The quality of radiometric models for generation of synthetic scenes is difficult to evaluate because of the cost and complexity of experimental data collection. In addition, the number of variables to be controlled or monitored is prohibitive. This paper suggests the use of a cold hohlraum and calibrated LWIR imaging systems as a method for testing certain components of radiometric models. The model tested includes self-radiance from the source, reflected background radiance, specularly reflected sky radiance, diffuse sky radiance, diffuse and hemispheric angular emissivity, atmospheric transmission, path radiance, sensor spectral response functions, sensor instantaneous field-of-view, and various types of sensor noise. The laboratory measurements are able to evaluate those aspects of the radiometric synthetic imaging model used in generation of the radiance field toward the sensor as a function of view angle. The final modeled image can be compared to the laboratory observed radiance field propagated to the sensor. This method allows for evaluation of the model's ability to effectively account for changes in emissivity with look angle, the effects of changing "sky temperature," changing emissivity of background surfaces (e.g. water roughness in naval scenes), and the interplay of target and backgrounds in the formation of infrared "halos." Both image comparisons for phenomenological assessment and quantitative radiometric comparisons are included in the discussion.