Fix SAPI4 prosody regressions#15502
Conversation
See test results for failed build of commit ee2ef6db67 |
|
Thanks @LeonarddeR I confirm this PR has been fixed #15500. |
|
I fixed one additional issue, a generator's bool value is always True. |
See test results for failed build of commit b57305ddca |
8f70ef1 to
a5f81e3
Compare
|
Hi @LeonarddeR Steps to reproduce
|
|
I've seen more errors like this. I guess we should just log these errors. |
4aee33d to
a682aa1
Compare
seanbudd
left a comment
There was a problem hiding this comment.
thanks @LeonarddeR, changes generally look good
| try: | ||
| self._ttsCentral.AudioReset() | ||
| except COMError: | ||
| log.error("Error cancelling speech", exc_info=True) |
There was a problem hiding this comment.
is there a reason why this logs at a different level to the pause changes
There was a problem hiding this comment.
Errors for pause just passed, I didn't want to introduce logging at the error level for this, though it made sense to me to at least something. Cancel, on the other hand, didn't catch any errors, so it made sense to log them at the error level.
…ce (#15583) Fixes #15580 Follow up of #15502 Summary of the issue: In #15502, I fixed some regressions regarding capital pitch changes introduced in #15271. This meant that we'd add prosody commands to every speech sequence to support a specific version of IBM TTS that seemed to need that. However, it introduced a regression in that it was no longer possible to change speech parameters using the gui. Description of user facing changes It is again possible to change SAPI4 parameters using the GUI. Description of development approach Only add blank prosody commands when there is any prosody in the sequence. IF not, don't do anything.
CC @cary-rowen
Link to issue number:
Fixes #15500
Follow up of #15271
Summary of the issue:
Some SAPI4 synthesizers reset all their prosody values to their defaults when they come across prosody in a sequence.
Description of user facing changes
Capital pitch changes no longer cause some SAPI4 voices to reset their rate and volume.
Description of development approach
When a sequence contains any prosody command, for the other prosody commands supported by the synth, we add commands to set them to the currently configured value.
Testing strategy:
Tested str from #15500.
Tested a sequence with a prosody command in the middle of a sequence, ensure that the prosody change is reflected in the speech.
Known issues with pull request:
None known
Code Review Checklist: