There is no such thing as
basic voice recognition. You need a full supporting system to recognise any word.
Voice recognition does not work on words, so you can't have a system that recognises a limited vocabulary.
To recognise the word "left", you need to recognise 4 phonemes, and then ascertain that those 4 phonemes consitute the word "left"; it could be numerous words based on the sounds recognised. There are 43 phonemes required for a full vocabulary, and your examples cover the full range of phonemes. For example the letter "H" is not one phoneme, it is many.