Gear wear and tooth root cracks are the most common faults in gearboxes and can occur simultaneously. Vibration-based techniques have been developed for gear crack detection, but without considering wear, which can affect gear dynamic responses and crack detection. To fill the gap, this work studies the different characteristics of wear and cracks, and the interplay between them, using degradation tests under both lubricated and non-lubricated conditions. The study employs several advanced techniques based on vibration analysis, transmission error (TE) measurements, and analysis of the gear tooth surfaces using moulding and image analysis. It demonstrates that cracks can generally be detected using ‘conventional TE’, while cyclostationary analysis of the vibration signal can reveal changes in the surface roughness of the gears, suggesting its potential for identifying gear wear mechanisms and enhancing wear monitoring and prediction strategies. More importantly, this work is the first to report that gear wear can potentially mask the presence of tooth root cracks. In part to address this, a novel TE-based method has been developed to measure the rate of compliance change caused by gear wear. Together with the previously developed TE method, this approach shows strong potential for effective, real-time gear condition monitoring by simultaneously assessing gear compliance—which reflects the current load-bearing capacity—and gear wear depth, which reflects the total amount of material removed.