draganratkovich/newfon

Multilingual speech synthesizer

newfon's Issues

Translate code responsible for number processing into english

@zstanecic it is your contribution.

Please make code to conform overal standard of an addon

Licenses mess

Need to fix the mess with declaring source code licenses:

1. The repository contains the text of the MIT license with unexpected copyrights. There is a suspicion that this file was copied from the project https://github.com/fongandrew/hydeout/ without any editing.
2. The .py files declare the GPL license, not an MIT license, which contradicts the statements in the newfon_license_*.txt files.
3. If you plan to license the speech driver under the MIT terms, this violates the NVDA's license (see detailed explanations below).

NVDA License Violation

The add-on's source code must be licensed under a license that is compatible with the NVDA's license.

• The NVDA core's source code is licensed under the GPL v2+
• The add-on's source code inherits classes of the NVDA core's source code.
• Inheritance creates derivative works. Also, NVDA's add-ons are considered derivative works, according to NVDA's license (this applies to any part of the add-on that uses functionality provided by NVDA such as functions, classes, and modules).
• The GPL requires that all derivative works be licensed as a whole under the terms of the GPL, an effect which can be described as "hereditary".
• The MIT license is not compatible with the GPL license.

Thus, the speech driver licensed under the MIT violates the NVDA's license.

Samples pair seconds haze issue and other Slavic languages.

In short, as it changes in the "Samples pair seconds haze" option, the depth of the voice and the speed of speech also change at the same time, so I think it should be separated.

We still need to add support for other Slavic languages - Czech, Slovakian, Belarusian, Serbian (Cyrillic), Macedonian and Slovenian -

and some sort of correct pronunciation of English and related languages,

add new voices, intonations and accents,

make Newfon available for the other Windows applications

and improve speech in terms of naturalness and quality.

For the latter, ESpeak NG's source code may be useful, because that synthesizer is still developing and improving more and more.

This would be quite enough for now; of course, can it be possible.

Thank you.

Where's the synthesizer code?

AFAICT this only contains the data files to translate English/Ukrainian/Polish to (approximate) Russian pronunciations. No code to synthesize the sound seems to be here... is it handled by some external library/program?