Tutorial 17: preparing illustrations. Part 4: go big
December 30, 2017
Matt just commented to me: “One thing I am realizing is that I have loads(*) of cleaned-up, ready-to-post photos in old talks, that I’ve never posted.”
I too have that experience. The problem is, they tend to be sized for a 1280×960 projector screen, which is not really good enough for modern illustrations. A useful rule of thumb is that if you want to be able to print at 300 dpi across the active area of a sheet of A4 — (which is 8.27 inches minus 2-cm margins each side, for a usable width of 6.69 inches), you need horizontal resolution of at least 2007 pixels. But you really want to go much bigger than that, since online journals can provide much better resolution than even the best printed journals.
Now, whenever I create an image for a talk or poster, I consider up front if I will one day want to publish it. If so, I make it at full-size, and only shrink the render.
One of our problems in the sciences is that we work in five different media (at least, I do): in roughly ascending order of formality:
- Blogs
- Talks
- Posters
- Preprints
- Papers
Not everything moves well between those forms. We learned from experience that translating the text of even quite formal blog-posts into prose for a paper is a serious piece of work. The good news is that illustrations are largely identical between all these forms: the hard work of preparing clean multiviews from photos need be done only once, and the later more formal versions need only the additions of scalebars and suchlike.
The moral of this story is very simple: always prepare illustrations in the highest resolution you will ever need, and in full colour. You can reduce resolution later, or reduce to greyscale; but if you prepare at low resolution or in greyscale, you can’t increase resolution later or add colour.
(*) He actually said “assloads”, but I censored it for this blog.
Sauropod Vertebra Picture of the Week 
December 30, 2017 at 9:52 pm
“The moral of this story is very simple: always prepare illustrations in the highest resolution you will ever need, and in full colour. You can reduce resolution later, or reduce to greyscale; but if you prepare at low resolution or in greyscale, you can’t increase resolution later or add colour.”
As an addendum, this is good advice even if you think you are never ever going to present a poster, talk, or blogpost. Most journals have very different standards for how they want figures to be submitted, and what may be acceptable for one journal may have to be resized if the manuscript gets rejected and has to be sent to another. This eats up more time than you think if you aren’t prepared for it.
As another addendum, if you do work in Illustrator, Photoshop, or a similar program always, always, ALWAYS, save a clearly named copy of your image with the layers and text unflattened in case you want to reposition text. Sounds like obvious advice but it’s really easy to forget.
As another anecdote/personal tip, it’s often a good idea to have large (“page-sized” versions of your image, even if you only intend to submit the picture at column width. I’ve had cases where some of the images in an article were resized by editor fiat, and I was not asked to submit a resized version of those images, meaning the published images were either grainy or shrunken.
On a related note, what do you guys do about large numbers of similar images on your computer from old papers? I save copies of images taken of specimens (e.g., original, cleaned for publication, etc.) but it ends up taking up a huge amount of space on my hard drive, especially as more papers get published. I keep thinking about putting them on an external hard drive so my computer isn’t so slow but I keep worrying about needing the images “just in case”.
December 30, 2017 at 10:29 pm
Thanks, Anonymous, lots of excellent points here. Your last question is an important and complex one, worthy of a post of its own. Thanks for the prod.