AQ - How does SAPI Output Work with the Latest Windows Build?

[Elliott Barnes]

Hi everyone.

I originally posted the below question to the AudioGames.Net Forum in
June last year; Matthew's message to this list reminded me that I
still don't know how this works! I've pasted the question here if
anybody can offer any advice.

Thanks.

Hi everyone. I realise that this may be better off in the AGRIP Google
group directly, but it doesn't appear that there has been activity for
several years so thought I'd ask here instead.
With the latest beta of AQ posted in October 2014, the game has been
packaged as a single executable rather than using Perl in the
background. When a new game is started, the following command is run:
zquake-gl.exe -window +set sensitivity 0
However, I can't figure out how SAPI is asked to output text. When
running the above command in a standard Command Prompt window, the
game launches but without any speech and when taking a look at the
source for the Python launcher on GitHub, there are no references to
SAPI output whatsoever. Does anybody have any ideas?
Thanks!

[matatk]

Hello :-),

When re-writing the GUI in Python on the Mac I found that it worked best when the code checked for the output from ZQuake and called the speech synth directly. I anticipate doing the same on Windows, though there are some issues with pipes on Windows, so will see how that goes. I'm hoping to be able to simplify the code a bit there too.

Currently there's no speech output on Windows. I have just got to the point where it's very close to running fine without speech output on Windows, but there are a couple of tweaks to filenames that I need to make cross-platform. I have got the launcher GUI running fine though, and the engine compiles well with Visual Studio Community 2019.

incidentally: VS2019 provided a smooth upgrade to the project format, so it's really easy to compile on Windows now, but the requirement is that VS2019 needs to be used to compile it now, as opposed to an earlier version - it's free so hopefully that won't be an issue. The project hasn't been updated to require the latest Windows SDK, so I imagine it'll work on Windows 7, maybe even earlier, but I can only test it on Windows 10.

I am not sure how long the self-voicing on Windows will take me to do, but hopefully not too long. I will try to keep you updated at a reasonable pace :-).

best regards,

Matthew

[Francisco Del Roio]

Hello all,

Can you try to use this library to output speech under windows rather to use SAPI directly. You can even integrate it under the engine directly and throw away the other speech stuff.

Cheers,

[matatk]

Hi Fransisco, and thanks for your suggestion. Those libraries look pretty cool.

At the moment I'm concentrating on getting it running, pretty-much as it was before, so I won't be able to look into this for a little while, but the idea sounds great, and I have filed an issue over on GitHub to track this: https://github.com/matatk/agrip/issues/37

Cheers,

Matthew

[Elliott Barnes]

Hi everyone, thanks for your replies.

@Matthew - thanks for explaining this. Apologies for not clarifying -
I was referring to the Windows beta that was posted in October 2014 to
the AGRIP GitHub account at the time (this now appears to have been
removed). Can you by any chance recall how you went about SAPI output
with this build?

Thanks.

> --
> You received this message because you are subscribed to the Google Groups
> "AGRIP Project" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to agrip-projec...@googlegroups.com.
> To post to this group, send an email to agrip-...@googlegroups.com.
> Visit this group at https://groups.google.com/group/agrip-project.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/agrip-project/7ed3852c-5100-4559-b4c7-84f0189a8f1f%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

[matatk]

Aha, right, the 2014 ones were using a library called pyttsx, though development on that seems to've stopped (there is a successor, apparently, called pyttsx3, though I am going to be looking into the ones that Fransisco suggested, as they can interface with screen-readers, which seems very useful.