Huck500 at 2:53 PM on December 2, 2013 [2 favorites]


phooky: "Hoax or not, I've never understood why you'd want to create pixel-perfect photorealistic art in the first place. I guess it's good training for original work, and there's something amazing about the meticulousness of the process, but if end result is the same picture of Morgan Freeman..."

Because you can stop at step like 10 of 200 and get an awesome impressionistic version of morgan freeman? Or step 100 and get a painterly but expressive version, etc.

I've also seen some amazing photorealistic paintings where the craft was critical to the intent of the art.

Being able to do it is just another skill in the toolbox.
posted by danny the boy at 3:03 PM on December 2, 2013 [2 favorites]


I've also seen some amazing photorealistic paintings

So frustrating looking at photo-/super-real paintings in a gallery as someone who dabbles in paint. It's like, OKAY YOU WIN ALREADY, HAPPY NOW?

It's really amazing, although for whatever reason I find it more impressive when it's done with actual paints.
posted by Hoopo at 3:19 PM on December 2, 2013 [1 favorite]


I, too, vote fake.
posted by chasing at 3:26 PM on December 2, 2013


David Hockney also does some iPad stuff that appears as New Yorker cover art every once in a while.
posted by LionIndex at 3:28 PM on December 2, 2013


I think a lot of what I appreciate about something like this, the reason I can see some artists striving for realism, is the appreciation of detail, the exact fine details of what makes reality look real. To render them in pixels or paint, you have to have noticed them in the first place.

No kind of real artist myself, but having done some fooling around in Photoshop, I can buy that maybe he was actually working with the photo underneath for most of it and just recording frames for the video with that layer hidden, but I'm not sure that really takes away from my thinking there's a lot of talent here. I certainly don't see how you could get to all the intermediate steps of this by just... removing detail somehow? It's not like it's just a blurrier version or something.

It's not like it's a Van Gogh or anything, but it's really fascinating to watch.
posted by Sequence at 3:29 PM on December 2, 2013 [4 favorites]


To me it looks best around a minute in, where you can see Morgan and the artist's brushstrokes at the same time. Then it becomes too detailed, a minute later you feel like you're licking his moles, finally it zooms in on a nest of glistening writhing shoggoths and I was all tekeli-li, tekeli-li
posted by fleetmouse at 3:58 PM on December 2, 2013 [3 favorites]


ten pounds of inedita: "Looks to me like he started with the photo and painted on top of it to obscure and remove detail until he was left with a blank slate, then reversed the playback."

So what you're saying here is that, at that part around 1:20 where the facial hair starts coming in, each little tendril one by one, what was actually going on was that that was a reversed video of him erasing each little tendril one by one, meticulously avoiding all the other tendrils very carefully so as to preserve the illusion that each was being drawn?

That's a hell of a lot of work to go to just to avoid work.
posted by koeselitz at 3:59 PM on December 2, 2013 [5 favorites]


As someone who does try to draw and paint realistically, two things make me sceptical: first, he gets the proportions and placement right the first time, perfectly. There's no shifting things over a bit halfway through, there's no "that cheek needs to come in a bit", it's "bam, there it is." This makes me suspicious because it's incredibly difficult to get those things perfect from reference even with continuous adjustment over the course of the painting, let alone first time. Very often what looks perfect early on looks subtly wrong after a lot of work. On the other hand, if you're tracing a photograph, it's dead easy.

Second, the detail in the hair and the pores of the skin goes far past photorealism or figurative realism, into the absurd, and scales up the work tremendously. You basically can't draw the skin highlights that perfectly match real pores without working in miniature, which causes exponential time increase. Imagine going to from "I'm carefully drawing his face" to "I'm carefuly drawing this square tenth of an inch". On the other hand, if you're a decent painter, you can create something without that level of detail, and then appear to fill it in by erasing your work over the photo, which is what it looks like to me when he actually gets the surface detail in--which comes in just stupidly fast.
posted by fatbird at 4:04 PM on December 2, 2013 [8 favorites]


I was a beta tester on Procreate (back when it worked on the iPad 1) and this looks legit to me.

Customizable brushes are one of Procreate's great strengths. You can take a high contrast photo of pore highlights, or hair or whiskers and use that as the source for a texture. I made up a brush from an asphalt photo that they included as a preset called "bonobo chalk" in one of their releases.

The Procreate team are amazing developers.
posted by bonobothegreat at 4:26 PM on December 2, 2013 [2 favorites]


That's a hell of a lot of work to go to just to avoid work.

True. But yeah, nailing the proportions perfectly especially makes no sense. Even if it took him multiple attempts. And he nails it right out of the gate. Is he "tracing," i.e. projecting an image onto the iPad? I don't know, I don't have an iPad. To me that would still be legit but a caveat would be in order.
posted by phaedon at 4:26 PM on December 2, 2013


Clearly, the whole story isn't being presented in this video.

There can be no question that the painting is based on the photograph. This comparison makes it even more clear.

One possibility is that the artist looked at the photograph and recreated every single hair, highlight, and freckle with perfect precision in exactly the same position simply by eyeballing it. This seems impossible, especially if you look at the exact curvature of every single hair.

Another possibility is that the photo was on a layer below the painting, and the artist switched back and forth, referring to it from time to time, but still painted everything by hand. He then edited out all appearances of this lower layer. This is at least plausible. It's still impressive (a bit like the hypothesized Vermeer technique), but a bit disingenuous.

Another possibility is that elements of the photo were actually copied, manipulated, and mixed with legitimately hand-painted/traced elements. This seems to be the most likely explanation. You'll notice that each hair is not individually drawn one at a time. Every white hair appears suddenly in perfect position at 1:20, as if they had been copied, isolated, and pasted in together as one layer. This muddies the waters considerably. If you can paste in elements of the photo, there is no bright line between that and simply copying and pasting the entire thing. For me, this crosses the line into misrepresentation. I believe that this image was created on an iPad, and it may have taken 200 hours, and involved a lot of brushstrokes. But it is photo-realistic primarily because it is a manipulated photograph, not because of any superhuman ability on the part of the artist.
posted by designbot at 4:36 PM on December 2, 2013 [2 favorites]


If I wanted to fake something like this I would do it this way:
1. start with the final photograph
2. remove detail from a particular area or feature, either by blurring/smoothing or just painting on top of it; take a snapshot
3. repeat until the whole photo has been effectively erased and you have a bunch of snapshots
4. record a movie going *forward* from each snapshot to the next; so each movie just needs to be focused on a particular area, and your goal is to get the painting close enough to the next snapshot that you can just splice the movies together without the jump being too noticeable

Of course even if he was up to something like that, it would still be pretty impressive.
posted by equalpants at 4:45 PM on December 2, 2013


Hadn't bothered to look at the overlay and it seems pretty damning but it looks to me like the image in the middle may not actually be the painting. I'm on my iPad right now and can't really compare side by side.
posted by bonobothegreat at 4:52 PM on December 2, 2013


soundtrack ripped from: "Tunes to get robo-apocalypsed by."
posted by es_de_bah at 5:21 PM on December 2, 2013 [2 favorites]


From the conclusion of James Elkins' What Painting Is:

"In the end, what is painting? Is it the framed object, with its entourage of historical meanings, the gossip about its painter, and the ledgers and letters and files and reports and reviews and books it inspired? Or is painting a verb, a name for what happens when paint moves across a blank surface? Neither is complete without the other... The fundamental fact tht argues for the importance of the act of painting, is that painters spend their entire lives working with paint. There must be a reason: the practice of painting cannot possibly be just an annoyance, or an effiecient way to get images on to canvas. (As a way of telling stories or depicting objects, it is almost outlandishly inefficient. Practically anything else would be faster.)... No: painters love paint itself, so much that they spend years trying to get paint to behave the way they want it to, rather than abandoning it... It is the paint that is so absorbing, so deeply attractive, that a life spent in the studio can be a bearable life."

This isn't a painting. It's a decent tech demo.
posted by nímwunnan at 5:50 PM on December 2, 2013 [4 favorites]


I do photo-realistic portraits as well (but hand-drawn, not on a device) and I totally buy this as real. In the comparison from the overlay link, the two pictures are highly similar but not identical. For example, take a look at the white line coming off of the group of three moles in a row to the left of his right eye in the painting that's not in the photo (or in the overlay, interestingly). Hair is also a good place to look for differences because it's quite difficult to reproduce a head of hair with each individual hair drawn precisely right. You'll notice that it often looks off in portraits of people with lighter or straighter hair, since it's easier to see the individual strands. In this picture, the overall pattern of light and dark spots in the hair is the same, but the tiny spots are different. Finally, yes, he got the proportions right pretty quickly, but he also seemed to do some minor alterations of them as he went on. It's also not totally unbelievable for an experienced painter to be able to do that.
posted by quiet coyote at 5:55 PM on December 2, 2013


Finally, yes, he got the proportions right pretty quickly, but he also seemed to do some minor alterations of them as he went on. It's also not totally unbelievable for an experienced painter to be able to do that.

For my own edification, let me ask a question. Have you ever come this close to pixel perfect duplication in choosing to independently draw something super-realistically, given any amount of time? I retouch photos for a living, so coming from a sister form of digital editing, this just does not seem possible to me. Maybe it's not clear to me how photorealist painting works. Are you taking it as a given that he's either working over the photo or masking certain areas back in? I mean, maybe this is how this type of art is normally created, and I'm making a big deal out of nothing. It's just that the video strongly suggests that this "started out" as a blank canvas and a series of splotches. I think if that's "not really the case," then that's the part that bothers me.
posted by phaedon at 6:14 PM on December 2, 2013


What would the motive possibly be for faking it? Is there some black market for hand-drawn photo-realistic digital Morgan Freeman pictures or something? I can understand why someone might do it for real, as a tech demo or just as practice to keep up the digital-art chops. But I just can't imagine why someone would fake this.
posted by Cookiebastard at 6:32 PM on December 2, 2013 [1 favorite]


What would the motive possibly be for faking it?

There is this thing called "getting attention on the internet".
posted by Halloween Jack at 6:53 PM on December 2, 2013 [2 favorites]


I'm hoping Penn Jillette makes a documentary about this.
posted by mecran01 at 9:40 PM on December 2, 2013 [1 favorite]


Look at the video at 1:25. You can see highlights of the texture of the bottom lip appear. This is the kind of thing that happens throughout the video. Those elements are drawn to what appears to be pixel-perfection, with every crease, pore, and hair being in the identical place to the photo.

Whether or not this is faked, it absolutely could have been.

Here's how:

With Corel Painter, you can use a wide variety of brushes, with each brush-stroke — whether it's a soft airbrush, of a thin oil brush, picking up the colour values of the source image. So, you can move from pastels to airbrushes, to thin brush strokes throughout the process (as is done in this video). You can pick up specific values, as well, moving from highlights to darker areas.

So, you can paint selected areas, whimsically moving from, for example, mole to mole, and hair to hair. But your brushstrokes are just replicating detail in the existing photo. This is why the texture of the lips (the increased speed of the video notwithstanding) seem to just spray across that area. The artist need only move a fine brush that's set to paint lighter colours across the area, which he had previously "painted" by cloning darker colours with a less detailed brush.

The process looks uncannily like the process shown in this video does.

Here is an example of an artist using exactly this technique. It'll give you insight into why, in the end, every detail is identical.

In case the above doesn't make it clear, if this is indeed how the painting was produced, it should indeed be classified as a hoax (or a deception). An hour or two of familiarity with Painter, and no artistic talent is required to produce such a result.

By the way, the early sketch stages of the painting look every bit like the way one would painting using the above method, and nothing whatever like what artists' sketching process looks like.
posted by huron at 9:51 PM on December 2, 2013 [2 favorites]


There can be no greater tribute to the artist's skill and dedication that it's literally unimaginable by most commentators. And, btw, Procreate, built by a team in Tasmania with just one developer, rules.
posted by oluckyman at 2:59 AM on December 3, 2013 [1 favorite]


I kept waiting for that animated swarm to add details to that murky gimp eye, but alas, that's apparently how it looked in the original photo. What is a photorealistic painting of a face without the bright glimmer and intricately woven tendrils of the iris?
posted by dgaicun at 5:13 AM on December 3, 2013


Am I the only one who is disappointed that the video was not a 200 hour time-lapse of Morgan Freeman fingerpainting?
posted by Walleye at 5:43 AM on December 3, 2013 [6 favorites]


The artist responded to the skepticism here:
I'm overwhelmed by the response to my finger painting video and happy that it has reached as many people as it has in such a short time.

"One of the big reasons I wanted to paint this portrait was to create a video that entertained, inspired and educated people about digital art. I've met a number of artists who were motivated to start painting after seeing my previous videos. I've also heard several cases where teachers are using my videos to inspire their students.

After using Procreate on my iPad (the app I created this portrait with) I realised that for the first time I could create a really highly detailed digital painting and have the whole process documented in a video. This was when I decided to start this painting.

The process began with me blocking in the essential colours and shapes of the portrait and then reducing the brush size to add features and details. Each stage of the process took a step closer towards the painting becoming more realistic.

To answer your question, no at no stage was the original photograph on my iPad or inside the Procreate app. Procreate documents the entire painting process, so even if I wanted to import a photo layer it would have shown in the video export from the app.

The key to panting accurate detail at this obsessive level is to break down the portrait into much smaller paintings that are more achievable. Once you do that, the detail becomes easier to manage. However, the consequence of this is that the overall painting then takes a lot longer to complete.

I trained as an oil painter and have been painting portraits and concept art using my Mac and iPad now for around 10 years now. Essentially, I've spent a lot of time learning how to paint realistic looking images. All I have done here is adapted my usual technique to work with my fingers instead of a stylus or brush. Other than the fact I am using my finger, the process of painting on the iPad is identical to all other digital mediums and in many ways similar to working with real paint.

The whole point of the video is that the final image looks almost indistinguishable from the reference photo. If it didn't, then the claim of it being 'the world's most realistic finger painting' wouldn't really be a valid one.
posted by designbot at 7:43 AM on December 3, 2013 [2 favorites]


If he had any other work that came even close to this, I might be convinced it's not a hoax.
posted by buriednexttoyou at 8:04 AM on December 3, 2013


designbot: "Clearly, the whole story isn't being presented in this video... One possibility is that the artist looked at the photograph and recreated every single hair, highlight, and freckle with perfect precision in exactly the same position simply by eyeballing it. This seems impossible, especially if you look at the exact curvature of every single hair."

I still reject the notion that this is a hoax, mostly because it seems completely and entirely possible to me. But I do agree that the whole story clearly isn't presented in the video. That's pretty obvious. The main reason is that there is nobody with fingers that small. But the work process he describes renders this not only possible but even easy, if a person actually takes the several hundred hours necessary to do it. As he says, one has to break the image into manageable chunks.

The main thing the video clearly isn't showing is zooms. If the artist broke his painting and the reference image almost immediately into exact quadrants and zoomed directly and exactly into those quadrants to do his work, it's not really surprising that he got the proportions right immediately. I can imagine that this Procreate app likely has a grid overlay that would make this even easier; that's pretty standard among painting apps nowadays. If the reference photograph is detailed and large enough, at that point it's just a question of progressively zooming in on particular chunks and reproducing details like hairs and texture. Anyone who's used an ipad to draw with their fingers knows that, though he brags about the "finger-painting" aspect, it's actually pretty easy to be detailed so long as you zoom in close enough.

At that level, in some ways my experience is that tracing actually makes the process more difficult than just having a reference image. A reference image can be held up and consulted closely, whereas with close-up layer tracing the two images can become difficult to tell apart.

Movie-poster artists have been doing these things for decades; I'm not sure if they still do, but I do know that ridiculously photo-realistic paintings are common enough, to the point where most moviegoers are a bit shocked to discover that even up to seven years ago or so most movie posters were done by "hand" by airbrush artists. If it works for artists working in a studio using cut-up reference images and a huge canvas, you can bet it works for a guy with an ipad app that gives him reference grids and infinite zooming capability.

The video seems a bit impossible because it appears to have cut all that out, showing only the full picture with none of the actual zooming or grid overlays. That may be a bit deceptive, although in his defense a 200-hour video that included all that stuff would be impossible to watch if you time-lapsed it down to something this short.
posted by koeselitz at 8:21 AM on December 3, 2013 [1 favorite]


phooky: "Hoax or not, I've never understood why you'd want to create pixel-perfect photorealistic art in the first place. I guess it's good training for original work, and there's something amazing about the meticulousness of the process, but if end result is the same picture of Morgan Freeman... well, I guess it's cool because it's an artisinal, locally-sourced picture of Morgan Freeman. Is it just considered a fun hobby?"

And yet, a HUGE portion of the planet thinks photorealistic is the highest expression of art - David's Pieta, the stray hairs on Mona Lisa, Chuck Close's pre-accident giant murals - and lose respect for a piece the further it resides along the spectrum towards pure abstraction (Mondrian and Pollack, for example).

Taste varies widely - not essentially from "good" to "bad", but from spicy to salty to cherry to burnt crusty bits...
posted by IAmBroom at 11:45 AM on December 3, 2013


It took 200 hours to draw that image on an iPad. The artist clearly used a photo as a source. Now, 200 hours is 720k seconds. If one types one RGB value per second, a picture with 640x375 resolution can be reproduced in 200 hours. The resolution in Lambert's work is larger, so the process of reproducing a photograph using Procreate on iPad with a crazy amount of skill is somewhat more efficient than having an unskilled person type RGB values one after another for five work weeks (no lunch or bathroom breaks).
posted by ikalliom at 12:17 PM on December 3, 2013


There is no sense in which the reproduction is "pixel-perfect." That's a cute term, but it doesn't really apply here.
posted by koeselitz at 2:28 PM on December 3, 2013


Comparison (not perfectly aligned). They score quite high on any conceivable image similarity measure, but are obviously not exactly the same pixel values. I'm sorry if I seemed to suggest so. I guess I was just trying to say that 200 hours is a lot of work. One brush stroke every three seconds. Just a couple of pixels in the final image for each stroke (of course, how else could you get detail like that).
posted by ikalliom at 3:46 PM on December 3, 2013


Well, I stared at that comparison for quite a long time, and I kinda want a Pizza Hut margherita now?
posted by lucidium at 4:02 PM on December 3, 2013


Or 2 hours to use the paint clone tool described by huron, with a little wobbly cloning thrown in for a subtle touch of authenticity.
posted by fatbird at 4:27 PM on December 3, 2013


On this fraudster's Youtube channel (yes, I'm elevating my position from a suspicion to a certainty), you can see many of his time-lapse "paintings" done with the same essential cloning technique. A followup to yesterday's comments:

• There is an enormous divide in artistic quality in the paintings he's done with the cloning technique, and those he hasn't. Regarding the latter, he has roughly the ability that you'd expect to find in a high school yearbook.

• Some of his non-cloned work calls for, and makes a very shoddy attempt at the kind of detail, specular highlights, textures, etc., that he produces through the cloning technique. He fails miserably.

• There's a video of him demo-ing his use of an iPad to paint. This is the single instance on all his many videos showing him drawing live. Look closely. He does a half-baked (despite surely being rehearsed) attempt at this blotchy, "intuitive" method. He then draws a few ham-fisted hair strands, and, with a sort of flecks brush, scribbles some background colour. Then end. That's it. Mr. Lambert, don't leave us hanging! Why none of the brilliant and inspired flicks of the finger, with which you create your genius, magically resolving gestalt pointillism? That work that at ever stage has the essence, proportions, and dimensionality of the final work? Why didn't you carry on to the meat of your method, and give us a few of your divine splotches? This demo makes the possibilities quite clear: either he prefers to show an audience his high-school-yearbook level of talent, and then stop, and show previous video captures rather than carry on for a few more minutes with the astounding divination of colour and composition, or this is a big lie by someone who craves approval and admiration.

• Notice that the recent reply to the skepticism was met with words, not with the proof with which he could triumphantly prove the naysayers wrong. He puts a lot of effort into overlaying his video captures onto a fake iPad. Why not take a half hour, set up a camera, and show us his process in the flesh?

• To reiterate my previous comment, this "artist" does nothing whatever like what an actual artist does. There are a few lines of sketching here and there at the outset, for credibility, and then BAM! he's into his magic splotch pointillism. He peppers in a few lines here and there, but he does nothing like the exploration of a real artist — who must grapple with form, shape and dimension. Using the cloning technique, he tries to add plausibility to his work by doing something that looks something like a creative process by, for example, drawing a large dot of white (as always, picked up by the clone method) for one side of an eye, another dot for the other, and a big coloured dot for the iris. Then, he draws a final dot for the pupil. Continuing on, that eye becomes more and more detailed, yet always somehow right, and in accordance exactly with the final result. Again, this looks exactly like it does when cloning from a source image, as in the above Painter example.

• Adding another layer of giftedness to his process, unlike all his human looking art (which ranges from amateurish to awful), some of his cloning techniques show lovely shifts in colour spectrum, with surreal combinations of hue that nonetheless create a cogent image. This can be done either by being an extremely talented artist, or simply by modifying the curves of the original photo in Photoshop or Painter.

• It's important to note that this cloning technique isn't in the class of "digital painting". With the exception of selecting brushes, and pretending to paint specific regions at a time in order to look like the work is an artistic effort, Mr. Lambert (and anyone else) can do this kind of process with his back to his iPad — or in this case, his computer.

• If you search for "Corel Draw" in Youtube, you'll see videos of people painting and sketching with Painter, and others of people using exactly this cloning-with-brushstrokes technique. Unlike an actual sketching process, you will see them, like Lambert, drawing large blotches of colour that were picked up from the source image, and doing the same, skill-free process with smaller brushes here and there. What you won't see, is them doing a little crummy line-drawing here and there over the image, in an effort to put us on. The thing is, those artists aren't claiming to be doing anything other than what they're doing. Lambert is claiming to be painting those images, and he's lying.

Kyle Lambert, you are the second biggest douche in the universe.
posted by huron at 7:52 PM on December 3, 2013 [2 favorites]


There is no sense in which the reproduction is "pixel-perfect." That's a cute term, but it doesn't really apply here.

I refuse to believe that any artist, given any amount of time, would be able to reproduce a photograph with almost exactly the same spatial arrangement on an independent canvas without the use of some kind of system.

I will grant you that "in theory" this could this be possible without the use of the system, in the same sense that the sun might not rise tomorrow, but even then, it would not be achieved in a linear fashion, again without the use of a system. There should be a great deal of trial and error involved, and there wasn't. Not just "chiseling," but actual mistakes. And yet there are only very minor proportional changes.

So the painter responds with the following comment:

The key to panting accurate detail at this obsessive level is to break down the portrait into much smaller paintings that are more achievable. Once you do that, the detail becomes easier to manage. However, the consequence of this is that the overall painting then takes a lot longer to complete.

I am not familiar with Procreate enough to understand how this is achievable on the canvas end, but at least this implies that he is mapping out a grid on the original photo (think a photorealist version of Chuck Close) in order to break it up into manageable pieces, and then the iPad application assists him in organizing and working on the grid blocks one at a time.

It's a pretty nifty idea, but to be honest, if you were to use such a system, you would think that the artist would draw each grid block independently to its greatest detail, thus giving rise to the complete photo. And yet it is clear in the video that that is not the workflow that he is using. He jumps around the photo the way a natural drawer would, completing entire facial features as if he were dealing with the entire photo all at once.

So he admits to breaking down the photo into bite-size pieces, but then doesn't actually draw as if he's broken it down into bite-size pieces! This is a perplexing decision.

Comparison (not perfectly aligned). They score quite high on any conceivable image similarity measure, but are obviously not exactly the same pixel values. I'm sorry if I seemed to suggest so.

Ok, looking at that photo, (scale the top layer to like 99.5% and move it up a hair, it matches better) clearly we are looking at two different images. There is definitely color variation in the skin splotches and very slight spatial variation in the hairs. I just don't know if this concretely proves anything.
posted by phaedon at 2:38 AM on December 4, 2013


I don't know this particular program he used to know if he's fake or not.

200 hours is a long time to spend on one picture - using photoshop I could produce something as photorealistic in a 10th of that

I've never understood why you'd want to create pixel-perfect photorealistic art in the first place.

You learn a hell of a lot artistically doing it - especially things like representing texture etc. Then again I never do stuff that's exactly pixel-perfect, there's always going to be a few tweaks here and there.
posted by fearfulsymmetry at 5:17 AM on December 4, 2013 [1 favorite]


phaedon: "I refuse to believe that any artist, given any amount of time, would be able to reproduce a photograph with almost exactly the same spatial arrangement on an independent canvas without the use of some kind of system."

Yeah, I agree. Like I said before, it seems obvious that a system is being used. Hell, it seems obvious that more is happening than we see; there must be zooming at least. The zooming is removed to show a smooth picture so you can see the whole as it develops. Presumably the grid overlay was removed, too.

"So he admits to breaking down the photo into bite-size pieces, but then doesn't actually draw as if he's broken it down into bite-size pieces! This is a perplexing decision."

I'm not sure this is true, actually. There are cases where he clearly zooms back out and works in broad strokes, but they're cases where that makes sense - where he produces a texture and then clones it, for example, or where he's working on an area where there's something that's the same for a larger area. It actually doesn't make much sense to completely isolate oneself to single blocks and do them completely; when you do that you end up with a quilt-looking series of blocks that aren't homogenous, and you have to recreate everything (textures, hair brushstrokes, etc) from whole cloth every time. In a sketching program where you can zoom in and zoom out while maintaining the grid overlay for reference, the kind of workflow he's using makes sense.

huron: "Kyle Lambert, you are the second biggest douche in the universe."

You really don't seem to understand how ipad art actually works. This is completely possible, and Lambert's workflow in this video makes sense.

"Mr. Lambert, don't leave us hanging! Why none of the brilliant and inspired flicks of the finger, with which you create your genius, magically resolving gestalt pointillism?"

Er - you do realize that this is time-lapse, right? You know how time-lapse works? It is faster than the picture was created. This means that "flicks of the finger" are generally ten-minute chunks of time.

"That work that at ever stage has the essence, proportions, and dimensionality of the final work?"

Grids. He said it, I said it, you didn't listen, so I'll say it again. Grids. This is very, very simple stuff. Copying a photograph is not painfully difficult; it just takes time.

"To reiterate my previous comment, this 'artist' does nothing whatever like what an actual artist does. There are a few lines of sketching here and there at the outset, for credibility, and then BAM! he's into his magic splotch pointillism. He peppers in a few lines here and there, but he does nothing like the exploration of a real artist — who must grapple with form, shape and dimension. Using the cloning technique, he tries to add plausibility to his work by doing something that looks something like a creative process by, for example, drawing a large dot of white (as always, picked up by the clone method) for one side of an eye, another dot for the other, and a big coloured dot for the iris. Then, he draws a final dot for the pupil. Continuing on, that eye becomes more and more detailed, yet always somehow right, and in accordance exactly with the final result. Again, this looks exactly like it does when cloning from a source image, as in the above Painter example."

Again, I gather you don't understand how pictures like this actually work. When you're working from a grid, and when we're looking at a time-lapse of two hundred hours, we won't see "grappling with form, shape and dimension." It's not even necessary. You reproduce what's in the grid on a specific square. You go from square to square drawing the outline, and then it's perfect. From the outset. No "grappling" necessary. Artists reproducing photographs don't "explore." They replicate, as closely as they can. There is something creative about that process, but if you're suggesting that he should be drawing a dozen outlines of everything when he's already got a grid laid out, I'm not sure why.

The descriptions of this video that you're giving aren't really correct, either:

"There are a few lines of sketching here and there at the outset, for credibility, and then BAM! he's into his magic splotch pointillism."

Sketching is not a process that takes two hundred hours. If you sketch for two hundred hours, you'll never get anywhere. It would make sense to sketch for a few minutes, maybe even an hour - and it looks like he did indeed sketch outlines for about an hour, from the final video. "Magic splotch pointillism" is also not a great description of how he goes about this. He's kind of working in layers, and one can note that he'll work on a texture for a while and then blend it back down smooth. How would one do this reversed - make a texture arise slowly and magically out of a smooth background? Some sort of precision sharpen tool? This seems implausible to me.

Please note that I don't know Kyle Lambert, and I was not there when he did this, so who knows what exactly he did. All I'm saying is this: nothing you've said is very convincing, huron, and what's being done here is not only possible but easily achievable by most artists if they put in the work.

One question: if you're so convinced that this is fake, can you explain how Kyle Lambert created his "final image" out of the photograph? They're obviously very different, particularly if you look closely, as phaedon noted above. But you're saying this is a cheap knock-off. So how did he get from the photograph to his "final image"? What magic did he apply?
posted by koeselitz at 8:43 AM on December 4, 2013 [1 favorite]


fearfulsymmetry: "I don't know this particular program he used to know if he's fake or not. 200 hours is a long time to spend on one picture - using photoshop I could produce something as photorealistic in a 10th of that"

Yeah, that's pretty much my attitude.
posted by koeselitz at 8:44 AM on December 4, 2013


The flaw with the claim that he worked on grids is that, when you work on a grid, you literally work to a grid--you fill in a square, you fill in the next square. The nature of grid copying is that you don't have to do any "overall" work where you build up the larger forms, and then narrow your focus, working in a progressively more detailed fashion. If this was simple grid copying, it would fill in by grids.

Now, he could have taken it from high level to mid-level and started with the grids then, now that he needed to zoom in that tightly... but then you're back the problem of how he laid down the major lines and relations perfectly the first time, without a grid. Put a grid down, work out the major forms, work overall, and then go back to the grid? If you're working by grid, there's no value to large scale work because the whole point of the grid is to take care of that for you. It's wasted effort.

Again, this is not done how artists normally work, either in the large or by grid. He's putting on a good show, but it's a show.

Not to mention that any use of a clone tool pretty much disqualifies this as "photorealistic painting".
posted by fatbird at 11:09 AM on December 4, 2013


fatbird: “The flaw with the claim that he worked on grids is that, when you work on a grid, you literally work to a grid--you fill in a square, you fill in the next square.”

Like I said, that is not really how artists work on grids.
posted by koeselitz at 2:27 PM on December 4, 2013


On rereading Lambert's comment, I see that grids are a derail because he doesn't claim to use grids--which would have to be visible in the video anyway, in order to be of benefit as a set of reference points. You either grid both the photo and your canvas, or it's useless.

What he claims is that he's progressively zooming in and working at a smaller scale, going down to the pore level. Fine as it goes, and broadly correct as general process, except now we're back to the initial problem of everything going in right the first time with virtually no adjustment.

Here's a video where you see the full process by another artist: http://vimeo.com/69303413. Notice that grids have nothing to do with it, that there's constant adjustment throughout the process as the level of detail increases, and that the very realist effect at the end depends not in minute detail like pores but on a coherent whole that's local to the image.
posted by fatbird at 2:55 PM on December 4, 2013 [1 favorite]


More generally, artists today tend not to use grids. If you're going to rely on some form of aid to get lines and relations correct, these days you project and trace. A grid is a very wishy washy step where you're neither doing it by sight, nor getting a perfectly correct map from which to start.
posted by fatbird at 2:58 PM on December 4, 2013


Presumably the grid overlay was removed, too.

Then the video is an untrustworthy record of the process. If he can remove the grid, he can likewise remove the semi-transparent layer with the photograph on it.
posted by fatbird at 3:35 PM on December 4, 2013 [1 favorite]


fatbird: "On rereading Lambert's comment, I see that grids are a derail because he doesn't claim to use grids--which would have to be visible in the video anyway, in order to be of benefit as a set of reference points. You either grid both the photo and your canvas, or it's useless."

He says he's "breaking it into smaller paintings." I think you could read that as grids - or as something a little more vague. Also, I'm not sure grids would show. Procreate records brushmarks and imported images, according to what I can find on their app (since I'm not paying the six bucks to try it myself.) A lot of apps have transparent grid overlays that I'm betting wouldn't show in the video.
posted by koeselitz at 4:17 PM on December 4, 2013 [1 favorite]


There's a variety of ways to use grids. At one end of the spectrum, you have slavishly copying the values of a grid square from reference to canvas/iPad, as an exercise in transferring values and shapes. High School art classes often have this as an exercise just to illustrate how representation is independent of understanding the subject. At the other end, you have putting a grid up to have a regular set of reference points against which to check the subject.

Lambert is obviously not doing the former because that way of working literally proceeds square by square, and can achieve tremendous fidelity just to the extent that you manage to treat each square as an abstract copying exercise.

I remain very sceptical because, if Lambert is using a grid, he's doing the latter method and zooming a lot. Well and good, but it doesn't change the fact that one still works continuously, in that mode, adjusting and balancing things lines and relations. Watch the video I linked; watch any of the hundreds of process videos on youtube where painters demonstrate how they go from blank canvas to very realistic depictions. It's not simply a matter of getting more detailed; there's continuous adjustment, and at the end, the realism is internal to the image, not fidelity to the subject. Lambert largely makes pass after pass after pass, going from thick blotchy lines to small lines until, with a magical swipe of his finger, pores and accurate hairs appear. And at the end, the verisimilitude with the original photograph is almost perfect, and the variation is easily achieved by adjusting the color curve or jittering a clone stamp.

And all that discussion above ignores that he gets the colors right, too, with very little mistake, including relative colors. That's incredibly hard. Being able to reproduce colors exactly from sight takes tremendous knowledge of optical effects and color theory, and Lambert largely just gets it correct as he goes along, without needing to lighten or darken or shift it along the spectrum. And a grid is zero help for this, arguably the more difficult part of painting.
posted by fatbird at 7:01 PM on December 4, 2013 [1 favorite]


And all that discussion above ignores that he gets the colors right, too, with very little mistake, including relative colors. That's incredibly hard. Being able to reproduce colors exactly from sight takes tremendous knowledge of optical effects and color theory, and Lambert largely just gets it correct as he goes along, without needing to lighten or darken or shift it along the spectrum. And a grid is zero help for this, arguably the more difficult part of painting.

That's actually another great point. My guess would be he picked a swatch of colors up off the digital photo before starting and would say this was besides the point.

Koeselitz, the painting and the photograph are indeed "very different" on a molecular level but I should point out that the general faithfulness to the original photo overwhelms any notion that this was drawn independently on a separate canvas. At this point, I think where the breakdown in our conversation is happening is - if Lambert is breaking the photo down into a hundreds of smaller canvases, he does not in fact show us what his brushstrokes looks like on that smaller level. You know what I'm saying? Despite being a "clear demo," it very quickly turns into a black box scenario. Which is what I think is driving the amount of speculation.

Here is a closeup of the beard for reference. At this point I couldn't even tell you which frame is the original. I retouch photographs for a living, and this informs my perception of the situation albeit photorealistic creations are not my specialty. All things being equal, my immediate thought is, from this perspective, this was traced over, color balance is slightly off, some small liquifies, a slight high pass sharpen has been applied, and boom you're done.

I then concede that this was built from scratch for the sake of argument. I ask how this is possible because it's too fucking accurate. I am told that this was broken down into smaller blocks. At this point the reverse kicks in. If you are slavishly duplicating sections of the photograph, why is there any variation at all? This again suggests to me the use of a reference.

At this point, I am asked to choose if the reference is a photo overlay or a grid system. I am very willing to appreciate the ingenuity of breaking this up into smaller paintings, but again, I'm surprised that the time lapse shows him painting all over the place instead of one block time at a time. You're right, totally reasonable.. but you have a grid system that is idiot proof! It would seem insane zooming in and out from grid to grid working on "layers of the face" like a normal painter would.

This would require switching back and forth and referring back to multiple grids over and over again, when what such a system would seem to avail you is the ability to focus on recreating an incredible amount detail on one given place at one given point of time. And that's where I get lost again, throw my hands up in the air, and say I find it hard to believe that I'm hearing the whole story.
posted by phaedon at 10:59 PM on December 4, 2013 [1 favorite]


What really bugs me at this point is that I can't figure out how he did it or why. It somehow seems like it wasn't simple, whatever the hell it was. And I can imagine that the Procreate people would be mildly upset to find out this was faked. But it is hard to make sense of no matter how I look at it, I confess. Possible? Yeah, probably. As fearfulsymmetry said above, something that appears that photorealistic would be possible in Photoshop. But that exacting? Eh? Who knows. Why even try? And why fake it?
posted by koeselitz at 1:57 AM on December 5, 2013 [1 favorite]


I am waiting for the man-hours spent debating the truthiness of the portrait work to exceed the claimed time on the portrait. Then we'll have a new portrait!
posted by jscott at 3:28 AM on December 5, 2013 [1 favorite]


I am waiting for the man-hours spent debating the truthiness of the portrait work to exceed the claimed time on the portrait. Then we'll have a new portrait!

Hilarious.

As fearfulsymmetry said above, something that appears that photorealistic would be possible in Photoshop

And only with a stylus.

What really bugs me at this point is that I can't figure out how he did it or why.

Well and that's the interesting thing. I'll be damned if I don't download ProCreate in the next week or two.

The thing no one talks about in photography, photoshop tutorializing etc is my god.. the education business is booming. Everyone wants to learn how to photograph, draw, etc. Everyone wants to be on top of the guy that "I can't believe he did this or that." It's crazy. I've been doing this for more than a couple of years now and there is just no end to the places you can spend money "learning how to _____." Every other articles is about how the photography world is collapsing, but photo education?! Money!!
posted by phaedon at 9:05 AM on December 5, 2013


I'm going to forgo spending Christmas with my family and produce a photorealistic portrait of this thread
posted by fearfulsymmetry at 9:09 AM on December 5, 2013 [1 favorite]


I'll stipulate at this point that it's possible it is just what it claims to be: Lambert looking at a photograph and spending 200 hours painting it in every final detail. I'm very sceptical, but I'm not saying it's impossible.

That throat-clearing out of the way, I suspect Lambert is at least moderately talented enough to produce figurative images that get him positive responses, and found a tool/method that allows him to create these awe-inspiring videos that have people all over the Internet gushing over him. An hour of so of noodling with the more interesting features of Procreate and suddenly, he's a micro-celebrity! It's fun! And Vermeer used a camera obscura, so this modern equivalent is kosher! It's just a shortcut! I could totally do it if I actually spent the 200 hours!

For my parent's 45th wedding anniversary, I promised a watercolor painting of a boat from the harbor by their house; Dad provided the picture. On their 49th anniversary, I actually delivered the damned thing. On the one hand, it's approx. photorealistic--from 10 feet away you might mistake it for a photo. Closer up it's obviously a painting.

Purely as a labor saving device, I projected the photo and traced the lines, to start with an accurate map. That took an hour; cleaning up/firming up the drawing added another hour. It was a 24"x36" block, so doing sight/grid drawing from the photograph would have taken days, and likely been much less accurate overall, and gotten a lot pencil and eraser marks all over the surface that I'd have to work around after. For what I consider to be my real work, I don't do that, because I consider draughtsmanship to be central to what I do. But this was a "make Mom happy/shut her up" effort, and they're deliriously happy with it, and anyway the college stuff I did that it replaced is happily gone now. When they die, I'm the closest to their house, and my siblings will arrive to a bonfire of my paintings where Mom asked for "this picture, but in a red that matches the couch."

Anyway, my painting on that piece was sophisticated coloring book stuff. It looks great, and appears to be a demonstration of great skill, and it's not. But my parents are happy, and their friends gush at me. And part of me is happy to look at it, and say "I did that. I made that. I'm responsible for bringing that into the world, however it was made."

That's all by way of saying that actually making something that's recognized as art has motivations that are much wider than purely artistic concerns, and if you look into the lives of artists, their methods, and their character, you're in for a carnival of human misbehaviour. Durer used to lowball commissions to get them and then spend years wheedling more money out his patrons with claims that only the finest lapis lazuli would do justice to his majesty's greatness.

Besides the video I mentioned above, there's an interesting set of videos by a painter named Scott Waddell, who sells full length videos on classical painting technique; his video channel with previews actually gives a pretty clear idea about normal process, and this one in particular, on painting an ear, gets into the technical discussion of light and how a realist painter needs to understand it. Another artist who gets into interesting discussions of this stuff is James Gurney, whose blog is just a joy to follow, and he does direct stuff on the optics side to explain it (such as this, on gamut masking).

Lastly, in my youtube tours looking for counterexamples on this video, I came across this guy, who was a painter making fakes, who starting selling them as forgeries in co-operation with a dealer, and went to jail for it, and is now the host of an ITV Granada show called The Forger's Mastercraft, where he teaches students how to work through a particular artist's style and learn from it.

There's so much unbelievably awesome stuff out there for artists, from higher quality materials than previous artists every dreamed of having, cheaply available, to videos by real experts offering the kind of guidance I'd have killed for in art school, to endless blog posts discussing the science of it all, and repositories of images of the great works, that previously you needed to tour Europe or the far east to see.
posted by fatbird at 9:24 AM on December 5, 2013 [3 favorites]


« Older Anyone Who Wouldn't Buy "Frolik Navidad" Should...   |   "Are these the shadows of the things that will be... Newer »


This thread has been archived and is closed to new comments