Support Free or Commercially-Available Formant Synthesis

Originator:listsebby
Number:rdar://20419433 Date Originated:03-Apr-2015 08:38 PM
Status:Open Resolved:No
Product:Accessibility Product Version:iOS and OS X
Classification:Feature (New) Reproducible:Always
 
Summary:
For those that need or prefer it, please support the development and use of alternative formant synthesisers.

Steps to Reproduce:
The only voices shipped with or available for OS X are concatenative (except MacinTalk, which is less distinct than other synthesisers).

Expected Results:
Formant synthesis should be available for OS X from third-parties; developers should be able to make bindings for those that are already available as Open Source. Documentation for TTS authors should be available (not just for consumers of the speech platform for converting text into speech).

Actual Results:
There are no formant synthesisers available for the OS X speech platform. Some Open Source synthesisers are available, but they are not usable by applications on the platform because they are not integrated with the speech platform APIs.

Notes:
Eloquence, DECtalk and eSpeak are still popular on Windows; I prefer and use the former, even though it appears to be dying.

Configuration:
This was unsupported since the outset of OS X and iOS.

Comments

Sabahattin Gucukoglu, 19-Jun-2015 02:27 AM

I think we now agree that the documentation issue is solved, but there is still a dearth of rule-based synthesisers actually available on OS X, except for your own. I suppose one solution to this might be to make MacinTalk function better as a formant synthesiser, but I still don't see any actual alternative--nothing commercial, and no eSpeak or other OSS engine either (unless you know otherwise).

Would you consider making arrangements with a vendor to ship one (like Eloquence from Nuance) or make it available, or else binding to eSpeak? I think we need at least one robust pure rule-based engine for those occasions.

Apple Developer Relations, 19-Jun-2015 00:04 AM

There are already several third-party synthesizers that have been ported to the Speech Manager, including some open source. And several of the Macintalk voices are indeed formant synthesis. The process of porting a voice to OS X is well documented and supported. Is this what you are asking for?

Someone directed me to "MorseSynthesizer". That pretty much clears up the documentation requirement, assuming it's complete.


Please note: Reports posted here will not necessarily be seen by Apple. All problems should be submitted at bugreport.apple.com before they are posted here. Please only post information for Radars that you have filed yourself, and please do not include Apple confidential information in your posts. Thank you!