I spawned this separate thread because  I don't want to distract from  Matt's original question.

 

But...  although  I am using neither NatLink nor Vocola at the moment, I am very interested in doing so.

 

I would like to understand the relationship  between Dragon, NatLink, and packages like Vocola and DragonFly and ...

 

I was under the impression that packages like Vocola interfaced with Dragon via NatLink.

 

Is this correct?

 

Is Vocola working irrespective of the NatLink error messages that Matt reports because it is not using  anything that interacts with those parts of NatLink (or DragonFly, as the  error messages suggest).

 

Or is Vocola  more independent of NatLink than I understood? In which case how does it use Dragon's recognition engine?

 

-------------------------

DPG15.6 (also DPI 15.3) + KB, Sennheiser MB Pro 1 UC ML, BTD 800 dongle, Windows 10 Pro, MS Surface Book 3, Intel Core i7-1065G7 CPU @ 1.3/1.5GHz (4 cores, 8 logical, GPU=NVIDIA Quadro RTX 3000 with Max-Q Design.

This article is a bit dated but it gives a nice flowchart

Introduction to Voice Programming, Part Two: Open Source Speech Tools (explosionduck.com)

 

Thanks, very nice diagram

-------------------------

DPG15.6 (also DPI 15.3) + KB, Sennheiser MB Pro 1 UC ML, BTD 800 dongle, Windows 10 Pro, MS Surface Book 3, Intel Core i7-1065G7 CPU @ 1.3/1.5GHz (4 cores, 8 logical, GPU=NVIDIA Quadro RTX 3000 with Max-Q Design.

Vocola 2 relies on NatLink. Vocola 3 relies on WSR.

 

Does "Relies on"  mean that Vocola 3 can only be used for WSR, not with NatLink?   

 

That would surprise me. I expected it is WSR in addition to NatLink.

-------------------------

DPG15.6 (also DPI 15.3) + KB, Sennheiser MB Pro 1 UC ML, BTD 800 dongle, Windows 10 Pro, MS Surface Book 3, Intel Core i7-1065G7 CPU @ 1.3/1.5GHz (4 cores, 8 logical, GPU=NVIDIA Quadro RTX 3000 with Max-Q Design.

 

They are 2 different systems despite the similar name (although the syntax of the languages are very similar).  Vocola 2 is built on natlink/python/dragon, whereas Vocola 3 is built on VB/WSR.  (Mark can correct me if I am wrong here).

Also despite the numerals, Vocola 2 is more up-to-date than Vocola 3.  (afaik Vocola 3 has not been updated in a very long time and neither has  WSR)

If you are using Dragon, Vocola 2 is what you want.

 

 

And since I would like to get away from Dragon, moving towards an open source recognition engine like Kaldi  or the Facebook recognition engine that Talon uses,  looks like I don't want to use either Vocola version.

 

-------------------------

DPG15.6 (also DPI 15.3) + KB, Sennheiser MB Pro 1 UC ML, BTD 800 dongle, Windows 10 Pro, MS Surface Book 3, Intel Core i7-1065G7 CPU @ 1.3/1.5GHz (4 cores, 8 logical, GPU=NVIDIA Quadro RTX 3000 with Max-Q Design.

yes I think you want to be fully open-source your stack needs to be kaldi, dragonfly, caster based. I heard that Mark was working on a Vocola to dragonfly transpiler, but don't think it's high-priority, so that might not be available for a while.