Allow specifying subtitles in tape files

[domust]

Is your feature request related to a problem? Please describe.
I have tried using VHS for technical demos of terminal apps and while the output is great, without additional context it gets very hard for the user to interpret what they are looking at. A good example is the README.md where each feature is prefaced with an explanation that the user is expected to read, because it can't be embedded directly into the output.

Describe the solution you'd like
I was looking for ways to embed subtitles into the output, but then the problem becomes of mapping subtitles with actual timestamps in the output. For example, something like, SetSubtitle or similar command which would show text at the bottom of the screen until set to something else would work great. This way subtitles could be stored in tape files directly.

Describe alternatives you've considered
I have tried writing subtitles into external files of various formats and then embedding them into output of VHS. It does not work will all output types, or at least I wasn't able to make it work.

[adtyapnda]

Hey, I'd like to take a crack at this. Wanted to sketch the approach first and get some input before writing code, since it touches the ffmpeg pipeline.

For syntax I was thinking a sticky subtitle that stays on screen until you change it or clear it. Something like:

Subtitle "Welcome to the app"
Type "hello"
Sleep 2s

Subtitle "Now we open the menu"
Ctrl+K
Sleep 1s

Subtitle ""

A few things I'm not sure about. Is Subtitle a better name than SetSubtitle? Since Set is used for options like Set FontSize, a top-level Subtitle feels closer to Type, Sleep, and Hide, but I'm happy either way. Should an empty string clear it, or do you want an explicit Subtitle Off or Hide Subtitle? And on styling, do you want Set SubtitleFontSize and friends in the first pass, or should I ship sensible defaults first and add knobs later?

For rendering, VHS already builds the output through FilterComplexBuilder in ffmpeg.go, and frames are numbered per framerate, so I can record each Subtitle call as a (startFrame, endFrame, text) tuple while the tape runs. At render time I'd add a WithSubtitles stage to the filter chain that emits one drawtext filter per subtitle with enable='between(t,start,end)', positioned at the bottom of the frame inside the padding area.

I'd lean on drawtext over generating an SRT or ASS file and using the subtitles filter, because it avoids a libass dependency, reuses the font and color options VHS already tracks, and works the same way across GIF, WebM, and MP4 outputs. The main things to watch out for are escaping :, \, and ' in subtitle text for the filter syntax, and picking a default font path. Since VHS already ships fonts via embed.go I'd reuse those.

Scope for a first PR would be the Subtitle command itself (token, lexer, parser, executor), the drawtext stage in FilterComplexBuilder gated so there's no overhead when no subtitles are set, a sensible default style (bottom-centered, semi-transparent background, existing font), and an example tape plus a README entry. Styling knobs could follow in a second PR.

Does this direction sound right? Happy to adjust before I open a PR.

[gtmax]

#727