ALIZARIN Engine
The reusable framework and supporting development tools. The public repository documents the framework as LGPLv3-licensed.
In development.
Framework LicenseVoice synthesis / Open-source R&D
ALIZARIN Engine
Building Voices That Never Belonged to a Human.
What if a virtual character could have a voice that was designed from the ground up, rather than copied from an existing performer? ALIZARIN is my experiment in doing exactly that.
It's an original synthetic voice project exploring algorithmic sound generation, modern voice technology, and distinctive virtual voices without relying on unauthorized voice cloning.
Experimental voicebank beta available by inquiry.
Framework in development.
What is ALIZARIN?
A Different Way to Give Characters a Voice.
Many modern voice synthesis systems start with recordings of a human performer. ALIZARIN explores another path: constructing an entirely synthetic vocal source using computational methods, then developing that sound into a useful character voice.
Rather than copying an existing singer or voice actor, the goal is to design a new vocal identity with its own unusual characteristics. That creates opportunities for virtual singers, synthetic characters, speech systems, and experiments in expressive audio.
- Creator
- Vega Lyrae
- Umbrella project
- Constellation Virtual Media
- Primary focus
- Synthetic voice technology
- Development model
- Public, open-source framework development
- Voicebank status
- Experimental beta access by inquiry
A different starting point
No Human Voice Hidden Inside.
I like virtual voices that sound like they belong to their own characters. That led me to explore making the source sound itself, instead of recording and imitating a human performer.
ALIZARIN explores how older synthesis techniques and modern machine learning might work together. The goal isn't to replace voice actors. It's to make another creative option available.
The public README acknowledges missile_39's Adachi Rei as an inspiration. Synthetic origins do not remove every licensing or misuse concern; responsible use and clear terms still matter.
High-level technology overview
From Synthetic Sound to a Virtual Voice.
These are the project's publicly documented research areas, not a claim that every application has shipped.
- 01 / Source sound
Algorithmic Sound Generation
ALIZARIN explores creating its initial sound from computational processes, including formant-inspired synthesis and layered sonic textures. Formants are resonances that help shape the character of a vocal sound.
- 02 / Voice experiments
Machine Learning and Voice Modeling
The project investigates how synthetic sound can become more expressive and practical voice models, combining conventional synthesis concepts with newer machine learning techniques.
- 03 / Creative applications
Speech and Singing Applications
The long-term framework is being designed around multiple ways to use a synthetic voice, including speech generation and vocal music production. Compatibility and individual capabilities remain in development.
A conceptual development workflow, not a completed production pipeline. See the repository for public development information.
Engine, voicebank & character
One Project, Different Kinds of Creative Work.
ALIZARIN Voicebank
The experimental primary synthetic voice. Beta access is available by inquiry, and the voice data has separate licensing and distribution terms.
Ask about current access and intended use.
Voicebank InquiriesALIZARIN Character
The character identity associated with the primary voice. Its name, imagery, and creative assets have separate intellectual-property considerations.
Contact Vega about permissions.
Ask About Your UseThe open-source framework, voicebank files, and character assets are not all distributed under the same terms. Review the applicable terms before using or distributing project materials. Some wording in the voice and character usage guidance needs clarification; contact Vega with questions about your proposed use.
Current development
Where the Experiment Stands.
Available for inquiry
Experimental voicebank beta
Interested testers can contact me to discuss the current beta and whether it fits their intended use. Access is handled individually.
In development
ALIZARIN Engine framework
Ongoing work toward a reusable public synthesis framework, including synthetic source generation, voice modeling, tooling, and documentation.
Research & future goals
Additional speech and singing applications
Broader compatibility, more accessible tools, and further creative applications are longer-term goals, without fixed release dates.
The README currently describes Phase 1 foundation work. Experimental voicebank beta availability reflects my current update; the repository's status documentation still needs to be brought into alignment.
Beta voicebank inquiries
Interested in Testing ALIZARIN?
I'm looking for people interested in experimenting with ALIZARIN's early voicebank and helping me improve it.
Whether you work with vocal synthesis, make music, experiment with audio software, or simply enjoy unusual synthetic voices, I'd be interested in hearing what you want to try. Feedback about sound, usability, limitations, and compatibility is especially valuable.
Request Beta AccessALIZARIN Beta Application
Required fields are marked with an asterisk. Beginners are welcome. This application requests consideration, not automatic approval or access to files.
Standard submission uses the same security check and opens Formspree's confirmation page. If the security check is unavailable, please use the email inquiry link.
Application Received
Thanks for your interest in testing ALIZARIN. Your application has been submitted for review.
If your testing interests fit the current beta, I'll get in touch using the contact information you provided.
Prefer email, or having trouble submitting? Send a beta inquiry to [email protected].
Open-source development & collaboration
Help Build the Engine.
ALIZARIN's framework is being developed publicly, and I welcome interest from developers, musicians, researchers, and creators who want to explore synthetic voice technology.
There are plenty of challenges in building a practical voice system from unusual source material, from synthesis and audio processing to tooling, compatibility, and documentation. If that sounds interesting, take a look at the repository or get in touch.
Where participation could help
- Audio DSP and synthesis
- Speech and singing technology
- Machine learning
- Audio application development
- Voicebank testing
- Documentation and usability
- Linguistic testing and pronunciation
A few useful answers
Before You Start Experimenting.
Is ALIZARIN a voice clone?
No. The project's primary concept is to design a synthetic vocal identity from computationally generated source sound rather than imitating a particular human performer.
Can I try the ALIZARIN voicebank?
An experimental beta is available by inquiry. Contact Vega with your intended use and testing interests.
Is the entire project open-source?
The ALIZARIN Engine framework is developed as an open-source project under LGPLv3. The primary voicebank and character assets have separate terms.
Can I use ALIZARIN in music?
Music production is an intended use of the voicebank. Review the applicable voicebank terms and contact Vega with questions about your proposed use.
Does ALIZARIN already support speech and singing?
Speech and singing are central development goals. Individual tools and voicebank capabilities are at different stages, so confirm current compatibility before planning a project around them.
Can developers contribute?
Yes. The public GitHub repository provides development information and contribution guidance.
More experiments ahead
Want to Hear What Comes Next?
ALIZARIN is still taking shape, and I'm interested in hearing from creators and developers who want to explore what synthetic voices can become. Whether you're interested in beta testing, contributing, or collaborating, feel free to reach out.