The purpose of this paper is to determine the accuracy and reproducibility of two noninvasive methods of measuring half-value layer (HVL), ring and localization, compared with an invasive technique (suspending tube rotation). The ring method uses concentric aluminum rings about a CTDI ionization chamber at isocenter. Data were acquired using axial CT protocols (rotating x-ray tube, stationary patient table). The localization technique uses square aluminum sheets secured to the gantry shroud to filter the radiation beam, and a CTDI chamber suspended externally at isocenter. Data were acquired using localization image protocols (stationary x-ray tube, moving patient table). The invasive technique was similar to the localization technique except that the ion chamber was placed on the patient table and the tube rotation disabled using service software. Data for all techniques were collected on the same CT system. Independent data sets were collected to determine reproducibility. Sensitivity to ionization chamber lateral displacement from isocenter was investigated. Measured HVLs (mm aluminum, mean+/-std, n=4) were 7.19+/-0.03 (ring); 7.17+/-0.04 (localization); and 7.24+/-0.02 (service mode), which were not significantly different (p = 0.05). Displacing the chamber from isocenter changes the HVL, depending on the bow-tie filter, by as much as 5 mm aluminum. Aluminum filter to ion chamber distances of 25-35 cm provided accurate results. Both noninvasive techniques were accurate and reproducible at isocenter. However, the measured HVL was dependent upon the bow-tie filter and the lateral displacement of the ionization chamber with respect to isocenter. Greater than 2 cm off of isocenter, the ring technique did not provide accurate HVL measurements.