Define the Need, Then Set the Resolution

Define the Need, Then Set the Resolution

“Use the highest resolution available” is not a meaningful imaging requirement, tempting though it may be.

A good specification starts with the decision the image has to support. An engineer may need to distinguish a failure feature. A researcher may need fine structure and the context of the whole specimen. A curator may need detailed documentation with minimal handling. An instructor may need enough detail for students to explore independently, without an unnecessarily heavy dataset.

Four measures hide inside “resolution”

The word is often used for four different things:

• Final pixel dimensions: the width and height of the finished image, summarized as megapixels or gigapixels.
• Subject sampling: how much of the specimen one pixel represents in a given configuration, usually expressed in microns per pixel.
• Resolvable detail: the smallest feature the optical system can actually distinguish under suitable conditions, and that remains visible in the finished GIGAmacro Viewer image.
• Calibrated scale: the relationship between pixels and physical units that makes measurement in the GIGAmacro Viewer meaningful.

These quantities are related, but they are not interchangeable.

The Magnify2 specification lists resolvable detail as high as 1 micron. Nominal sampling can be finer than that. Calculated from Canon’s published sensor dimensions, a Canon R10 samples about 14.9 microns per pixel at 0.25X and about 0.37 microns per pixel at 10X. A Canon R5 Mark II, with its larger sensor, samples about 17.6 and 0.44 microns per pixel at the same magnifications. A sampling value of 0.37 microns per pixel does not mean every 0.37-micron feature will be optically resolved or measured accurately. Our lens-selection article covers how magnification, camera body, and working distance interact.

Define the feature, the area, and the use

The project brief should first identify the smallest relevant feature and the complete physical area that must remain visible. It should also state what the image is for: visual examination, calibrated measurement, comparison, documentation, teaching, publication, or a combination.

Coverage matters because magnification narrows the field of view recorded in each frame. A highly detailed image of the wrong region does not meet the requirement. Capturing an entire specimen at the finest possible sampling can add positions, focus levels, processing time, storage, and review work without improving the decision the image needs to support.

Pixel count is not a ranking

Three records in the GIGAmacro Viewer show why final pixel count alone is a poor measure:

• Bay Area Science Fair: 40 megapixels, assembled from 128 photographs, with a scale of 6,270 pixels per inch.
• Aluminum 8020 2020 Cross Cut: 276 megapixels, assembled from 168 photographs, with a scale of 18,058 pixels per inch.
• Coffee Bean (Roasted): 1.07 gigapixels, but its GIGAmacro Viewer record states that no scale is available.

The largest image is not automatically the most measurable, and the smallest is not automatically insufficient. The three records represent different subjects, scales, capture structures, and purposes.

Plan measurement before capture

The GIGAmacro Viewer includes tools for recording line and area measurements, but meaningful physical results depend on accurate calibration. If a project requires measurement, define the reference, the unit, the calibration method, and the acceptable error before capture. Then check the finished GIGAmacro Viewer record against a known scale. Including the scale measure in the image, where possible, makes that task simple.

Resolution has an operational cost

A gigapixel image is built from many source photographs through focus stacking and stitching, often hundreds and sometimes thousands. More positions, more focus levels, and more pixels per frame all add capture time, transfer time, processing time, working storage, and backup resources.

Test before committing

A representative test turns an abstract requirement into evidence:

1. Capture a small but meaningful area with the proposed camera, lens, magnification, lighting, and focus-stack depth.
2. Process it through the intended workflow.
3. Review the smallest required feature in the GIGAmacro Viewer, and test measurements where they apply.
4. Estimate the full project’s photograph count and storage from the test.
5. Accept the configuration only when it supports the stated decision.

The finished image may be a few megapixels, several hundred megapixels, or several gigapixels. The better measure of success is whether it preserves the required detail, features, context, and measurement capability at a cost the project can sustain. Our capture brief article sets out the other decisions worth making before taking the first frame.