For a transient thermal event, select radiometric JPEG when the required evidence is a measured still image, and select radiometric video when frame-to-frame temperature behavior must be retained.
The choice affects measurement repeatability, post-processing, storage, and analysis: R-JPEG keeps radiometric metadata with each image, while video must preserve temperature data across its frame sequence.
Choose by evidence requirement, not file extension
A radiometric JPEG is a still-image workflow for a point-in-time measurement. AV3 AEROVISUAL discusses differences among the .TIFF, .JPG, and .R.JPG image formats used to store thermographic images [S1].
A radiometric video sequence is the stronger choice for a transient event whose peak temperature, heating rate, cooling rate, duration, or propagation across adjacent components determines the outcome. A conventional visual video may show the event clearly, but the decisive criterion is whether quantitative temperature history must survive acquisition and export.
Compare the two workflows against four criteria
The same criterion can produce different recommendations. A radiometric JPEG is cheaper to review and report when one image answers the question. [S2]
A forum post reports that FLIR Research Studio could save radiometric JPG and ATS files that did not resemble the radiometric JPG files from LWIR cameras, and that exiftool yielded no thermal data or Planck constants from the JPG file [S3].
Radiometric JPEG is for repeatable point measurements

R-JPEG differs materially from ordinary JPG and TIFF workflows. AV3 AEROVISUAL discusses the characteristics and differences among the .TIFF, .JPG, and .R.JPG formats used to capture and store thermographic images [S1].
The principal advantage is that radiometric metadata remain associated with the smallest homogeneous image unit, the pixel.
Absolute temperature is not required for every thermal inspection. Relative differences can reveal a cooler solar panel among similar neighbors, warm brakes on one vehicle, or a person concealed in vegetation. Radiometry measures the amount of radiation coming from a target, usually in terms of target temperature, and is not always required for a given application [S2].
A useful workflow is to capture a high-quality still with the target filling a meaningful part of the frame, record the measurement inputs, and verify that supported software can change emissivity and palette after capture. For a detailed radiometric-image analysis workflow, see Radiometric JPEG Analysis: Preserve Measurement Data.
Radiometric video is for timing and thermal behavior
The recording must include the complete event, not only the visually obvious peak. Radiometric video provides an ordered set of measurements, but it still requires an export path that retains the underlying temperature data rather than only rendered images. File extensions such as ATS, SEQ, and CSQ cannot be treated as interchangeable without verifying the camera, software, and metadata structure [S3].
Frame rate alone does not establish event resolution. Record the acquisition frame rate, timestamp basis, duration, and temperature range, then verify whether the exported sequence includes per-frame temperature values and calibration metadata. If only an ordinary video remains, the result supports visual interpretation but not defensible frame-by-frame radiometry.
Measurement accuracy limits both choices

Thermal sensitivity and temperature-measurement accuracy are different properties. Sensitivity describes the ability to detect temperature differences, while accuracy describes the error in an indicated temperature.
Accuracy depends on the selected measurement range.
Emissivity, reflected apparent temperature, atmospheric effects, distance, viewing angle, focus, and scene motion can alter the result. Isotherm color does not improve measurement accuracy; the isotherm feature has the same accuracy as the temperature measurement [S2]. For a short event, verify the temperature range before capture and avoid relying on a display palette to communicate precision that the measurement chain cannot support.
Apply this decision sequence
First, define the required output: one defensible temperature measurement, or a temperature-versus-time record. If the acceptance criterion is a measured value at a location, capture radiometric JPEG and retain the raw radiometric file. If the criterion is peak timing, rate of change, duration, cooling, or propagation, record radiometric video covering startup, event, and recovery. [S2]
Second, classify the task by user need. Pure anomaly detection can often use visual thermal imagery without absolute radiometry. Repeatable absolute measurement requires a radiometric camera and comparison with established limits [S2]. Third, confirm that the chosen export preserves the needed data: R-JPEG for adjustable radiometric stills, and a documented radiometric sequence for per-frame analysis.
Fourth, test the complete chain before the field mission. Open one representative file in the production software, verify palette control, change emissivity on a still, inspect a video frame's measured temperature, and export the report evidence. A proprietary radiometric file that the analysis team cannot open is not usable evidence, regardless of the camera's measurement capability.
For a transient event whose outcome depends on when and how quickly temperature changes, select radiometric video. For an isolated condition whose outcome depends on an absolute surface-temperature limit, select radiometric JPEG, and verify that emissivity and background-temperature corrections remain available after acquisition [S1][S2].