Voice-onset time. The phonemes and both consist of a plosive (air escaping when the lips open) and voicing (vocal cords vibrating to make the ah sound). In /ba/, the plosive and voicing are nearly simultaneous — the voice-onset-time is about 0 ms. In /pa/, the voicing occurs after the plosive — the voice-onset time is about 50 ms. You can see this in the waveforms or the spectrograms.
These waveforms, which depict change in amplitude over time, show the repeated oscillation of the voiced sound in both the /ba/ and the /pa/. The plosive appears most clearly in the /pa/, where it precedes the voicing. These spectrograms, which depict change in frequency over time, clearly show harmonics of the voiced sound in both the /ba/ and the /pa/. The noisy plosive appears most clearly in the /pa/, where it precedes the voicing.