This study examined (a) the validity of two accelerometers (ActiGraph GT3X [ActiGraph LLC, Pensacola, FL, USA] and activPAL [PAL Technologies Ltd., Glasgow, Scotland]) for the assessment of sedentary behavior; and (b) the variations in assessment accuracy by setting minimum sedentary bout durations against a proxy for direct observation using an automated wearable camera in free-living environments. Time spent in sedentary behavior estimated from the GT3X, with varying identification methods (i.e., a hybrid machine learning algorithm, Sojourn, and activity count thresholds), and the activPAL were compared to the criterion measure with and without applying minimum sedentary bout durations. The activPAL assessed sedentary behavior most accurately followed by GT3X with the "Sojourn" method. The performance of the GT3X is improved when identifying sedentary bouts = 15 min with the "Sojourn" method or a threshold of < 150 cpm. The activPAL should be considered for studies focusing on sedentary behavior. The ability of the GT3X to assess sedentary behavior is improved when focusing on sedentary bouts = 15 min using the Sojourn method or the threshold of < 150 cpm.