(This also modifies image-source, obs-text, text-freetype2, and UI)
This improves source definition versioning. To do this, it now stores
two identifier names. One "unversioned" which is the original name, and
one "versioned" with the version number appended.
This fixes both backward compatibility with older OBS versions, and
fixes the inability to use "add existing" in OBS itself on sources
created from older version definitions.
Fix the image files not being freed, and remove the unused source
definition function 'update', which was additionally calling
image_source_load with the wrong number of parameters.
With the "instant-replay.lua" script, it would only play the last replay
buffer saved, meaning the user would have to save a replay with the
replay buffer hotkey, and then press the "Play Last Replay" hotkey to
play the last replay buffer.
Instead, combine the two and make the script save the replay
automatically first, and then play it back as soon as it's available.
This makes it actually qualify as an "instant replay" script, as its
name implies.
Currently the captioning code is a bit intertwined with the UI, and the
captioning is hard-coded towards microsoft speech API.
This patch abstracts captions to allow other APIs to be implemented
later.
Allows generating captions via the windows speech recognition API
(SAPI).
This is currently marked as experimental due to speech recognition
technology still being less than ideal. Speech recognition technology
in general is probably never going to be anywhere near perfect.
Microsoft's speech recognition in particular requires a bit of training
via the windows speech recognition tool to ensure it can dictate better.
Clear speech with a good mic is recognized fairly well, but casual
speech and/or speaking with a poor microphone will have some significant
issues. Captions can often be way off when speaking casually rather
than with clear diction.
