Introduction: Five components were studied in 4 subtraction waveforms derived from ERPs obtained in passive and active conditions of a 3-tone oddball task (common=70%, C, 0.8 KHz; deviant=15%, D, 2 KHz; 1.4 KHz=15%, t, also used as a target (T)). These waveforms reflect different stimulus-mismatch processes and thus their topography could be revealing of different brain regions mediating them. Methods: The following mismatches were studied: stimulus-mismatch (deviant - common, D/C, rarity and pitch confounded, known as the mismatch negativity, MMN), pitch-mismatch (T - deviant, T/D, rarity not target features controlled, known as processing negativity PN), attention - mismatch (T - t), T/t, controlled for pitch and rarity to show the influence of target features, known as the Negative difference Nd). These are compared with Goodin's procedure (G-wv, (T-common (active) - (t-common (passive)- the "Goodin-waveform"). Results: There were main site effects in normalized data in all cases (not P2 and N2 latency). There were separate frontal and posterior contributions to P1, with the former emphasized where target comparisons were involved. Frontal N1 peaks, largest in D/C (MMN), spread posterior and to the right where target matching was involved. P2 posterior maxima were also less localized where target features were involved in the comparison. N2 topography was similar between waveforms but spread slightly more to each side in the T/t comparison (i.e. Nd). Onset was earlier in the D/C comparison (i.e MMN). Parietal P3 peaks in waves based on target-ERPs showed a left temporal shift (vs D/C), though in T/D P3 was in fact maximal on the right (i.e. PN waveform). Conclusions: Thus an attentional effect(controlled processing) is evident as early as 60 ms. Target features modify the anteroposterior distribution of positivity and negativity for the early components and in the lateralization of P3-like positivity. A comparison of waveforms by latency of potential shift (running t-test) vs. peak identification (MANOVA) is illustrated and discussed. D/C (MMN) and T/t (Nd) waveforms, rather than T/D or G-wv (PN & Goodin waveforms) waveforms are recommended for distinguishing comparator mechanisms for stimulus- and task-relevant features.