Re: [orca-list] Orca speaks extra dot character



Thanks for looking into this!

For me, I don't hear the dot with

espeak-ng -v en-us -m '<speak><tts:style field="punctuation" mode="none">autogen.sh.</tts:style></speak>'

espeak-ng -v en-us -m '<speak><tts:style field="punctuation" mode="some">autogen.sh.</tts:style></speak>'

speak-ng -v en-us -m '<speak><tts:style field="punctuation" mode="most">autogen.sh.</tts:style></speak>'

I'm using eSpeak NG text-to-speech: 1.49.2. So perhaps it's a regression?

Since I cannot reproduce the problem (and don't have time at the moment to build and install espeak-ng) it would be great if someone who can reproduce the problem could file the bug.

Thanks again!
--joanie

On 4/29/20 17:11, Samuel Thibault wrote:
Hello,

I'm indeed seeing the issue at the espeak-ng level, so it should be
reported to espeak-ng:

espeak-ng -v en-us -m '<speak><tts:style field="punctuation" mode="none">autogen.sh.</tts:style></speak>'

will speak “autogen dot sh dot”. With "some" we get the same result. with
"all" we get “autogen dot sh period”. With en-gb and en we get the same
series of result.


With the fr-fr voice, with "none" and "some" we get “autogen sh”, and
with "all" we get “autogen point sh point”.

So this really looks like a bug in the en-us language.


In the latest speech-dispatcher developments we have indeed included
some symbol support to optionally process punctuation etc. before the
synthesis backend. In speech dispatcher 0.9.1, uncommenting
DefaultSymbolsPreprocessing 1
should indeed be making some/none/all behave as expected. Log files with
LogLevel set to 5 would allow to have an idea why that's not happening.

Samuel




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]