Paintings from Sunset Series by Spence Munsinger, Color Field + Blank White Canvas + Realism + Contemporary Abstract Art, original paintings for sale

"We all know that art is not truth. Art is a lie that makes us realize truth..."
― Pablo Picasso

9000 x 12000 pixels – long road home


I post-scan-processed “long road home”. This part is mechanical, just adjusting to bring out the best possible image. But the visual reward – to have truly printable high quality images – just the impact of that many pixels is incredible. The original canvas is 30″ x 40″.

This image links to the reasonable-sized detail image – the blue outline is the section that the detail image captured – the detail image links to a full size image of the section – you will have to scroll around to close it, it is larger than the resolution of most monitors (the image size is roughly 1200 pixels x 2200 pixels). It does give you a sense of what a 44 MB jpeg image can capture – distilled from a 618 MB TIFF file.

long road home detail

This was not sharpened except at the noise reduction filter level, and that only slightly. To print effectively it is likely to require more sharpening. There is a technique called high pass filter sharpening [1], which failed miserably in this case, but was outstandingly effective on “weathervane”. On this painting image, it turned the colors toward shades of grey, and I went back and started over again.

Back to top…


— spence


[1] Note:

High Pass Noise Filtering (Photoshop CS2)

On the Layer palette select your Background Layer and right click. Select Duplicate Layer.

  • With this new layer highlighted select Filter / Other / High Pass. Set the Radius to 10 and click OK.
  • Zoom into your image to Actual Pixels level so you can better see what you’re going to do next.
  • Go back to the Layer Palette and select Hard Light from the left drop down.
  • Now go to the Opacity Slider and select a level of sharpening that seems best to you. Usually something between 20% and 70% will be best.

Back to article…

images of art for print reproduction


What does it take to produce a reproduction of a painting? Not an image for a website, but an image of high enough quality to use as a print reproduction? At full size or as close to full size as practicable?

weathervane detail

Two measurements – two sides of the process:

  • DPI (dots per inch) is the measurement on the printer itself. That’s your print output.
  • PPI (pixels per inch). Pixels per inch is the measurement of the image resolution within the file.

Differences between DPI and PPI are interpolated in software – like upsizing or downsizing an image in Photoshop, done on the fly. Upsizing, making larger, is prone to errors, and is better done in Photoshop rather than by the printer driver; downsizing is not.

The acceptable file size for full print reproduction of a 30″ x 40″ painting (300 pixels per inch (PPI)) is 9000 x 12000 pixels. 240 PPI is usually acceptable. I just took a 6 x 7 cm medium format color negative, scanned in a Nikon film scanner at 4000 DPI/PPI, and with cropping the final resolution image is 8964 x 11016 pixels, and a whopping 493.22 MB in size (16bit). Increasing the resolution of the height from 11016 to 12000, and then cropping back to 9000 for the width will hit that target. I’ll increase the resolution in Photoshop to 12000 using bicubic smoother. The final file gets saved separately from the original scan file. Both are TIFF files at this stage, uncompressed.

The combination of Mamiya medium format RZ67 camera and Nikon film scanner result in an 89 megapixel image. Ten times the resolution of my Canon 30D digital. That’s exactly why film rules for this application…

Paintings are not photographs. They are imprecise by the nature of the medium. The final file will print very well at 300 DPI on canvas, or heavy paper.

The negative is scanned at 4000 DPI (sorry – really Pixels per Inch at this stage, but Nikon uses DPI, as does all the scanner literature). I multiscan the negative 4x to 8x (reduces digital noise, at least some). The original scan file is opened in Photoshop (slowly – 565 MB file). I use Neat Image, a noise reduction program to clear the digital noise out – it works both as a standalone program and a Photoshop plugin, and produces outstanding results in cleaning digital files, whether scans or digital photographs. Get the resolution correct, sharpen for printing, save the file, and that’s it.

I took a test roll of ten shots of a single painting to get exposure correct – in the lighting, with the filter and set up exactly as the actual shots would be. I varied the exposure from measured light amount (1-1/2 seconds at f16) upward 4 stops and downward 4 stops. It turned out the measured exposure was over-exposed, and dropping a full f stop down (less light) created a much better image. I also found that I needed to move closer to the painting, filling more of the frame, to get the largest capture possible (less cropping, more actual pixels).

This last Friday evening I took the test roll, ten shots, from set up to photographing to development of the film and hanging it up to dry. Saturday I scanned the resulting negatives, parsed through and determined the ideal exposure. I then took three rolls of 10 exposures each, developed all three rolls, hung up to dry, and broke down and packed away the setup.

Total cost in film and chemicals – $36. Equipment costs to photograph, develop, scan, adjust and store the images – $5500.00. Having control over the quality of the final image files – priceless…

weathervane section

Scanning and adjusting the images will take a couple of weeks. Test prints to validate the image – another week. I’ll post some more detail sections as I work through these.


— spence


too many images…

I caught up. I developed 6 rolls of black and white and two rolls of color, all medium format, by this last Sunday AM. Scanning those 80 B&W, plus another 25 color images, again, not too bad, I can keep up with that.

Keeping the images ordered and findable, that’s no longer simple. The sheer size of the storage required boggles the mind. A 6×7 medium format image file runs 296 MB. That’s a huge amount of detail. Images to print are even larger. But 296 MB is sufficient detail to only miss the original negative a little bit.

That kind of size isn’t needed for every image. Of the 105 images this past weekend, perhaps 11 were interesting enough to be true keepers. Eight of those were because I had noticed a quality of image in photographs of snow on medium format Kodak TMY 400 black and white film, the extreme contrast between the trees and branches, blacks and greys, and the blinding highlights in the snow. In most series of eight rolls of film, maybe three will be good photographs, maybe another 4 useful for form or inspiration or reference.

Most of this stuff could be scanned and saved at a much smaller size. But that only works for me if I can second guess that judgment later on, and I have preserved the original RAW file in the case of a digital photograph or the location of the negative in the case of a film image. That means I’m having to add in organization. Right now I could find the negative, eventually. Maybe. I’d like to be able to be certain.

Probably something like a thumbnail image ties into a unique index number tied into the location of the file on disk, tied into the location of the negative and/or DVD optical backups. Something like that.

— spence


20080527 Tuesday

After having thought about putting the organization of the images into a program like a wiki, or even a wordpress (this weblog) program frame, I decided that’s overkill.

Simplest would be:

A script, probably in perl. Uses ImageMajick and maybe exiv2 to read out metadata, we’ll see. In initial testing the exiv2 data was minimal, but maybe that’s a reflection of the data presented in the image.

Point it at a directory, it reads the files, renaming each to match a unique index number in the filename, then creates a thumbnail/contact print in a website directory, matching name and index number. Assembles a page that includes each contact image, plus metadata, file location, filename. The page name is the date _1 for the first, _2 for the second, etc. Each date-named page is indexed on a master sheet, where comments can be added…

Manually add the index numbers to the folders in which the negatives are stored, and add in any DVD backups and index number (separate, different dataCD database) for these. Add this information to the listing page.

All of this can be served out locally by a simple virtual webserver…