The “diphones data” refers to speech perception data from two unusually large perception experiments, which provided data on how listeners interpret phonetic cues as they hear the sound signal over time, in all sounds of Dutch and English, before and after all possible sounds of the language. Together, the two studies include approximately 1,000,000 phonetic identification responses from human listeners. This data can be used to answer speech perception questions and for modeling of spoken word recognition. We have made all of the data publicly available. To download the data, read the README, and find the references to published papers using this data, see the sections “Dutch diphones project” and “English diphones project” on the Douglass Phonetics Lab site. We are in the process of depositing the data and accompanying files at the University of Arizona Library Repository (as of summer 2023) so that it will remain available as web sites change, so check back for the UA Library archive information later.