I am looking for a free English data set to use for testing different ASR API's.
There are quite a few for money, but the only one that I know of that is free is the Buckeye Corpus. About 40 hours of sound and matching transcripts. You have to register, but there is no cost for non-commercial users.
do you know which ones (that may or may not cost money) are the easiest to use? I've had success with CMU_ARCTIC and LDC93S1 (TIMIT...) but am looking for more.– m81Jun 1, 2017 at 0:40
1I am not sure about ease of use, but the Linguistic Data Consortium has many data sets. Select the button for Catalog, then search. On the search window at lower right select "speech recognition" under recommended applications and if you want add a language restriction. You should get many options.– G5WJun 1, 2017 at 0:52