Category: Technology

Ideas about the mechanics, techniques, and technology behind image making.

Posted on

My Favorite Lens

In general, we photographers love our equipment, Especially our lenses. It is not uncommon to have a favorite one. You can always get a discussion/fight started among photographers when you talk about lenses. I would like to discuss what has become my favorite lens.

Lenses

The lens is a critically important piece of equipment to photographers. Sensors are improving dramatically and lenses have to improve with them to achieve all the sharpness and resolution the sensor can capture.

Modern lenses constantly improve in resolution. Look at DxO image tests of current best lenses vs. the best from 20 years ago. Our lenses now enable us to capture more information and be able to produce wall-size prints that are extremely sharp.

The lens determines the point of view that is captured in our frame. It establishes the field of view, the width of the scene we are capturing. Some of us naturally have a telephoto view. Others have a wide angle view. This refers to the lens choice we tend to select to frame our subjects. This is just personal preference. The lens is a tool to help express our esthetic.

Many photographers feel they need a whole bag full of lenses of various focal lengths from extreme wide angle to super telephoto, with macro lenses and tilt/shiftes thrown in. Because, you never know what you might find. 🙂 Personally I have simplified my life a lot over time. I generally only carry a 24-70mm and a 70-200mm in my kit. But that is just me and where I am at right now.

So what we want is a lens or lenses that allow us to capture all the information we want (resolution, sharpness, dynamic range) in the field of view we want. A big ask, but doable.

Digital workflow

Most of us are in the digital world now. The digital workflow is quite different from the analog workflow.

What we call the analog workflow – the film days- involved developers and enlargers and prints and lots of chemicals and time. Personally, these are days I don’t miss. I am a big fan of the power and freedom and flexibility we have now.

There is a corresponding workflow for digital processing, though. It includes loading images on our computer, viewing them, culling or grading them, processing selects with our software of choice, etc. Each of theses steps is time consuming. Especially since we tend to shoot so many more frames now that they “don’t cost anything”. And each step requires software and considerable training.

The result, though, is that we spend a lot of time in front of our computer now. We probably spend more time in the digital workflow than we did in the analog workflow.

My favorite lens

What does this have to do with a discussion of my favorite lens? Well, in a sense the “lens” I use the most and that has the most impact on my work is my computer monitor.

This is where I view all my images. Zoomed in to 100% I look at individual pixels. Here is where I crop and color correct and adjust tones and contrast and saturation. This is where I view and edit the image when I convert it to black & white. When I create new images by compositing others together, that is done entirely though the monitor “lens”.

Yes, all of the things I just said are actually done through specialized software. In my case it is primarily Lightroom Classic and Photoshop. But metaphorically and to me, the monitor is the lens into the process.

Now days the monitor is where we view everything we do. Regardless of what the original image looked like, what I see in the monitor at the end of the edits is what counts. The result could be a complete re-imagining of the starting image.

The new primary lens

I spend more time in front of my monitor than I do outside shooting. More and more it is coming to dominate my workflow. If I lost or broke a lens, that would be terrible, but I could continue doing my art with other lenses with only minor re-adjustments to my vision. I had this experience recently. My 24-70 lens dropped and shattered the polarizer filter. I was up in the mountains and I did not have a filter wrench with me to get the jammed filter off, so I had to switch to using an alternate lens. A little frustrating, but not a big deal.

But if my computer died, although I could continue shooting, I could not view or process a single image until I fixed it. Eventually things would back up to a critical point and I would have to get the computer back. I also couldn’t select images for galleries or process images for printing. Dead in the water.

So in a sense, the focal point of the digital workflow is the monitor. That is the new lens I use to view and do most of my work. The monitor is the lens for the increasingly important part of the digital workflow.

The future

In the future will this trend increase or will we return to simpler times? What do you think?

My money would be on the increase of digital processing. We will trend more toward an attitude that the camera and lenses are used to gather raw material, but pictures are actually made in the computer, looking through the monitor. Increasingly, the final image may look less and less like the original capture. Better processing software opens up new possibilities. And viewers are more willing to accept that photography should create something more than a true representation of reality.

So the next time you are lusting for a wonderful new lens, it might be better to upgrade your monitor instead.

Posted on

Created by Me

ICM blur of dead tree. Take that, generative AI.

Generative AI is all the rage now. I suppose there might be some applications for it, but you will not see any of it in my work. What I show is entirely created by me, and I have no plans of ever changing that.

It’s all around

The news is full of hype about ChatGPT and Bard and, for images, DALL-E 2. Tech companies are inventing hundreds of billions (yes, “billions”) in it, so it must be about to take over everything, right?

It is hard to read anything without seeing references to the coming revolution. It is the “next big thing” in tech. MIcrosoft, for instance, has invested huge in ChatGPT and says it will embed it in its browser and all of its applications. With so much press and money and interest, it must be true, right? Maybe.

But do you understand what it is?

What is AI?

I have said before that I am a reforming Engineer. Well, I must admit that at one time I was involved in AI applications. I even believed in it at the time. That is just to say I have some technical background in the subject, so I am not just quoting press releases.

“Artificial Intelligence” is a weird term. It is definitely artificial. Whether or not it represents intelligence is debatable. To me, there is no real “intelligence” involved. It is just a fancy computer algorithm with a lot of data embedded in it.

The AI that is hyped today is called neural networks. It is based on a fairly simple structure that tries to mimic the way the human brain is organized by simulating neurons and synapses. Then they train the network with huge sets of data. The connections and values of the neurons and synapses are adjusted to give a desired output for a given input.

To over-simplify it, imagine a patient teacher trying to train a neural net to recognize an egg. They “show” it a picture of an egg and say “this is an egg” and let the network adjust its values to give a positive output. Then they show it a picture of something else and say “this is not an egg” and again let the network adjust its values to give a negative output. Repeat it over and over thousands, maybe millions of times with different pictures. Eventually the neural net would get pretty good at identifying an egg, if the training data was good enough and extensive enough.

But so what? The AI does not at that point know what an egg is. It just classifies shapes as being one or not.

What is the good of it?

We are discussing generative AI, so I will try to focus on that. Generative AI takes a request to make a picture or song or some such work, maybe based on the style of another artist. You could say “make a picture of a tree in the style of van Gogh”. It would make one. It would probably look like something Vincent might have done.

If you were generating the image for an advertisement, you might be able to simulate a certain style without the encumbrances of creative fees or intellectual property laws. For you, the user of the image, you get to bypass paying the artist. Or maybe, charitably, you get something you wish the artist had created, but they did not.

Many companies are very eager to have AI trained to be able to produce minimally acceptable results faster and cheaper than a human. Be aware of those companies that want to get rid of their people and replace them with minimal acceptable results. Have you used an AI-based chat agent to try to get support from a company? My results have been way below minimum acceptable. Maybe search engines is the best application for these bots. Most of the search results already can’t be trusted.

So for someone wanting something cheap for a practical use, it can be a good thing.

Is it art? I have my opinion, but let’s get to that in a minute.

What are the limitations?

Neural network-based AI only “knows” what it is trained to do. Its abilities are limited strictly by the data it is fed. And I used “know” in quotes because, one of the great limitations of this system is that it doesn’t know what it knows. It doesn’t even know what knowing is.

AI cannot explain it’s actions. The data compressed into its network has been stripped away from its source. This is going to become one of the major limitations that will cripple it or stop it’s use. So, for instance, when an AI system turns you down for a loan, you cannot force it to explain why. All it can say is that you just didn’t meet the pattern. Lawsuits will come of this.

And it may produce wonderful seeming results, but it is a cheap trick. AI products are a regression to the average, at best. That is, a large set of training data defines the average of whatever domain is being learned. This is all it knows. It does not understand the difference between unacceptable and acceptable and exceptional results. It does not understand the concepts behind what it is doing at all.

So when you ask it to make a picture of a tree in the style of van Gogh, its data bank has many images of trees. It has encodings of parameters describing patterns of van Gogh’s style. It can mix them and make something. But it can’t step back and say “Wow, that is great. I’m proud of that! That is good art.” There is no more feeling than a tax form.

Where does the training data come from?

This is a little off topic of the quality of the results, but have you considered where this huge volume of training data comes from? Google, Meta, Amazon, Microsoft and many, many others, including your Government, collect and use all the information they can find . This includes public data like Wikipedia or the Library of Congress, but also everything they can scrape up about you. So every network search you have ever done, every web page you have ever visited, all of your email, all of your pictures, your contacts, your contacts contacts, every post you have ever made, your facial images, your job, your salary, your spending habits, all of your telephone calls, everything is just free data to them.

This is all used without your permission or control. So for an artist, for example, all of their online works can be used to train the AI to do better to try to replace them. And with no compensation or attribution.

There is currently no accountability for AI or the companies profiting from it. It has been proven that much of the training data used was biased or incorrect, producing bogus responses from ChatGPT. And Google’s Bard got a black eye the day it was announced when it gave false information to a query about the Hubble telescope. No accountability, no ability to explain.

A passing fad?

One part of me thinks AI is just another passing fad. It has come and gone before. AI was going to revolutionize the world about 20 years ago or so. It died. Now the pundits are enamored with it again. Most of them are too young to realize it died of natural causes already. But venture capitalists and tech gurus are very quick to throw billions of dollars at “the next big thing”, even if it has been unable to generate any money.

But no, I’m afraid we will have to live with this for a while. Too many billions have been invested for it to die soon. And it can show some limited tricks. Either you believe AI is a higher and more perfect form of life that will make the world better or you don’t. I don’t.

Not on my watch

Lots of rambling, but back to the adoption of generative AI. As far as I can see, I will never use this in my art. This is not like the introduction of digital imaging, where film purists wailed about the passing of a wonderful era. This is not a technology shift, it is a tool that plans to eliminate artists.

I will use useful tools, like sky selection in LIghtroom or Photoshop, but that is just a force multiplier to get my job done quicker. I could do the same thing myself and I can often get better results. It is like a woodworker using a planer to smooth a tabletop quickly rather than spending hours sanding it. You don’t say the tool created the piece of furniture.

When you see images from me, they were created personally by me. I don’t and do not plan to use AI to create my art. I don’t think art created by AI is really art, but that gets into the argument of what art is. What I call art is only created by humans.

Call me a Luddite, but I believe only humans can actually create.

Posted on

Perfection

A freight train speeding by

For an artist, I believe perfection is a false goal. It can lead us to spend our energy in the wrong places. This seems especially true for photographers. Our technology-based art can lead us to believe technical perfection makes good art. It doesn’t. At least, not by itself.

An absolute

I am a recovering Engineer. I know a lot about specs and technical details and I am naturally drawn to “perfection”, whatever that is. As photographers, we tend to be pulled this direction. Are there any overexposed highlights? Do the shadows contain some information and very low noise? Did the lens and sensor resolve every bit of detail that could be used? Was “proper” technique used to maintain total sharpness and low noise? Did it follow the “rules” of composition?

More and more I am convinced these things are relatively unimportant compared to the impact of the image on the viewer.

Normal people view and enjoy prints at a distance of about 1 1/2 times the diagonal measure. Photographers tend to press their nose right against the print to try to see any imperfections. Yes, I do too at times. But this is not realistic or very important for normal viewers.

One of the themes I enjoy at times is images that have super high detail. Images about complexity and texture and the details of the material. I have good equipment and I am OK at using it, so I can do that whenever I want. Some subjects seem to lend themselves to it. But I don’t think I have any images that I consider great solely because of their technical perfection.

A moving goal

And what is “perfection”? Who defines it? Is such a thing achievable?

Our technology is constantly improving and pushing the boundaries back. The camera I use now is vastly better than the one I used 10 years ago. It has higher resolution, lower noise, and wonderful new features like live histogram view. These things let us achieve ever better results with our craft.

Likewise with printers, drop sizes get smaller, allowing for sharper prints, inks get better permanence, and printers get larger. Along with that software technologies improve all the time. We can upscale and sharpen images with much less artifacts. New algorithms can reduce noise without materially damaging sharpness. It’s a great time to be a photographer.

Is this perfection? Sure, a photographer can put his nose up to a large print and see “perfect” detail, low noise, great edge sharpness, and smooth tonal gradation. Is that what perfection is?

So is perfection the absence of any artifacts and a hyper-realism that looks sharper than real life? That is nice, for some images, but I do not believe it is what perfection is.

Why perfection?

Before I attempt to get in over my head, I will ask why we need perfection? Does it make better art?

I have seen prints by Ansel Adams, Edward Weston, Stiglitz, and many others. Many are stunning and have amazing presence, but not all are technically perfect. At least, by today’s standards. I have also seen many paintings by Monet, da Vinci, Rembrandt, PIcasso, O’Keeffe, etc. Again, I would say that the great ones may not be because of perfection in the sense I have been discussing it.

Craft trumps perfection

Take Ansel Adams for example. He shot mostly 8×10 negatives and spent many hours producing a print. But, the film technology he had was arguably not as good as modern high-end medium format sensors. And his lenses were not particularly good compared to modern designs. Some of his prints are not as technically “perfect” as many artists at their studio today making a print on their Epson or Canon printers.

But there is something else that overrides the technical limits. There is a magic in what he brought out in the printing process. He was a marvelous craftsman. He knew how to work an image until it changed from an average original to a stunning final print. “Moonrise, Hernandez” is a classic example of that. He shot it quickly because he was losing the light. So quickly that he couldn’t find his exposure meter, so he had to guess. Because the negative was badly exposed, it was very difficult to print. It required many hours of work in the darkroom to create a rendition of it. But it became one of his masterpieces. The final print is far superior to the original capture.

As he himself said

A photograph is not an accident it is a concept. It exists at, or before, the moment of exposure of the negative. From that moment on to the final print, the process is chiefly one of craft; the previsualized photograph is rendered in terms of the final print by a series of processes peculiar to the medium.

Ansel Adams

Adams was very good at all the aspects of photography. But in my opinion, it was his craftsmanship that made him rise above most everyone else. He would work a print until it glowed and had a life in it. The tones and contrasts and lighting were amazing. The results he created went far beyond considerations of technical perfection.

Story trumps perfection

I include story here because I believe it is powerful. But in general I struggle some with the concept. In a sense, story happens automatically. If you pause to examine a print for more than a couple of seconds it is natural to build a story. Humans naturally seek meaning and story. Guiding the viewer into seeing a more interesting story is a plus, both for the viewer and artist.

I do prints. Generally single images, meant to hang on a wall. To me, it is difficult to tell an extensive story with one isolated image. Not that it can’t be done, but I don’t see it happen as much as critics and some artists want us to believe. Probably the Engineer in me is still too literal.

But I see examples sometimes that make me wrong. A great one is Henri Cartier-Bresson’s “Place de l’Europe, Gare Saint Lazare, Paris“. Long name, but you’ve seen it:

Place de l’Europe, Gare Saint Lazare, Paris

I think this is a great story captured at, what Cartier-Bresson called “the decisive moment”. As you look at it, you tell yourself a story about what is happening, why he is doing this, what he is thinking, what happens next. It is still a memorable image today, even though it is under exposed, the subject is slightly blurry, and, depending on your notion of proper composition, the guy being about to disappear out of the right frame can be a problem.

It is a great and famous and beloved image. Being technically perfect would not have improved it at all.

Emotion trumps perfection

Emotion is something I have struggled with for a long time. I now believe that if I can’t make you feel something about my print, it is cold and sterile. I believe it so much that I feel that emotion far outweighs technical perfection. This is one reason I have been doing more ICM (intentional camera motion) projects lately. It throws out all notions of sharpness and detail and focuses mainly on capturing a feeling or impression.

Emotion in art has been written about a lot lately, but let me repeat and reinforce it. If I can’t make you feel something of what I felt when I made the image, I have probably failed.

There are techniques for creating an emotional response, but I am not concerned with them here. The fact is, we have to do it. As an artist, I have to share my feelings in an image or there is not much interest for the viewer.

Sharing and being transparent is a challenge for some of us (me). I am learning. The results are apparent to me. An image with a depth of feeling has more impact and staying power. Sigh. I will just have to force myself.

But the point is that an image that touches you emotionally is more meaningful than one that is just technically perfect.

Table stakes

So perfection, mostly technical perfection. Where does it fit? Am I saying it is not important? No. A technically perfect image may be excellent, but not just because it was perfect. It also has to embody emotion and story and excellent craft. Perfection is a table stake. It has to be there in order to get in the game. It is not the game itself. Art has to go beyond technical measures.

Today’s image

Earlier I mentioned ICM as a tactic I have been using occasionally to break away from the feeling I needed technical perfection. This image is an example of that. It is from a series I did called Speeding Trains. The intent is to capture the sight and feeling and power of a huge train speeding by. I hope you like it. Read the artist statement and see the rest on my web site.

Posted on

It’s Just Data

Smoky Ouray sunset, looking up the box canyon

Digital images are just data, pixels, digital values. Yes, but… That’s like saying paintings are just pigment smeared on canvas. It can become something more.

It’s data

Every digital image is data. I won’t go into film. It is the same but different. But a piece of exposed film is just data, too.

What comes off my sensor is a rectangular array of pixel values, red, green, and blue tuples. Tuple is just a mathematical term for a small set of numbers you keep together and in order. In this case (red value, green value, blue value). This is just numbers. Data.

When this data is brought into my computer it is still data. The manipulations I do on it in Lightroom and/or Photoshop are mathematical operations. Things computers are good at dealing with. An image may be gigabytes in size, but it is still nothing more than data.

Data just is. It doesn’t mean anything.

Interpreted by our minds

When the data is displayed on screen, I can view it and interpret it as something. This is the key. It means nothing until a human interprets it.

A particular set of contrasting tones and colors in a region looks like a tree to me. Even if the computer uses an AI classifier to identify it as a tree, that is just a meaningless label to it. The computer does not know what a tree is. An image of a tree cannot invoke memory or symbolism or meaning in the computer. It can in our minds.

So the data we see on screen, that is just variations of intensity and color, becomes meaningful to us because we are human. The data itself does not encode hope or despair or memories or associations or pain or beauty. That is what we make of the data.

The pen

There is an old expression that says “the pen is mightier than the sword.” This is true, but unpack it a little. A literal pen (do you remember what those are?) is not stronger or more forceful that a literal sword. The expression is metaphorical. The force of words conveyed to people’s minds can do more than the threats of swords can.

This is the case with images. The data making up the image means little. What we interpret from the data when we view it is everything.

I do not get political in this blog, but from a sociological interest, the protests going on in China (as I write this) are fascinating. Censorship is so strict that the symbol of the protests is a blank piece of paper, representing that they can’t say anything. From an engineering point of view, the amount of “data” in a blank sheet of paper is zero. It is the meaning ascribed to it that makes it powerful. An empty of paper can say volumes.

What elevates some?

Back on track, how is it that some data creates a far different effect than others? It’s much more than just the data. For example, here are 2 histograms. This is important data about the color information and distribution of pixels in each image.

Mona Lisa histogram Random flower image histogram

Their shapes are not that different. The one on the left has more warm dark tones and is darker overall. The one on the right has a lot of bright reds. Both have a red spike at the high end. But these are just 2 sets of data.

Would it make a difference if I told you the left histogram is the Mona Lisa and the right one is a random flower image from my reject bin?

So we cannot take an engineering view of the data to infer the value or user reaction. Our human perception makes all the difference.

More than data

The famous photographer Edward Weston once said: “This then: to photograph a rock, have it look like a rock, but be more than a rock.” It is a little bit of a stretch, but I don’t think it breaks badly to say when we press the shutter, we collect data, but it is more than data.

We could look at the data as an engineer and analyze histograms, tonal distributions, edges, area balances, and 100 other parameters. I know. I have. But my conclusion is that matters little. It gives us ways to describe the superficial data, but it says almost nothing about what the image means to us as humans.

So what?

I think we always have to ask “so what?” when we learn something new. Let me share 2 takeaways I get from this.

The first is that the data doesn’t care. I spent years trying to optimize the perfect histogram, ensuring total, crisp sharpness, capturing and preserving perfect color balance. At this point in my journey I will say that none of that really matters. All that matters is the effect I bring to myself and my viewers. Is it pleasing? Does it make us think? Is there a larger idea behind the surface scene?

The second takeaway is even harder for me to really grasp. It is just data, and the numbers don’t really matter. This means that what the original scene looked like (the captured data) should have little bearing on what we do with it. Process the data as much as necessary to create a great image. I need to stop being limited in my thinking by the reality I started with.

If it was an average, sunny scene but I feel it should be dark and moody, fine. If it was a colorful scene but I feel it should be presented in black & white, fine. Crop it. Add texture. The original data should not limit our artistic interpretation. This is one reason I often find it valuable to let images “age” a bit before I process them. I loose much of the association to the real scene and can take a more artistic view of the result I want.

Today’s image

I love this image and this place (Ouray Colorado). The sunset was almost blown out from a haze of wildfire smoke. Contrast was challenging. But I had to get something. It was too beautiful to stand idly by.

Besides B&W conversion and cropping it, the pixels have been bent quite a bit. The image is a couple of years old. I find that every time I go back to it I push it a little more to the extreme. Each time I do, it becomes a little more what I remember of the event. The less I remember the actual original scene, the more I feel free to make it match what I felt.

Posted on

How Big Can I Print It?

A VERY low resolution image (3 MPix) that would print surprisingly well

One of the things we have to wrestle with when we want to make a print is how big can I print this image and get good results? And how large should I print it? There is a lot of advice out there. Some of it is good.

Film vs. Digital

Virtually all images have to be scaled up for printing. The print you want to hang on your wall is many times larger than the sensor or piece of film you start from. Hardly any of us are shooting 8×10 negatives these days. Even if we are, we still usually want to make larger prints.

The technology has changed completely from the film days. Enlargement used to be optical. By adjusting the enlarger lens and the distance from the film carrier to the print surface, the image was blown up to the desired size. If the lens is good, it faithfully magnifies everything, including grain and defects. If the lens is cheap, it enlarges and introduces distortion and blurring.

Digital enlarging is a totally different process. A digital image is an array of pixels. My little printer at my studio likes to have 300 pixels/inch for optimum quality. So if I want to make an 8×10 print and I have at least 2400×3000 pixels, it will print at its best quality without changing a thing. Digital enlarging is a matter of changing the number of pixels.

Digital enlarging

But usually I want to print a larger size than the number of pixels I have. Here the digital technology gets interesting. And wonderful. Going back to my example, if I want to make a 16×20 print and maintain best quality, I would have to double the pixels in each dimension. It would have to go to 4800×6000 pixels.

Photoshop has the ability to scale the number of pixels in your image. There are several algorithms, but the default, just called “Automatic”, usually does a great job. Here is the difference from film: software algorithms are used to intelligently “stretch” the pixels, preserving detail as much as possible and keeping smooth transitions looking good. Lightroom Classic has similar scaling for making a print, but it is automatically applied behind the scenes. Smoke and mirrors.

The result is the ability to scale the image larger with good quality.

Print technology

In a recent article I discussed a little of how an inkjet printer makes great looking prints using discrete dots of ink. There are other technologies, such as dye sublimation or laser writing on photosensitive paper, but they are far less used these days.

It should be obvious, but to make a really big print, you need a really big printer, at least in the short dimension of the print. Really big printers are really expensive and tricky to set up and use. That is why most of us send large prints out to a business that does this professionally.

Why do I say the printer has to be big in the short dimension of the print? Past a certain size, most prints are done on roll feed printers. They have a large roll of paper in them. Say you have a printer that prints 44″ wide. The roll of paper is 44 inches wide and many feet long.

We want to take our same 8×10 aspect ratio image and make a 44×55 inch print. If it was film, we would require an enlarger with at least a 44×55 inch bed and a cut sheet of paper that is 44×55 inch. But an inkjet printer prints a narrow strip at a time across the paper. The heads move across and print a narrow 44 inch long strip of the image, the printer moves the paper a little bit, and it prints another narrow strip. Continuing until it has printed the entire 55 inch length. Then the printer automatically cuts off the print.

But if we naively follow the recommendations for optimum quality, we have to scale our poor little 2400×3000 pixel image up to 13200×16500 pixels. Even the best software algorithms may introduce objectionable artifacts at that magnification.

Viewing distance

Maybe we don’t have to blindly scale everything to 300 (or 360) pixels/inch.

A key question is: at what distance will the image be viewed? Years of studies and observation led to the conclusion that people are most comfortable viewing an image at about 1.5 to 2 times the image diagonal length. This lets the natural angle of the human eye take in the whole image easily. For the example we have been using of the very large print, people would naturally choose to view it from about 105 to 140 inches.

Along with the natural viewing distance there is the acuity of the human eye. I won’t get into detail, but the eye can resolve detail at about 1 arc minute of resolution (0.000290888 radians for the nerds). Simply, the further away something is, the less detail we can see.

Going through the calculations, if our audience is viewing the large print from 1.5 times the diagonal, it only has to be printed at 33 ppi! Finer detail than that cannot be seen from that viewing distance.

I have heard photographers who have images printed for billboards or the sides of a large building talk about inches/pixel. It would look like Lego blocks up close, but it looks sharp from where the viewer is.

Nature of the image

This is true unless the audience is photographers. They are going to get right up to the print, as close as their nose will allow, to see every blemish and defect. 🙂 But normal humans will view it from a distance.

There are modifications to the pixels vs. viewing distance calculations depending on the nature of the image. If the image contains highly detailed structure it will encourage viewers to come closer to examine it. If the image is very low contrast, smooth gradations, it could be even lower resolution.

Printing at the highest possible resolution that you can for the data you have is always a good idea.

Your mileage may vary

How big of a print can you make? It depends – don’t you get tired of hearing that? It is true, though. The real world is messy and simplistic “hacks” often don’t work well. It is better to understand things and know how to make a decision.

When it comes down to it, these are great times for making prints, even large ones. My normal print service lists prints as large as 54×108 inches on their price list. I know even larger ones are possible.

How big should you print? How big can you print?

Conventional wisdom is that scaling the pixels 2x each dimension should usually be safe. My camera’s native size is 8256×5504 pixels. Scaling an image 2x would be 16512×11008 pixels. This would be a “perfect” quality print of 55×36 inches on a Canon printer. I have yet to need to print larger than that.

Given the perceptive effects of visual acuity, I am confident I could create much larger prints. Larger than is even possible by current printers. And they would look good at a reasonable distance.

A key question is who are you printing for? A photographer or engineer will be right up to the print with a magnifying glass looking at each pixel. Most reasonable people will want to stand back at a comfortable distance and appreciate the image as a whole. Who is your audience?

Learn how to scale your image without artifacts and how to use print shapening to correct for problems. Know the perceptual effects of human visual acuity. This is part of the craftsmanship we have to learn in our trade. Given those tools, the rest is artistic judgment. With today’s equipment and careful technique and craftsmanship we can create wonderful results.

Your mileage may vary.

The image with this article is very small – 3 MPix. I would not have a problem making a 13×19 print of it. I doubt you could see the pixels.

Have you tried to make large prints? How did it go? Let me know!