[bug]: Non-Latin characters are rendered as "???"

[ramen]

Operating System

Windows

Other Operating System

No response

Architecture

amd64

ReaSpeech Image

reaspeech (CPU)

What were you trying to do?

This occurs when processing audio in non-English languages that cannot be represented using the Latin character set, such as Russian, Japanese, etc.

What actually happened?

Characters outside of the Latin character set are displayed as "???" due to a limitation in ReaImGui: cfillion/reaimgui#5

When writing to the Reaper project (create markers, create regions, etc.) the original characters are preserved. This issue mainly affects the table view within the ReaSpeech UI.

This is a non-trivial problem that may require changes to dear-imgui: cfillion/reaimgui#5 (comment)

Steps to Reproduce

1. Load a Reaper project with media items containing non-English speech audio
2. Process the media items in ReaSpeech
3. Browse the resulting transcript table

[ramen]

A few workarounds we've considered:

• Rendering the text using Reaper's "gfx" library, and capturing the output as an image (so far, we haven't gotten this to work)
• Rendering the images from Python and providing an API endpoint to retrieve images of transcript text
• Using a different UI toolkit besides ReaImGui - none of the other options have a comparable table widget, though a table view may not be an absolute requirement