I want to know more about the latest attempts to create synthesizers for singing. I would ideally like to know more about both the most widely respected algorithms/models as well as mainstream software or programs that implement them (whether free or proprietary). Can someone offer me resources that would give me a decent grasp of the above?

I'm very new to this area of research so any suggestions for where to start would be welcomed.


  • Have you heard of "vocaloids"?
    – NReilingh
    Commented Dec 6, 2014 at 0:08
  • Yes. In many disciplines, people write articles summarizing key historical discoveries and inventions as well as latest developments and research efforts. I was hoping for something like that so I can understand what is out there more quickly. That was one of the topics I would hope to learn more about. Commented Dec 6, 2014 at 0:14
  • Ah, sure -- I'll bet there are some interesting peer-reviewed journal articles on the subject.
    – NReilingh
    Commented Dec 6, 2014 at 0:22
  • Ideally, I want stuff I can make music with. To my knowledge, no singing synthesizer exists that sounds anything like a real person (eg Selena Gomez) singing. If such a thing exists, I'd love to know. In the absence of that, I'd like to know what researchers and engineers see as the biggest impediments to creating such a synthesizer. Commented Dec 6, 2014 at 0:23

3 Answers 3


Since Max Matthews pioneering piece Daisy, singing synthesis has been informed by research in speech synthesis and phonetics. A voice may be modeled as an excitation (the glottis) filtered by a resonant cavity, which can be done by analysing recorded vocal sounds with linear predictive coding and then regenerating the sound with independent control of pitch and timbre.

It is crucial to be able to create the right spectral formants for each vowel, and hopefully also make the transitions between vowels or phonemes sound convincing. There are a few methods of formant synthesis, such as FOF. Diphone synthesis is a method concerned with the transitions between steady state segments. An early method of speech synthesis was to use short fragments of recorded spoken words and automatically fit pieces together into words and sentences. This kind of concatenative synthesis has found renewed interest in the last 15 years with the work of Diemo Schwartz, Bob Sturm and others. By transposing the segments to the required pitches it would be easy to use concatenative synthesis to simulate a singing voice. Johan Sundberg has done extensive research on the singing voice.

For an advanced speech synthesis software, you might have a look at Praat, but be warned that it has a steep learning curve and is not made specifically for musical purposes.


Vocaloid has already been mentioned and it's the most popular solution to date, although it's still quite far from a real voice. There are other apps like VocalWriter, quite old but interesting, or Text to Sing. Lately Realitone released Blue, not a full vocal synthesizer per se, but it includes a word builder and sounds pretty great. You may also want to check SING by Komponant. It's the technology I'm working on and it aims at creating the first virtual vocalist that actually sounds human. It's still in early stage! Demo: www.komponant.com

  • komponant's site moved to emvoiceapp.com
    – rraallvv
    Commented Oct 3, 2017 at 21:10
  • Vocaloid is arguably so far away from a real voice that each Vocaloid has become a distinctive artist. I've often read stuff like "Senbonzakura is a Hatsune Miku song", while Hatsune Miku is a Vocaloid and the source never mentions the real composer(s).
    – Dekkadeci
    Commented Oct 4, 2017 at 15:12

NeoSpeech is also working on singing text-to-speech voices. They have posted some samples for Japanese, but they will be adding more languages. There isn't a commercially available tool or application yet.

They are trying to get more natural sounding singing text-to-speech voices through the use of HMM-based text-to-speech as explained in this post.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.