Objective: Direct real-ear measurement to the 4–6 kHz range can be measured with suitable accuracy and repeatability. This study evaluates extended bandwidth measurement accuracy and repeatability using narrowband and wideband signal analysis. Design: White noise was measured in female ear canals at four insertion depths using one-third and one-twenty-fourth octave band averaging. Study sample: Fourteen female adults with reported normal hearing and middle-ear function participated in the study. Results: Test-retest differences were within ±2 dB for typical frequency bandwidths at insertion depths administered in clinical practice, and for up to 8 kHz at the experimental 30 mm insertion depth. The 28 mm insertion depth was the best predictor of ear canal levels measured at the 30 mm insertion depth. There was no effect of signal analysis bandwidth on accuracy or repeatability. Conclusions: Clinically feasible 28 mm probe tube insertions reliably measured up to 8 kHz and predicted intensities up to 10 kHz measured at the 30 mm insertion depth more accurately than did shallower insertion depths. Signal analysis bandwidth may not be an important clinical issue at least for one-third and one-twenty-fourth octave band analyses.