Printing Grab Shots Large

Thom Hogan is solely responsible for making photography my primary hobby. I am indebted to many other photographers for ideas and technique, but were it not for reading Thom’s review of the Nikon D70 when I was looking for a camera, I wouldn’t be here today. I recently had occasion to experimentally verify his maxims for printing when working with an old vacation snapshot.

Thom’s standard line on printing is “If you can’t get satisfactory prints of the largest size a desktop inkjet printer can produce, [usually A3+/13″ x 19″] it’s not the camera that’s the problem.” Let’s try that out on this picture:


This was a quick grab shot while strolling through the National Shrine of the Immaculate Conception. I was already taking several demerits on Thom’s list: handholding, no remote shutter release, no mirror lockup, not base ISO. I was also working at wide aperture, not the best for a subject that wants large depth of field. There was no way to do otherwise, of course; light was dim, my family was waiting for me, the Shrine wouldn’t take kindly to J. Random Tourist setting up a tripod, etc. However, I carefully placed focus, gave good exposure, and used good handholding technique, which was enough for the next step.

When I got home I processed from RAW and cropped a little bit; the image then sat around until I got the crazy idea to print it at my Epson R3000’s largest size, which is A3+. For printing I use a pixel density of 300 ppi, which means the short side should be 3900 pixels for my chosen paper size. I upscaled it in the GIMP using a Lanczos-windowed sinc for the interpolation.

I then did output sharpening thus:

  1. Duplicate the image, and run GIMP edge-detect on the copy (Sobel algorithm with parameter 6).
  2. Convert edge mask to grayscale.
  3. Use Levels to drive the true edges to white and the false-positive edges to black.
  4. Gaussian-blur the edge mask with a radius of 5 pixels.
  5. Repeat Levels to trim out as many remaining false positives as possible.
  6. Decompose the original image as HSV.
  7. Copy the decomposed Value channel as a new layer in the original image, with blend mode set to Value.
  8. Add the edge mask as a layer mask on the Value layer. Now, the Value layer is only being composited in where the edge mask is white.
  9. Apply Unsharp Mask to sharpen the Value layer. With A3+ at 300 ppi I’ve had good results with a radius of 3, and amount seasoned to taste.

This somewhat involved process does two important things:

  • Sharpening is done with regard only for value and not color, thus any latent chroma noise or irregularities are not amplified.
  • Broad areas without edges that do not need sharpening are not sharpened.

I then ran it through my standard printing sequence:

  1. Image -> Mode -> Convert to Color Profile… to apply ICC profile.
  2. File -> Print with Gutenprint…
  3. Adjust paper settings and margins.
  4. Print.

The results blew me away; it looked far better than I had any right to expect. Moral of the story? Thom is right: you don’t need 18 bazillion pixels to print large, if you capture good pixels to begin with.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: