Recently, I’ve been very busy with POV-Ray and Matlab. This led me onto a website of a (POV-Ray) visualisation specialist, Paul Bourke. Here, I found useful information on making and using a 3D density file.

A 3D density file is a file that contains density values for each volume element (voxel) in a volume. Magnetic resonance imaging (MRI) is one example of that, but also gas clouds (e.g. as in space), or simply electron density volumes.

Rendering this by means of POV-Ray takes an enormous amount of time, but eliminating the “scattering” behaviour of the medium saves much time.

I’d advise all parties interested in data visualisation to take a look at Paul Bourke’s website. It is filled with many goodies and many good images depicting the many capabilities of the POV-Ray raytracing software.

On a related note, there is a nice article out called “Digital Doctoring: How to tell the real from the fake” by Hany Farid. It discusses the artefacts created when manipulating images. It is a nice introduction to image doctoring effects, but unfortunately is very light on details. Rumor has it that Adobe is working on integrating the details in Photoshop®.

A while ago I gave a small three-hour workshop to introduce the functionality of the POV-Ray Raytracing software. Last week, a participant of that workshop came with the intention to build an animation for his Ph.D. project.

I was pleasantly surprised by the number of participants to that workshop, and this indicates a need for graphics generation instruction in the current curriculum.

Another reason is the need for scientists to express themselves not only in words understandable to their peers, but also in images, which are much more accessible to the general public.
Allow me to substantiate this claim by a quote from Prof. Atkins:
“Computer graphics are opening up whole new methods of understanding and communication in science, and a well-selected image can break through the fog of incomprehension and shine out with its message.”

POV-Ray is a worthy program for making scientifically oriented, aesthetically pleasing imagery. It is open-source (and thus free), multi-platform and uses script-able input files. Furthermore, it offers true four-dimensional (3D+time) animation options. Due to its raytracing nature, it can render lifelike imagery.

My lecture handouts are available here: Some POV-Ray examples / tutorial, if there is interest in more information on the subject, please do drop me a line.

In 1999, a little before my time, an article was published by Miller et al. in the International Journal of Biological Macromolecules, which couples results from AFM (a type of scanning probe microscopy) measurement results to SAXS patterns.

The captivating bit is that they used the height profile obtained from the AFM results to simulate (by means of Fourier transformation) a small-angle scattering pattern, which they compared to the measured SAXS scattering pattern.

My professor indicated that this is not unheard of, and that comparisons between small-angle scattering patterns and electron micrographs are performed in this fashion.

This also works the other way around. So instead of validating the scattering pattern, a micrograph can be deemed to show a representative part of the sample when the scattering pattern matches with that of a SAXS experiment. This will add more weight to the obtained micrograph and can transform it from a pretty picture to a picture representing the actual morphology.

It is important to do this, for trying to describe an entire surface by looking at a minute fraction of it (as is done with microscopy), is said to be similar to describing the surface of the earth by looking at a square millimetre. This quote is not mine, and I will find out whose it is, for it accurately defines the limitations of microscopy techniques and highlights one of the values of the scattering technique.

Nevertheless, I like the comparison by Miller et al., because it not only shows the authors to be familiar with several techniques, but also shows that they do try to verify their results obtained from one technique, with results from other techniques. In my eyes, this self-evaluating nature is not nearly being practised often enough.

There is a nice article (short communication) out, co-authored by one of my heroes of SAXS: W. Ruland [1].

This man is well versed in the principles of the field, having specialised in it for over 40 years now. Over the years, he’s kept himself busy with both practical and theoretical issues. One of the aspects that captured my interest in this man is that he is not afraid to check someone else’s work. Previous work of his (which I may elaborate upon at a later stage), provided a point by point discussion of the work by another researcher.

That particular aspect is one I respect. I strive towards having a similar characteristic in my research. He does not focus on his work alone, running in a single direction with eyes closed. On the contrary, he opens his eyes, looks towards the work of others and joins it with his own research.

If scientists (in this field, but not limited to this alone) would recognise the work of others, co-operate with them towards the common goal of the advance of science, or even discuss the work of another, the scientific world would seem a lot more professional. But I digress. The point is that discussion is good, and in my eyes, at least 50% of ones time should be spent looking at the work of others and discussing their results.

The referenced article [1] is a good example of that. As it turns out, two conflicting methods were reported by Kratky (1933) on the one side, and Ledbetter and Norris (1979) on the other. The topic was a relationship between the axial orientation distribution and the equatorial intensity profiles. Both relationships are widely used, but are not equal. Burger and Ruland noticed they are significantly different, and stated that they cannot both be correct.

During a short discussion, it is found that there is a mix-up of variables in the Ledbetter and Norris version, and that the Kratky version is the right one. Whilst for small orientation distribution widths, the deviation is not so bad, it is rightly stated that there is now no longer any need to use the wrong equation.

In my eyes, this says it all. This is the goal one should keep in mind. With this one short work, a method is shown to be flawed, it is shown what the effect is and no question now remains. People that have used the erroneous derivation in the past, might want to determine their errors and should switch to the right derivation as soon as possible.

Simple steps like this ensure progress in the field. I for one will be watching for more gems like this.

[1] C. Burger and W. Ruland. Evaluation of equatorial orientation distributions. Journal of Applied Crystallography, 39:889–891, 2006.