Late to the party for Shepard Fairey

The Blognosticator in Munich

It gets late early here in Munich in winter, and sometimes I find myself being ready for bed only to discover that it’s 8:15. So, I have taken to watching documentaries on Hulu (and elsewhere) or refining that day’s effort on the New York Times Spelling Bee puzzle.

I was watching the Hulu Original documentary about Shepard Fairey a few nights back, and found it very compelling. I have been an subconscious admirer of Mr. Fairey’s work for years, though unlike the rest of the world, I became aware of the artist’s name only after his Barrack Obama poster became so well known.

This is the SWM building in Munich that features a wonderful mural
by street artist Shepard Fairey.

I was also aware of the lawsuit filed by Associated Press claiming copyright infringement, a suit that the cooperative spent almost $15 million prosecuting (according to the documentary). At the time of the suit I sided with AP, because I saw the artwork as being an almost-literal copy of the photo taken by Mannie Garcia, who is a contributing photographer for the Associated Press.

If you watch the documentary – and I recommend that you do – you will learn that Shepard Fairey didn’t just “paint-by-numbers” using software, as was argued in AP’s lawsuit. Mr. Fairey used Rubylith* film to hand-cut the separations for his poster, working from an enlarged version of the photo. It was, by reason of artistic merit, a derivative work.

That phrase “derivative work” is vaguely described in the most recent U.S. Copyright law, excepting artists for works that are “derived” from the work of others from infringement claims. After watching the film, I flipped to Fairey’s side of the argument.

In the film there is a touching scene where Mannie Garcia, the photographer, talks about going to lunch with Shepard Fairey during a break in depositions. To the horror of all the lawyers present, the two went off together and shook hands after discussing the photo and Fairey’s interpretation of it.

It’s important to note that Shepard Fairey is sincerely repentant about not licensing the photo, which he should have done, and which would have obviated the need for the action in the first place.

AP and Fairey settled their suit in 2011 on undisclosed terms.

I have long admired, but misunderstood Shepard Fairey’s work. I really like his Soviet-style constructivist art (often modeled after actual Soviet-era artwork), and I love his political posters, especially those he created after the election of Donald Trump in 2016. To see tens of thousands of Fairey’s posters being carried by the millions of people who attended protest rallies after the inauguration in 2017 was moving.

His style is unique. His skills are extraordinary (he is a graduate of RISD), and his ability to use images to move crowds and change minds is really impressive. His work has become timeless.

Sure, he has been arrested on numerous occasions (who hasn’t?**) and certainly he is guilty of putting stickers on all sorts of municipal properties in his career. All of that I have forgiven because his artwork transcends graffiti and vandalism.

I am just now finishing a project with my students about street art in Munich. As part this, I have experienced a conversion that lets me to see street art as something entirely different from graffiti or “tagging.”

While on a tram last week traveling to a local museum, my wife and I saw an amazing work of street art: a building-size paint can with an environmentalist message. I made note of the cross-streets and returned the next day to photograph that art. It covers an entire side of a building in Munich, one that houses the workshops of SWM, the local public utility. It is a dazzling work of art, and I wondered who did it (I should have noticed the style and the similarities to his other works).

It was that evening that I watched the Shepard Fairey documentary, and saw in one scene, Mr. Fairey painting the wall in Munich. A version of that same image was used at the Paris Climate Conference on a sphere suspended under the Eiffel Tower. Wow! It is Shepard Fairey’s work, sponsored by a Munich arts organization called Positive-Propaganda.

That work is complemented by an electric vehicle charging station (appropriate because the message of Fairey’s art is anti-Big-Oil).

So, call me a late-to-the-party Shepard Fairey fan. I have now seen so many fine examples of his work to call him an inspiration to fine art around the world. I’m sorry that I didn’t become an admirer sooner. I apologize for my tardiness.

* Rubylith is a red (or amber) colored gelatine film that is used to prepare masks for photographic and plate-making activities in the graphic arts. You cut through the gelatine layer and peel-away the parts that you want to be transparent, leaving the rest behind to create a photo-opaque mask.

** I have never been arrested, though I was escorted and questioned once by military police after taking a panoramic photo on military property. They let me go without charging me, because the boundary of the property was not clearly marked.

Addendum 9 January, 2022:
Another gorgeous mural created with a funds from Positive Propaganda can be found on Dachauerstraße, near the corner of Lothstraße in Munich. This was painted in 2018 by an artist named Liqen.

This extraordinary work, entitled Exhuman, is by artist Liqen, and sponsored by Positive Propaganda. It was painted in 2018.
Posted in Art, Graphic Arts people, People, Photography, Printing and Printing Processes | Tagged , , , , , , , , , , | Leave a comment

A visit to FOGRA

The Blognosticator in Munich

A few days before Christmas I rode a subway to a regional train to another subway, then walked a bit to reach the headquarters of FOGRA, the Research Institute for Media Technologies in the town of Aschheim, just outside Munich.

FOGRA’s office in Aschheim, Germany. In the adjacent office is the German Printing Industry Association.

This for me was like a Springsteen fan walking on E Street, but the street in Aschheim is named after Einstein – fitting for the location of this organization.

My purpose in walking on Ensteinstraße that day was to make some measurements using FOGRA’s spectrophotometer. The similar device at my university in Munich will not work with the most recent version of the profiling software. I needed a more modern tool for this work, and FOGRA welcomed me to use theirs.

Berthold Oberhollenzer of FOGRA welcomed me at the entrance and took me up to the measurement lab where I rolled out my press sheet and cut the test panel out of it to put on the bed of the X-Rite i1IO 3 table.

The press sheet I had created was printed on a Landa Nanopress, the first in Europe, also located in Munich. The owner of that press, Blueprint, ran the test sheet for me to give me an opportunity to test the resolution, register, and color qualities of the Nanopress.

The X-Rite i1IO 3 instrument at FOGRA

I have made a lot of test sheets in my career, usually with color targets, with which to measure the behavior of various offset presses, wide-format ink-jet printers, electrophotographic printers and paper combinations. This was my first full size digital printing press (the Nanopress prints on B1 sheets – 1,000 x 700 mm or 39.37 x 27.55 in.). It is the same size as a Heidelberg 102.

The press has a native resolution of 1,200 ppi (472.4 px/cm). To test that, I used several line art scans I have made at various resolutions: 600, 800, 1,200 and 2,400 ppi. I also made straight and diagonal line sets at various thicknesses from 0.009 pt. (0.00317 mm) to 2 pt. (0.705 mm) and Helvetica Thin type from 1 pt. to 7 pt. – both black-on-white and white-on-black.

Landa’s Nanopress prints with seven colors of ink: CMYK plus orange, blue and green. This is one of the largest color gamuts of any printing press. And, since it uses ink-jet for printing, and because the ink is dry as soon as it touches the paper, there is little to contaminate the colors or reduce the glorious color possibilities of this machine.

This is my test sheet for the Landa Nanopress (before printing).

I wanted to measure that color space, so I put an RGB color target called an IT9.18 target on the page. The colors in this patch set are defined with RGB values, so I thought that, unhindered, I would get an expanded color gamut on my press sheet from the RGB target, one that would show off the three extra colors on the machine.

And, that is what brought me to FOGRA’s offices outside Munich.

A month earlier, I had picked up some press samples from Blueprint, and was dazzled by the colors on those press sheets. Glorious greens, zesty oranges, and boisterous blues were on the pages that I brought home from my first visit to the plant.

I wanted that same color gamut for my current project, a book about Munich street art, and the seven colors of the Nanopress were calling! I put an RGB image of a colorful hot-air balloon on the press sheet in five variations. Opening it from Camera Raw, I made one version of the image in Apple RGB, sRGB, Adobe RGB, and ProPhoto RGB (increasingly larger color gamuts). I expected the colors of the ProPhoto version to jump off the page because the color gamut of the Nanopress would closely match that of the image.

What happened is interesting: the sRGB photo looks “brightest” while the Adobe RGB image has the richest gamut. The ProPhoto photo turned a hideous shade of purple; it looks awful.

And, after reading the IT8.19 RGB target as interpreted by the Nanopress (the colors are actually processed by a version of the EFI Fiery RIP at the front-end of that press), the profile that I created is almost exactly the same shape and volume as sRGB.

I obviously did something wrong.

I tested rich black (20, 20, 20, 100) with white type reversed-out; that worked perfectly! There is almost no possibility of register error on the Nanopress because the ink is deposited onto a belt, then transferred to the paper as a complete image. Interestingly, the density of single-color 100 percent black is excellent – much denser than a similar ink on an offset press. I question whether rich black is even needed when printing to the Landa press.

I put a rainbow gradient in RGB across part of the sheet, and that failed in the range of oranges and reds on the left side. I don’t understand what happened there. Conversion of red, green and blue type looks great, and a photo of an orange pick-up truck reproduced beautifully, as did an image of a spiral Aloe plant that has some almost-alien greens in it.

My overall impression of the Nanopress is that it has incredible potential. The color gamut is huge (though I am not sure how to measure it), and the technical qualities of the press are extraordinary.

I will report more about the Nanopress in another blog coming soon.

Posted in Color Management, Printing and Printing Processes, Software | Tagged , , , , , , , , | 3 Comments

It was a dark and stormy night…

The Blognosticator in Munich

Two of my students and I were searching for Daniel, one of the important people in a community of artists in Munich. I say he is “one of the important people,” but I cannot tell you his title, nor how many other important people there might be. That information will come later, when I read the interview transcript.

Looking a bit like a sketchy meeting under the railroad tracks, my students conducted an interview for a book we are publishing. Daniel, center, describes his arts community Bahnwärter Thiel, in Munich. On the left is Ana Margarida, and on the right is Sebastian, two of my Master’s degree students. This is the uncorrected photo.

It was getting dark, it was snowing, and it was cold. Sebastian, my student interviewer, had tried calling Daniel a few times, and Daniel had tried calling back, but two-way communication had not yet happened. The other student, Ana, and I were shivering.

Finally Sebastian’s phone rang and the two had a conversation. Daniel was across town, across the river. Could we postpone until next week? No. Our deadline is very soon, and we couldn’t wait a week. Could we get on a bus and meet him across town? Yes, we could!

When we stepped off the bus Daniel found us. We then had two choices for interview location: inside a warm bar/club (requiring a rapid antigen test and a 20-minute wait) or under the railroad trestle with high-speed trains whizzing by every few minutes. We chose the railroad trestle (that isn’t really the word for a concrete structure that supports several train tracks overhead).

…and this is the corrected version. Lots of shadow enhancement, color temperature adjustment, and 60 units of Noise Reduction make the photo perfectly acceptable for our project. I am amazed that the amount of light in the scene allowed for this photo to be enhanced to this degree.

I had my Canon R camera with me, but it was too dark for photography. Our séance looked like a clandestine transaction, but it was an interview where Sebastian asked questions of Daniel, and he responded with animated, detailed answers. Some people are great interview subjects, and he was one of those people. We wound him up and let him go. He talked for an hour.

While the interview was continuing, I took a few photos with my iPhone 12 Pro. These cameras are well known for low-light photography. But this situation stumped my iPhone’s cameras and software. The photos just didn’t work.

So I decided to give it a try with the Canon. I turned the ISO to the highest setting: 32,000. That, combined with my f2.8 lens, and I had an exposure that was at least possible: 1/13 sec. Our subject moved his hands and arms a lot, but I was still able to get a surprising number of good photos.

Then there is the noise. At ISO 32,000, the noise is visible. Sometimes, though, you use the photo that you have rather than the one you wish you had.

In Adobe Camera Raw I lowered the color temperature about 200 degrees, then I added a small amount of sharpening. I suppressed the highlights and emphasized the shadow detail quite a bit, opening up the dark areas at the bottom of the photos.

Then I added 60 units of Noise Reduction and 60 units of Color Noise Reduction. The two combined to give me a few very acceptable photos. Once again, I am impressed by the combination of the camera and Camera Raw to deliver usable photos in situations where I would never have thought it possible.

It wasn’t a low-light situation; it was a no-light situation. I’m dazzled that I got anything usable at all. When printed, these photos will be a quarter-page or less, so I don’t think I will have any visible noise in the final product.

Posted in Adventures, Art, Color Management, People, Photography, Photoshop techniques | Tagged , , , , , , | Leave a comment

An analysis of the Image P3 color space

The Blognosticator in Munich

About four years ago, Apple changed the color space of the iPhone from whatever it was (maybe sRGB?) to a new color space called Image P3. This is the color space used by flat panel TV screens.

This was a smart idea on Apple’s part because the capabilities of the newer devices – iPhone, iPad and various Macs – are so much greater than sRGB and other small gamut color spaces. The potential of the Retina Displays is immense and Apple responded by making the official color space on these devices much, much larger with P3.

This is the 2-dimensional plot of the gamuts of three common ICC profiles: ProPhoto RGB, Adobe RGB, and the newer Image P3 color space that is now default on iPhones and iPads.

But in all my studies and classes I have never taken the time to compare Image P3 to other big gamut color spaces. I have now done that, and I am impressed.

In 2019 I made a presentation at the annual Color Conference in San Diego on fine art reproduction. As a part of that, I compared the color space of my Canon digital camera and studio strobe lamps to the Adobe RGB color space, and found that it (ARGB) didn’t cover some of the critical colors in the painting I was trying to reproduce. By changing to the ProPhoto RGB color space I was able to capture those elusive colors, and the resulting print was much more accurate.

I’m teaching in Munich now, and the subject came up in class this week.

Using the latest version of ColorThink Pro, I opened three profiles: Image P3, Adobe RGB and ProPhoto RGB. I graphed them in both 3D and 2D to get a sense of both the size and the volume of these profiles.

ProPhoto is by far the largest, and Adobe RGB and Image P3 are close competitors. Image P3 captures measurably more color along the red-green axis than Adobe RGB, the same area where I gained ground using ProPhoto RGB. These are important colors: peach, yellow, orange, and many images will benefit from having these colors within P3’s gamut.

Adobe RGB has greater area on the green axis, and has brighter reds and magentas, while Image P3 shows greater bright color volumes on the green end of the green-red axis.

So, is one of these color spaces “better” than the others?

Here, the 3-dimensional view shows how the Image P3 gamut compares with its closest counterpart: Adobe RGB. They each have strengths and (comparative) weaknesses. Adobe RGB has a slightly greater area overall, and the two have similar total color volumes.

ProPhoto has more volume and more area – more total colors and more brightness range than either Image P3 or Adobe RGB. This is probably most valuable for photographers and designers who are going to print on wide-gamut ink-jet printers and presses.

Adobe RGB is likely the most popular color space among professional photographers now, but for those preparing images for delivery on flat-panel TV displays, iPhones, Androids and iPads (especially the newer models with Retina Displays), P3 is a better choice.

Planning the color as you open it from a digital image (the color profile is assigned as a Camera Raw file is opened) will help to match the image to its destination to take advantage of the full range of colors and brightness available on different delivery systems – print or electronic.

Note: I used ColorThink Pro to make these charts. That software, from Chromix, is the best product for looking at color images and profiles in mathematical space for purposes of comparison and problem-solving.

Posted in Color Management, Photography, Printing and Printing Processes, Software | Tagged , , , , , , , | Leave a comment

A grateful message to all of you

The Blognosticator in Munich

This is my occasional update on traffic to The Blognosticator with statistics and commentary on my art of blogging.

…as of 26 October 2021. Thank you!

In an era when podcasters are doing fabulously, I have to face facts: I am a print guy. I tried video, and I did pretty well with it, but I don’t think in moving terms, instead preferring still photos and words that stay fixed on a page. I guess that describes my blog.

I started writing blogs here ten years ago in July. Prior to that I had a regular assignment with Graphic Arts Monthly magazine to write a minimum of four blogs every month. That publication went out of business in 2010, and they were kind enough to return the copyright of my work to me with no restrictions. I was free to republish any or all of my work there.

I did repurpose a few of those early blogs, but I found that most had lost their timeliness.

So I started writing fresh material on July 26, 2011 here.

In the ten years since then I have written 289 blog posts, and have received 589 comments on the posts.

In addition to the legitimate comments, I have received over 900,000 spam comments – attempts by spammers to infiltrate and hack my blog. These spam comments have been successfully filtered out by software called Akismet. I pay $100 per year for their service, and it is flawless. It searches all incoming comments and quarantines/deletes anything suspicious.

I take breaks from my blog, something experts say I should never do, but I am busier now than I have been in recent months and years, so I don’t sit down to write blogs very often.

Despite those breaks, I keep writing, and over 400,000 visitors have come to The Blognosticator to read my musings about printing, photography, word puzzles, cameras, graffiti, and more.

I am humbled by your continued support. Thank you.

Posted in Business | Tagged , , | Leave a comment

An obsession with words

The Blognosticator in Munich

Though it has nothing in common with my usual posts, I want to say a few words about my obsession with the New York Times’ Spelling Bee puzzle.

This is the format of the Spelling Bee puzzle in the New York Times.

It’s fun! It’s cause for insanity. It’s enough to make you either a genius or a maniac, or both.

I play it every day, and for the past year I have been able to get to Genius level every day. Sometimes that takes me ten minutes, other times it takes me, on and off, all day and late into the night to reach that goal.

And, three times in the last year I have made it to Queen Bee status, where I have gotten all of the possible words in that day’s puzzle. This morning I made Queen Bee status with the word “inbox.” I was thrilled (and I am being quite modest about it).

In the approximately half year that I trained to be a genius every day, I occasionally cheated by using a web site called There, you can put in the seven letters (I put them in twice each) and it will tell you all the words possible with those letters. Some of its words are not allowed by the New York Times for various reasons, but it will often get you out of a jam.

One year ago I got good enough at the puzzle that I stopped cheating, and pledged never to use WordHelp again. I have been true to that pledge, and have not been back.

Now I rely entirely on my wit, my memory, and luck to get to Genius every day. I also use the built-in dictionary on my iPad to check words as I enter them (this isn’t cheating because it is only verifying that a word is spelled correctly, and it is available inside the Spelling Bee puzzle itself). If I touch and hold, the word will be highlighted and an option to “look up” arrives on-screen. I will occasionally look up words in the solution list because I have never seen them before and I want to know what they mean. I used it for fettuccini recently.

A few months ago, a man in New York made a web site for people like me who are obsessed with the Spelling Bee puzzle. It’s called I like it, and I look occasionally to see how many words it will take to get to Queen Bee status. I also check to see words that are in Webster’s, but not usable in the puzzle that day. There are a lot of these words. Yesterday I wasn’t able to use “nonunion” even though it is a perfectly valid word in the dictionary.

In you can also cheat and see the words for that day’s puzzle. I never do this. I would rather wait until the next day to see the words I missed. My puzzle day begins at exactly 9:00 a.m. European Central Time when the Times posts the new day’s puzzle. I am often waiting at that minute to check my words against the official list, and to start work on the next day’s challenge.

I have developed the skill of Genius-making on the Spelling Bee, and I thought I would share my techniques with those of you who are obsessed with Spelling Bee, or might become obsessed.

Every day there is at least one pangram, defined as a word that uses all the available letters. Sometimes there is more than one pangram; those are great days.

I treat the pangram as my first stab at getting a good score. I stare at the letters and play anagram games in my head, trying to make words, or prefixes and suffixes. Sometimes the pangram will jump into my head, while other times it takes much longer. I usually stick with it until I get it, and only then move on to the easy words.

The rules are simple: you must use the center letter in every word; you may use any letter
one or more times; each word you spell must be in the official list of words for that day.
There is at least one “pangram” in each day’s puzzle, that being a word that uses all seven letters.

If you get a pangram, you get lots of points for the word.

Four-letter words get 1 point each; more letters earn 1 point more per letter; pangrams get 7 extra points in addition to their word score.

Proper nouns are not allowed, nor are obscene words. This is a carefully curated puzzle.

I bang-out the obvious words first, getting as many points on the list as possible. Then I try adding prefixes like “un” or “in” or “en” or “re” and I read down my list of words trying to add the prefix if I can. “Noble” might become “ennoble” or “done” could be turned into “undone.” When you have exhausted the prefixes, look for suffixes, adding “ed” or “ing” to every word you have made so far. In the process you will find words that you weren’t looking for. Some days you can nearly double your score by adding “ing” to most of the words on your list. Those days I call gerundific days.

When you have reached Amazing level, and it seems that you cannot find any more words, look for opportunities to use:

  • Pasta: rotelli, tortoni, fettuccini* etc.
  • Other food words: blini, falafel, tortilla, tapa, naan, bialy, focaccia, onion**
  • Musical terms: trio, quartet, nonet, octet, tutti, largo, rondo, duet
  • Geometry terms: nonagon, octagon, decagon, dodecagon, septagon, hexagon, et ceteragon
  • Body parts and medical terms: ulna, alveoli, atria, and related plurals
  • Common chemical and biological terms: anionic, ionic, amoeba, oleic, niacin
  • Japanese gates: torii (who knows why, but it shows up often)
  • Japanese cushions: tatami

*Note that “fettuccini” was only allowed with this spelling, and not the other five options possible in Italian and other languages.

**If you start participating in the Spelling Bee, you will use “onion” several times a week. Naan is another regular in the Spelling Bee.

Most of the time these categories, and the words that you will memorize over time, will help you succeed every day. You’ll also start to add words that are used in the puzzle regularly. Eventually you’ll get to the Genius level, and say, “What’s next?”

This is the screen that pops up
when you earn Queen Bee status.
It would be nicer if the word “All” were on the same line as the words in the next line, and the word “worth”
found its way down to the next line also
(the typographer’s comments).

What’s next is Queen Bee, where you get every possible word in that day’s puzzle. To reach this level you should try on days when Genius is achieved with fewer than 30 words. The difference between Genius and Queen Bee is daunting, and with more words, there are more needed to get to Queen Bee.

Yesterday I got all 27 words to reach Queen Bee. To reach Genius only required (for me) 24 words, so I had to find only three more, and that was comparably easy. On other days, when there are 35 or 45 words to reach Genius, the distance to Queen Bee is immense – sometimes 10 or more words. And, where on Earth am I going to find ten more words?

The satisfaction of reaching Queen Bee status is incredible. I walk down the street with a big grin on my face (missed entirely by the neighbors who pass me).

I hope that these tips will help you to reach Queen Bee level soon so that we can celebrate together, or at least you can smile as you pass me on the sidewalk.

Note: Playing the Spelling Bee requires access to the New York Times online. You can get a regular subscription, or there is a special subscription that gives you access only to the puzzles. This includes the famous Times crossword puzzles, Spelling Bee, and several others. That subscription is just a few dollars per month.

Posted in Language and grammar | Tagged , , , , , , , | Leave a comment

A refined approach
to repositioned panoramas of Munich street art

The Blognosticator in Munich

Four years ago I was teaching in Munich, and during that time I photographed a number of pieces of street art on Tumblingerstraße in the city. You can read that story and its associated comments here.

The photos were taken in the urban landscape, on a street with cars parked along the curb, making it impossible to step back far enough to take photos of the works in one frame. Also, single photos are not high enough in resolution for my projects. I need hundreds of megabytes in order to print on wide-format ink-jet printers or large pages on offset presses.

This is one of several test images I created using the technique described in this article.
I like the way the real motorcycle and scooter blend into the mural as if painted there.

In response to this need for high resolution, I used a technique I call repositioned panoramas where the subject is photographed as a series of images taken horizontally at fixed intervals, then stitched together in Adobe Photoshop using its Photomerge function.

I have used this technique to scan strips of historic photographic film, creating multi-gigabyte images. The technique works really well.

In the field, it’s slightly more challenging to shoot “panoramic” images if you move the tripod between shots. In fact, if you are shooting anything with visible perspective, it makes images that are impossible to stitch.

My street art photos pose very little perspective trouble because they are painted on walls, and those walls have almost no relief, therefore not much chance of perspective shifts. In 2017 I used a technique where I took a shot, then took three careful steps to the right, then took another shot, and continued until I had a stepped record of the scene.

It worked well, and the images I made were excellent. They exhibited some small flaws, but I attribute these to my handling of the camera with no tripod.

Another example of the street art on Tumblingerstraße in Munich.
The works are fluid, and are often replaced by new works put up by other artists.

I am back in Munich now, teaching at Hochschule München, and on my second weekend here I ventured back to Tumblingerstraße to see how it has changed. All of the artwork is new – it is constantly refreshed with new art – and the location has expanded considerably. Now there is an entire village on Tumblingerstraße dedicated to street art and related art. There are buildings inside the wall with more art, there is a café, and it appears that people live and work in the village (I’m still learning about the village).

My students tell me that there is an occasional concert there, and on Sunday last week there was a flea market there, open to the public. People were lined up to buy fleas.

To make my repositioned panoramic images better I knew I had to be more accurate in taking the source photos, and for that I needed a team of photographic helpers.

As part of the course I am teaching, I have the students documenting Tumblingerstraße, interviewing the artists, and taking a large number of photos. These will become part of a book we are writing and producing as part of the class. One team worked on taking new repositioned panoramic images of the artwork.

Our new technique relies on using a tripod, using a higher-resolution camera, and taping a long piece of string to the sidewalk, marked with increments for the camera (we tied knots in the string every meter). We set up the camera, put it on manual focus and manual exposure, then locked the camera facing the artwork. We then hung a plum-bob from the center column of the tripod, and moved the tripod and its centering tool from one knot to the next for the photos.

This is one of over 100 knots we tied in the string to mark the increments for photography.

The string was taped to the ground at three meters’ distance from the wall, so the subject-to-camera distance was constant. This meant that we were photographing the artworks with much more precision than I had when I first did this. We took two photos in each position – just in case – and we moved a total of 130 meters along the wall to capture all of the images.

Two of my students positioning the string three meters from the wall.
This step was critical to getting the shots to be consistent.

The images on this wall are not level, nor is the wall. It goes downhill, and so do the images on it. We set the camera parallel to the base of the wall to photograph each image as straight to the wall, regardless of the slope of the wall. As a result, the slope is removed from the images.

By eliminating almost all opportunity for error, we made the exposures and then processed them through Adobe Bridge to convert them to DNG* as we downloaded them from the camera card. Then we selected a series representing one piece of work and opened them in Adobe Photoshop.

From there, we chose Automate from Photoshop’s File menu, then Photomerge. In Photomerge, we selected the open images, chose Reposition as the technique for the merge, and let Photoshop do its work. Running on a new MacBook Air with the M1 processor, each combined image too about one minute to process. The resulting images are about 250 MB each, and show no significant errors of stitching.

These will make excellent photos for inclusion in our book, which we plan to print with long, wide pages to accommodate the wide stitched panoramic images we made at Tumblingerstraße.

*You can read my essay about converting Raw files to DNG here.

Posted in Adventures, Art, Panoramic Photography, Photography, Photoshop techniques | Tagged , , , , , , | Leave a comment

AppleScript solves my grid mask challenge in Adobe Photoshop

I pride myself on being a competent user of Adobe Photoshop. I know how to make my photos look good by sliding sliders and pushing buttons and applying filters and adding Layer Effects and all sorts of visual things.

Occasionally I need to do something out of the ordinary, and in these situations I combine my skills in Photoshop with my skills in programming with AppleScript. I taught myself AppleScript back in the 1980s when I had insomnia. I simply couldn’t get to sleep at night. So, instead of counting sheep, I wrote programs with AppleScript.

This is the collage of images I wanted to make in Adobe Photoshop. I used an AppleScript to make a checkerboard mask for the photos, then used that mask (and an inverse of it) to mask each photo separately. This made it possible to create an image without a single pixel out of place.

For the uninitiated, AppleScript is a scripting language on the Macintosh computer. It is a full-bodied computer language that can send and receive instructions to and from various applications. You can, for example, query a FileMaker database, copy a line of information from that database, then move that copy to Adobe InDesign and paste the text into a document, while applying a specific type font, size, leading and color. It’s a very clever language.

The greatest advantage of AppleScript is its “natural language” quality. It has English-like commands (other languages are also supported). You can issue a command like:

tell application “Photoshop”

…and the script will open Photoshop. Many applications have sophisticated AppleScript-ability. Some have none, so you cannot control them with AppleScript.

AppleScript wasn’t my first programming language. I learned COBOL in college, had learned a language called WPL on the Apple II, and had taught myself BASIC and some Pascal, all of which I found intriguing. I believe that once you learn any programming language you can use your skills to learn another. This is because programming languages have so much in common. It’s mostly form and syntax beyond the basic control functions. You just have to learn the localized grammar, and then figure out the details.

Yesterday I was working on a photo mural and I needed a set of absolutely precise masks to make a checkerboard collage of images. I started out with 24 selected images from a project I did in 2016 called the Bishop Peak Portrait Project. That project was a year-long photographic study of a mountain in San Luis Obispo. I took about 200 photos every day with a camera in a weatherproof box on the roof of the Kennedy Library on the Cal Poly campus.

The camera was trained on Bishop Peak, the most popular of the nine mountains that run from my city out to the ocean at Morro Bay, where Morro Rock is the second-to-last mountain in the chain. These glorious peaks were once part of the rim of a huge volcano that formed this part of California. We got the west half of the rim; the east half is located about 100 miles south of us in another part of the state.

My project was supported by a grant from Cal Poly’s College of Science and Mathematics. I started in March, 2016, and ended on the last day of February, 2017. Between those dates I took about 70,000 photos of my favorite mountain.

The camera was a Canon T5 with a standard lens zoomed to about 50mm. The camera was powered by two 12 volt motorcycle batteries and they were kept charged by a couple of small solar panels. The camera was controlled by a Raspberry Pi microcomputer on a circuit board of my own design. It had two small power supplies and a transistor circuit that triggered the camera when the Raspberry Pi signaled for an exposure. I had it set to take one photo every five minutes from 5:00 a.m. until about 10:00 p.m.

I had a 256 GB SD card in the camera which allowed me to leave the camera unattended for weeks at a time. I would occasionally climb the six floors to the roof of the building and retrieve the camera card, replacing it with another.

Then I would discard the nighttime shots – many of my photos were black rectangles – then rename the files in Adobe Bridge (I used a six-digit numbering scheme). Then I scored the photos in Adobe Bridge, giving them four stars if I liked them, and five stars if I loved them.

Eventually I had to pick only one photo to represent the best image taken each day for 365 days. It was surprisingly difficult because there were so many to choose from.

The end product is a mural on permanent display at Cal Poly of the 365 photos, each printed on a 5 x 5 inch square of aluminum, mounted on a 22-foot panel in calendar order. It turned out really well.

This year I was approached by a representative of a local law firm to provide some images for their new office. It was suggested that I could pick some of my Bishop Peak photos to make a photo collage for a wall in their foyer.

I looked through my photos and picked about 30 of my favorites. Then I opened them (they are all DNG raw images) in Camera Raw, made adjustments for contrast and vibrance, and converted them into Photoshop images. Each is about 12 x 9 inches in size at 300 ppi.

I wanted to make a grid of photos six across and four down for the collage. The total size of the final image is about six feet wide by three feet tall. The image will be printed and mounted on stiff Gator Foam board, then attached to the wall with a wooden cleat.

I made a layout in Photoshop with blue guidelines separating the master image into a grid, then I started bringing in the images and arranging them in an attractive layout.

After the second and third images I realized that it is very difficult to get the edges of these photos to align with pixel-precision and not have any gaps or corners that show errors of position or size. I needed a mask!

So, I drew one, and then I copied it and propagated it across my canvas. But it was slightly off, and the corners didn’t work perfectly, and at the end of the first row I had a handful of pixels that ran off the end. I needed more precision.

I decided to draw the masks using AppleScript. To do this I made a template in Illustrator and put the coordinates in pixels onto the template.

This is my annotated grid for the mask script. Each pair of numbers is the X,Y position for that corner
of each mask in pixels. The entire canvas is 21,516 X 9964 pixels in size.

Photoshop can be programmed with Visual Basic on Windows machines, with Javascript on Mac and Windows, and with AppleScript on Mac. I have considerable experience with the AppleScript version, and I approached it by opening a script I had written a few years ago when I was making the 365-photo mural for Cal Poly. That script uses pixel coordinates to make a block of black on which I put the date and time of each photo. I saved a new copy of that script, stripped it of its text components, and kept the rectangles part for my masking adventure.

The idea was to make solid black rectangles of exactly the correct size for one photo, followed by an equal amount of white space, followed by another black rectangle until it filled the canvas with a checkerboard of solids and white spaces. This would be the primary mask. I created a canvas in Photoshop and had it open, then I tested my script on a single rectangle. It worked, so I built a whole row and tested it again. That failed due to a simple math error, so I fixed that error and got a row of three black rectangles with equal white spaces between.

Here is an example:

This is a single line from the AppleScript that describes the coordinates of one of my rectangles.
It uses a Cartesian coordinate system of X,Y points, then in a subsequent line it
fills the rectangle with black.
In the line of code you’ll see that the bold words are commands, the green words are variables, and the integer values are shown in bold. Sets of brackets hold pairs of X,Y coordinates made up of variable names
and integers. In the script I first set the variables to 0, and then increment them for the various rectangles.

Then I incremented the vertical starting point and made the second row of rectangles. That worked also, so I made all four rows. After testing it, I saved the canvas and began inserting the 24 photos into the document.

I only need two master masks for this project, and one is the inverse of the other, so I only created one checkerboard.

After I dragged a new photo onto the canvas, which results in that image being on a new layer, I selected one or the other mask, then created a layer mask for the photo, isolating it from the adjacent images (one only has to avoid those whose corners touch). The next photo used the opposite mask, and gradually I built the collage with its 24 images, 12 using the positive mask, the other 12 using the negative.

Note: It’s important to unlink the image from its mask on each layer so that the mask stays put while you move the photo around inside it. This is done by clicking on the chain icon between the image and its mask in the Layers menu.

The photos are very consistent, with the mountain appearing almost exactly in the same location in each image (I use another script I wrote in 2016 to crop the images to the correct ratio, and to put the mountain top in the same position from image to image.

It took several hours to build the final collage, and it is pixel-perfect. There is not a single corner that is not correct; no hairlines show; no gaps exist. It is absolutely accurate.

The Scripting Guide published by Adobe gives (mostly) clear examples and instructions for controlling Photoshop. They also publish similar guides for Illustrator and InDesign. Automating tasks and producing visuals with absolute precision are the byproducts of the scripting supplements for these three applications. The syntax can be confusing at times, but I have always been able to find online examples that clear-up ambiguities when I get stuck in an AppleScript program, or an Adobe-specific implementation of AppleScript.

Regardless of your choice of programming languages – VisualBasic, Javascript or AppleScript – if you commit to learning how to automate with scripting, the benefits are many. It takes a bit of time to master it, but the results are worth the effort.

Posted in Art, Imposition and Pagination, Photography, Photoshop techniques, Software | Tagged , , , , , , , , , | Leave a comment

Illustrator graphs show me how my solar system is working

Greetings Blognosticator readers!

I (sort of) apologize for my long absence from this blog. I retired from my teaching position at Cal Poly in mid-June, and have been relaxing since then by not contributing essays to this forum. Instead, I have been working in my wood shop, doing home repairs, and doing a lot of reading. I have also been doing quite a bit of kayaking on Morro Bay, in San Luis Bay, and on Santa Margarita Lake, all of which are close to my home. I like retirement so far!

Seven years ago I installed eight solar panels on the roof of our home. I did a lot of the installation myself, and hired an electrical contractor to handle the connection to the utility box and working with the city inspector to get the work approved so that the electric utility would allow me to turn it on.

At the time I estimated that the return on investment would be pretty fast: four years.

I got a federal tax credit for about 30% of the cost of installation. I also got the do-it-yourself discount. I did all the planning, permit applications, site drawings (I went overboard on these, making realistic drawings in Adobe Illustrator), the roof preparation, installation of the rails, panels and inverters. My electrician put in the conduits, and drilled through the roof into the garage to deliver the power to my electric panel. He and I pulled the wires together, did all the finished wiring on the roof, and celebrated the completion of the project when we got the permit approved.

This is a sample of the illustrations I created to apply for my building permit.
The local officials were impressed by the detail.

In San Luis Obispo we get a lot of sunlight; it almost never rains, and it never snows. My pre-installation estimates showed that the panels would generate enough electricity to break-even in about four years. After we turned the system on, it generated more power than my estimates, and it paid-off in three years and three months.

But I still had an electric bill.

I had foreseen this, and I had the electrician put an extra stub conduit on the roof with a weather cap on it. I had also received my building permit for 12 panels (even though I installed only eight in the first round). 

So, adding four more panels – after three years – was really easy. I bought the panels, installed the rails and attached these to the roof. I ran a new conduit to the stub, pulled the wires, and connected everything together at the junction box. Then I turned the system on and was suddenly generating 50 percent more energy!

And my electric bill dropped to zero (though our utility requires a minimum $10 per month to be connected to the grid).

On the “net metering” agreement I have with the utility, I get one power bill each year, and that bill arrived last week. For 2020 I paid $10 each month, and, after all charges and debits, I owe the utility a whopping $18.06 for the year. My cost for electricity for the year was $138.06.

The investment of money and labor to install these solar panels was obviously a good one. For the first three years I checked the host web site every day, then, after pay-off, I stopped looking every day. After installing the additional four panels I was back to checking it every day, but I lost interest after a couple of years.

Last week I decided to check up on my system: Was it still working correctly? Am I still generating enough power to break-even? How did 2020 compare to previous years? Are the panels degrading in efficiency?

(And this is where this article takes a turn toward the graphic arts…)

I use Enphase inverters on my system, and they report through a hardware portal to the Enphase web site. This site allows me to monitor my system, to check for equipment failure, and to monitor my power generation. That site allows me to see my generation graphically, and it also allows me to download spreadsheet files for analysis offline.

I downloaded two years’ worth of data in CSV format. Then I opened those files in Microsoft Excel, deleted some text that was redundant, and saved the file in text-based tab-delimited format (this may not have been necessary). Oddly, I couldn’t get Excel to find-and-change the entries – it just refused to work – so I placed the tab-delimited text into an Adobe InDesign document and did the editing there (strange, I know, but it worked). Then I exported it back to a text-ony file.

The amount of data is not overwhelming – two years of two columns of data – but I knew that I would need to make a relatively large page for my graphs in Adobe Illustrator to show the information adequately.

Adobe has relegated the Graph tools to the Illustrator backwater, so I had to move them to the current tool bar (this is done at the bottom of the tool bar pull-out menu). Once there, I started with the line graph tool and made a 36-inch-wide graph rectangle with the tool. This opens a table into which you can upload data in several formats, among them tab-delimited text.

I imported one year as a line graph and plotted the daily watt-hour values on the vertical while plotting the days of the year on the horizontal. 2020 was a leap year so it has 366 data points; 2019 has only 365; this creates a single-day problem of alignment that I decided to ignore (it’s gestalt).

This is the line graph version of my solar data for two years. It’s too busy and not very attractive.
It reminds me of the financial graphs presented in the Wall Street Journal.

Graphs can be tricky. It’s difficult to show clearly what is important, and do it in a way that is attractive. Illustrator’s graph tool is also a bit strange in its automatic formatting of text and indices. It chooses a default font (perhaps it’s the one you’ve been using most recently?) and scrunches mountains of data (the horizontal date information) into a pile of overlapping entries. I selected these text items with the Direct-select arrow and changed them from 36 point Myriad Pro to 4 point Myriad Pro Condensed. That made them legible at least.

The text indices for the Watt-hours were too small, so I changed them, and also edited them to Kilowatt-hours by removing three zeroes from each one (the vertical values are still valid). On a good day in summer we generate 22 KWh of power; on a cloudy day in January we generate about 5 KWh.

One pleasant thing about the Graph tool is that elements in the graph are easy to select en masse. The line graph is made up of 365 separate line elements, each ends with a small rectangle at its intersection with the next. To color, stroke and fill these individual items I use the Select menu to choose all items of the same stroke weight, or all items with the same stroke and fill. In the process I can customize the graph to make it more attractive.

As long as I don’t Expand Appearance on the graph, the data in it are interactive. If you change an element in the table, the graph will be updated automatically. Most of the time I like this feature, sometimes it gets in my way. Deleting, grouping, averaging, or joining elements will result in one of Illustrator’s obsequious dialog boxes telling me that I cannot do that.

My final graph of the year 2020 looked pretty good, until I showed it to my greatest critic – my wife – who was confused by it. After taking her thoughts under consideration I decided to redraw the graph as a vertical bar chart. This gives the data body, showing that the solid areas represent electricity being generated. The line graph shows just a wiggly line that is not as informative visually.

This is my modified bar graph, showing electricity generated in 2020,
with 2019 generation data overlaid as a line graph.

But having a graph of my power generation for one year didn’t answer my questions. How did my panels perform compared to another year?

I drew a second graph with 2019 data as a line graph. I selected the line elements only from that graph and pasted the data on top of the bar graph for 2020 (this is where the 366-to-365 data points became an issue). With this line graph superimposed atop the bar graph I can see that my electricity generation from year to year is approximately the same.

Taking the hours of daylight into account (a simple curve with its peak in July), my system generates a more or less constant amount of power relative to exposure to the sun. Thus my charts are annual descriptions of sunlight falling on my roof.

Curiously I made more power in 2020 than I did in 2019. This answered a couple of my questions: are the panels still performing efficiently? and do they perform about the same as they did the previous year? I was pleased by both answers.

Eventually I added a gradient to the vertical bars with red at the top (for high output) and yellow at the bottom for low output. Superimposing the 2019 data on top of the 2020 bar graph is noisy, but it does illustrate the output for the year in a way that I can see it clearly.

In 2020 we had our house fumigated, and thus the panels were covered for 1.5 days. That is the only time that I have not generated power with the system.

I like Illustrator’s Graph tools; they put the data on the page in a way that I can customize it to make an attractive (and effective?) graphic illustration of numerical data. Adobe has not updated that tool in many years. Perhaps it is in need of a facelift to make the graphs more attractive or easier to manipulate.

Posted in New technology, Software | Tagged , , , , | Leave a comment

Saved by the wire (and the fiber-optic cable)

You would never think that a wire could slow Internet speed. Well, perhaps you would think of that, but I hadn’t thunk about it much. I was once warned by my friend Eric that the wrong kind of wire could bring an Ethernet network to its knees. I knew he was right, but I didn’t act on it because it didn’t affect me. Until last week.

Ethernet, the electrical and electronic architecture that runs the Internet, runs on copper wire (mostly). Inside our offices and homes are kilometers of cables that deliver digital signals to our computers from nodes of the Internet far away.

In my home I have numerous computers, two printers, two Apple TV units, one Roku, and a device that reports from my roof-top solar panels. These are the wired connections, all of which use Ethernet. Unfortunately, our home was not wired with the Internet in mind. Though it’s a modern house, it has an older telephone wiring system. That has not been helpful in my networking efforts.

I have solved two problems with network wiring by using devices called Tendas. These take an Ethernet signal and put it onto the household electrical wiring. They cleverly deliver Ethernet where it is impossible to get wires; you plug a Tenda device into any electrical outlet, and then plug an Ethernet cable into that device. Elsewhere in the home you put another Tenda and pick up the signal there. With these devices I have Ethernet at my electrical utility box, and thus I can monitor my solar panels.

Apple’s USB-to-Ethernet adapter. It’s rated up to 100 Mbs. (Image from

A couple of weeks back my neighbor told us that she and her husband were on the new fiber-optic Internet service provided in our city by AT&T. I had seen men on fancy trucks pulling the fiber along the overhead lines a couple of summers ago, but I didn’t know it was ready for consumers. I was excited! Our Internet has been provided for the past 15 years by a local Internet Service Provider. It was bonded DSL, and the fastest speed we ever got was 32 Mbs down and 6.7 Mbs up, which is pretty poky. The alternative was the local cable TV supplier, whose offering is much faster, but famously unreliable.

Mbs is Megabits-per-second, the measurement unit of network speed. Divide that by 10* and you have Megabytes-per-second – approximately – and that tells you how fast your Internet connection is. On a 100 Mbs connection you can send and receive 10 megabytes per second. If you have a 25MB Photoshop file, it will take approximately 2.5 seconds to transmit the file over the Internet. That is acceptably fast in today’s world. AT&T’s fiber-optic is ten times faster than that, so the same 25MB file would take only 0.25 second to transmit. I was salivating!

I jumped. A call to AT&T got me started. Limits? Throttling? Extra fees for non-AT&T suppliers? No, no, and no. It seemed too good to be true.

I signed up, and three days later a team of skillful workers showed up at my house. They pulled the fiber through an underground conduit into the phone box on my house, then they drilled through the wall into a discreet spot behind my television and installed the gadget that converts the optical signal into an electrical signal. Then then routed that back out to the phone box, and somehow got the signal upstairs to my office, where they installed a modem and WiFi router.

In three days I went from 32 Mbs to 1000 Mbs signal speed. I was thrilled. The AT&T service is also half the price of my former service. I now get 30X speed at .5X price. You can’t beat that.

Apple’s Thunderbolt-to-Ethernet adapter. This one is rated to speeds up to 1000 Mbs. (Image:

Everything was moving faster except my primary computer, a Mac Pro (cylinder). With that machine I was getting only 67 Mbs after the fiber-optic install. My MacBook Air was getting a similar speed, but its Ethernet comes through a USB adapter (there is no Ethernet port on a MacBook Air). Even the new WiFi was faster than either of these, coming in at about 350 Mbs.

I started troubleshooting. Was it my 16-port Ethernet switch? Was it my Mac Pro? Did I have a defective connector? I started dragging wires around the office, draping them over tables and chairs, connecting this to that, and then that to this, and I discovered that my wife’s Mac Pro was running at 900 Mbs-plus while mine was clocking-in at a paltry 67 Mbs.

It was the wire.

I looked it up. There are three common wires used for Ethernet wiring: CAT-5, CAT-5E, and CAT-6. They have eight wires inside, each twisted in pairs, then co-twisted at very specific intervals to prevent cross-talk and electromagnetic induction. Though this is pretty esoteric, I realized that I must be a victim of induction-related trouble.

I installed a new Ethernet cable between the Switch and my computer, and it suddenly soared to 938 Mbs! Wow!

My cable had also been way too long, so I had coiled the left-over part into a nice 8-inch coil on the floor behind my computer. It turns out that this is a very bad idea since making a coil in a wire whose purpose is to prevent induction creates a big inductive opportunity. It’s better to have the cable go across the floor and never coil it up.

I also discovered that Apple makes an adapter to bring Ethernet into a MacBook Air through the Thunderbolt connector rather than through USB. I ordered one. When it arrived, I plugged it in and my MacBook jumped to a phenomenal 824 Mbs. This is 12 X the speed I got with the USB adapter. (Apple’s specification on the USB adapter says it goes to 100 BaseT which is only 100 Mbs, which explains the speed problem.)

Now all of our computers are running at full speed. Our TVs are happier, my solar panels are exactly the same, and my iPad is also happier. Everything is faster, and I am now a part of the digital community, no longer limping along with old technology

*The geeks reading this will quickly counter: No it’s not! It’s only eight bits per byte, etc. They are correct, but with what are called parity bits and checksum characters, it works out closer to ten, so I always figure that I should divide by ten. It works just as well, and it’s easier math.

Posted in Gadgets, New technology, Technology | Tagged , , , , , , | Leave a comment