We can hear and monitor different aspects of our speech, including our formants which convey vowel identity. If what we hear does not match our intended speech, we tend to compensate to ‘fix’ the speech we produce. In a lab setting, this compensation response has been investigated using two different perturbation scenarios. In one scenario, the perturbation is introduced unexpectedly and temporarily, which can be introduced at utterance onset or transiently at mid-utterance. These perturbations cause speakers to have a transient compensation response. In another scenario, if this ‘unexpected’ perturbation is constantly repeated, the perturbation becomes expected and an adaptation response will likely take place. Here we explored three questions: 1) the transient compensation responses to transient mid-utterance formant perturbation, 2) the relationship of (1) responses to formant perturbations initiated at utterance onset, and 3) whether this compensation response is crucial for driving for formant adaptation. We asked participants to phonate either the word ‘head’ or ‘hid’ for 1.5 seconds while we applied a real-time shift of the first formant frequency (F1) to participants’ utterance. The shift was applied either at utterance onset (whole-utterance), for only 400 ms after a jittered delay (mid-utterance), or constantly over several trials (constant shift). We found that participants do compensate for mid-utterance formant perturbations, and this response is highly correlated to their responses to whole-utterance formant perturbations. However, their transient compensation responses did not correlate with their adaptation responses, which supports the idea that different mechanisms drive transient compensation and adaptation.