After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 626197 - Orca often fails to associate the correct voice with the correct string when multiple voices are present in an utterance
Orca often fails to associate the correct voice with the correct string when ...
Status: RESOLVED FIXED
Product: orca
Classification: Applications
Component: speech
unspecified
Other All
: Normal normal
: 2.32.0
Assigned To: Joanmarie Diggs (IRC: joanie)
Orca Maintainers
Depends on:
Blocks: 543157
 
 
Reported: 2010-08-06 12:34 UTC by Joanmarie Diggs (IRC: joanie)
Modified: 2010-09-20 10:57 UTC
See Also:
GNOME target: ---
GNOME version: ---


Description Joanmarie Diggs (IRC: joanie) 2010-08-06 12:34:50 UTC 
In the current Orca code, typically there is at most one voice being used to speak a given utterance. When multiple voices (i.e. default, link, capitalization) are needed to present a given string, the string tends to be passed in as separate utterances.

Orca's relatively recent formatting strings make it possible to create a single, functional utterance in which multiple voices are used. Thus we could have a voice for roles, another voice for state info, a third voice for position info, a fourth voice for tutorial info, and so on. Unfortunately, when this is attempted, Orca regularly fails to do the correct thing and instead speaks many of the strings using the wrong voice.
Comment 1 Joanmarie Diggs (IRC: joanie) 2010-08-06 12:38:40 UTC 
This, btw, is going to make it rather challenging to fix bug 543157. :-(
Comment 2 Joanmarie Diggs (IRC: joanie) 2010-08-06 12:50:17 UTC 
http://git.gnome.org/browse/orca/commit/?id=27044564cd4b2e7abd1d77d39f04505a36bb76b8

Ultimately, we should clean up all the speech code. In the meantime, this gets things working properly.

Fully regression tested. We can't really add a new test for this until we have multiple voices. Perhaps bug 543157 would be a good place for such a new test.