NVDA add-on
Experimental Addon: eSpeakVoiceDesigner_0.2.43
20:19
Manolo
Hi everyone,
I would like to share a new experimental NVDA add-on:
eSpeak Voice Designer 0.2.43
The purpose of this add-on is to make the creation and editing of eSpeak voice variants much more interactive.
Instead of repeatedly editing a text file, reloading the voice and testing it again, eSpeak Voice Designer lets you change voice parameters while NVDA is running and immediately hear the result.
You can start in several ways:
from the standard eSpeak voice;
from one of the factory eSpeak variants;
from one of your previously saved personal variants.
So loading an existing variant is only one possibility. You can also start from the standard voice and gradually build a completely new variant.
The designer exposes many parameters used by eSpeak variants, including formants, pitch-related parameters, voicing, consonants, echo, breath and breath weighting, together with a number of Klatt-related parameters for experimentation.
Changes are applied live, so while designing a voice you can continue navigating with NVDA and immediately hear how the voice changes.
Personal variants can be saved as work in progress and reopened later, so it is possible to design a voice over several sessions rather than completing everything at once.
Some useful shortcuts are:
Alt+S:
Move to the Sections control.
Alt+V:
Move between the Factory variants and Personal variants controls.
Alt+L:
Move to the parameter list.
Ctrl+S:
Save the current voice as a personal variant.
F2:
Rename a personal variant.
Delete:
Delete a personal variant.
Ctrl+Shift+S:
Install the current variant into NVDA's eSpeak variant directory.
The last command is intended for when you are satisfied with the voice and want to use it as a normal eSpeak variant in NVDA.
Because the eSpeak data directory is normally protected by Windows, administrator permission may be requested during this operation.
NVDA does not immediately refresh the list of installed eSpeak variants. After installing a new variant with Ctrl+Shift+S, NVDA therefore needs to be restarted before the new variant becomes available in the normal eSpeak voice settings.
The general workflow can therefore be something like:
standard voice -> edit parameters -> save personal variant -> continue working on it later -> install it into NVDA -> restart NVDA -> select it as a normal eSpeak variant.
Existing eSpeak variants can also be used as starting points.
One important goal of the current version has been preserving the original behaviour of loaded eSpeak variants. Parameters which have not been changed by the user should remain untouched rather than being unnecessarily reconstructed by the designer.
This was particularly important for parameters such as breath, breathw and formants, where small changes in interpretation can noticeably alter the character of a voice.
The add-on interface is currently available in English, Italian, Spanish and French.
This is still an experimental project, so feedback is very welcome, especially from people who already create or modify eSpeak variants manually.
I am particularly interested in hearing about variants which contain unusual combinations of formant, breath, voicing or Klatt parameters.
drive.google.com/uc?id=1kJIHy4…
Best regards,
Lo-lo
#nvda #espeak
aman
in reply to garo • • •garo
in reply to aman • • •aryan
in reply to garo • • •Jonathan
in reply to aryan • • •garo
in reply to aryan • • •