MBROLA and eSpeak NG are two speech synthesizers that can be used as MIDI instruments. MBROLA has been often uses for singing synthesis, because it allows you to control timing and pitch via its text interface. It became free software in 2018. Before 2018 I was already listening to a lot of VOCALOID and UTAU music, and I began researching how to implement my own singing speech synthesizer by reading "An introduction to text-to-speech synthesis" by Thierry Dutoit (author of MBROLA) and many other papers related to VOCALOID and UTAU. With a deep understanding how the MBROLA algorithm works I began to implement my own independant singing voice synthesizer, with eSpeak as an optional frontend.