How are these modern physics figures made?

I am curious to know how the figures in popular introductory physics/Engineering textbooks are made. As examples of what I mean, see the posted figures and my comments, all from various editions of the Physics book by Serway and Jewett book (I think).


Look at all of these color gradients throughout the metals, and the imitations of lighting on the spheres. Notice the relatively realistic nature of the force scale and pullies.


Notice here that we have proper, LaTeX-like font on the mathematical symbols.


Wow. So many complex shapes, good water drawings, and the ability to annotate the drawing freely.


Lastly, we see more extremely complicated, realistic objects, like the ship and the mountain. But this isn’t just pictures, it’s clear that the software used has a very precise coordinate system in it, as evidenced by the kinematics drawings overlaid.

I am aware of other posts on Physics SE recommending drawing softwares such as PowerPoint, TiKZ, aymptote, mathematica, etc. But I highly doubt these figures were produced in any of these. My suspicion is that the publishers have some specialized software for this task. So what is that software? Does anyone know? The Physics SE did not have the answer…

It is important to realize the stipulation that one must be able to type properly formatted mathematics (using some manifestation LaTeX) into the figures. I say this because I intend to use such figures in a book, and the fonts must match.


They are most likely drawn by a human on a computer. Manually, with some definition of manual. It always strikes me a bit weird how people this day and age can not draw simple line drawings themselves. So to begin to answer this question I would ask you how long would it take for you to draw the following image with a pencil and paper?

triangle angle

Image 1: How long would it take for you to draw this neatly with a ruler, paper and a compass?

I actually timed it, it took me ~5 minutes to find the tools. Drawing it neatly took me ~3 minutes. Now for comparison I timed how long it took to draw this in illustrator and it took me a few seconds to start and ~2 minutes to draw it neatly with no hurry. Now having done this also in Mathematica takes about 12 minutes once I layout stuff properly*.

I could certainly do better if my drawing app would have a constraints based solver and I could easily drop the time to a minute. Unfortunately no constraints based app exists that would be as neat as illustrators output, so I’m ready to take a 50% hit on drawing if i dont need to use 200% extra time to make the output neat.

The reason is that the easiest way to describe the problem at hand to a computer is to in fact draw the image. CAD apps do have technology to make this even faster in more complex cases but they are not adept at the illustration part. Even in CAD apps technically you draw something topologically correct and input values. This is the fastest we have come up with to this date, if you can come up with something better then I’m pretty sure I could make you a millionaire with some effort, entire branches of engineering would turn on their heels.

Now once you have the basic shape in a computer app, moving to the designed style is not very hard.

enter image description here

Image 2: Applying style, takes about as long as deciding what the style is going to be.

I the end, with some liberal reuse of iconography and styles. Those images do not take very long for a experienced illustrator to draw. It takes far longer to decide what to draw and where to fit the image than actually drawing it. So in the end its more of eliminating one humans procrastination to another humans procrastination, at last until we have effective AI. Software could be slightly more optimal but not worth the effort for one book especially as you’d need to beat current crop of software on most fronts.

Some points:

  • You can embed latex typeset parts in illustrator (this isn’t more tedious to manage than laying two windows next to each other). Not all parts of a image need to be made with one tool.
  • Not all images need to have been made with the same tool

* Mathematica would be great if its annotation tools just could snap to grid.

Source : Link , Question Author : MaanDoabeDa , Answer Author : Community

Leave a Comment