"The scariest moment is always just before you start ."

(Scroll to the bottom for the answer.)

The moment before you begin is often the most uncertain, which is why it can feel intimidating to take the first step into something new. But that same moment can also be the most energizing. A new year tends to feel like that. It is a pause between reflection and action, where you look back at what worked, question what didn’t, and start forming ideas about what comes next.

This January feels similar for me. Aya Takase built this CT newsletter into something genuinely useful and occasionally fun, and she is now passing that torch to me. My goal is to keep the spirit and structure you’re familiar with, while continuing to share practical, thoughtful content that helps you think clearly about CT experiments and the questions you are trying to answer.

That moment is familiar in X-ray CT as well. Right before starting a scan, you are making a series of decisions that shape everything that follows. What question are you trying to answer? What outcome would confirm or challenge your hypothesis? What are the trade-offs between resolution, time, and data quality?

Those questions are easy to rush past, especially when you are eager to see the data. But spending a few minutes on them up front often makes the rest of the experiment smoother, faster, and easier to interpret. Clear intent helps guide parameter choices, sets expectations for what “success” looks like, and reduces the urge to scan first and optimize later. With that in mind, I thought I would start the year offering a few practical tips.

You can find additional tips in our blog articles and X-ray Computed Tomography webinar series and workshops. We’ll have more coming this year, so stay tuned.

- Angela

To be efficient, inspired, and informed.

Here are a few practical tips for approaching your CT experiments.

1. Write the question first, then define “success.”

Before you touch parameters, write one sentence: “This scan will tell me whether ____.”

Then add a measurable success criterion, for example:

• “Detect cracks ≥ 30 µm through the full volume,” or

• “Measure wall thickness within ± 20 µm,” or

• “Separate pores from matrix with stable segmentation.”

If you cannot state success clearly, the scan is still a brainstorm, not an experiment.

2. Choose the smallest feature that matters, and back-calculate voxel size.

Pick the minimum feature size you need to resolve (not the smallest feature you hope exists). A practical planning rule: voxel size should typically be 3-5 times smaller than the target feature. Then, sanity-check whether that voxel size is feasible given the sample size and required field of view.

3. Set beam energy and filtration for the sample, not for habit.

Start with material class and thickness:

• If penetration is the limiter, raise beam energy and consider filtration to remove low-energy photons that mainly contribute to dose and beam hardening.

• If contrast is the limiter, avoid automatically increasing kV, because you can reduce contrast even while improving penetration.

The planning mindset: you are balancing penetration, contrast, and artifact risk, not chasing a single “best” kV.

4.       Budget signal-to-noise early (time is a parameter).

Decide what you can afford in scan time, then allocate it intentionally:

• If you need detectability (low-contrast features), prioritize signal through exposure and averaging rather than only tightening voxel size.

• If motion or drift is a risk, shorter scans with adequate signal beat long scans that look sharp in theory but smear in practice.

A useful mental check: if you expect segmentation to be the final step, plan for the signal-to-noise required to make segmentation stable, not just “pretty slices.”

5. Treat artifacts as diagnostics, and plan one “artifact check.”

Go in expecting at least one artifact mode and decide how you’ll recognize it:

• Beam hardening often shows up as cupping or density gradients in thick regions.

• Rings tend to be detector-related and show up as concentric patterns in slices.

• Streaks can come from high-attenuation regions and photon starvation.

Build in quick checks into your workflow: a couple of slices, a line profile, or a small ROI histogram that tells you early whether artifacts will degrade data analysi.

6. Fix acquisition fundamentals before tuning reconstruction or post-processing.

If the raw data is limited, reconstruction cannot rescue it. Before you start chasing sharpness or “cleaning up” artifacts downstream, confirm that the acquisition conditions are appropriate for the sample: sufficient signal, appropriate beam conditions, a stable setup, and a geometry suited to the part.

Then, and only then, refine reconstruction and post-processing.

That's a wrap. Please let us know how we can help you learn more about X-ray CT. We love to hear from you!