1) it depends on your usecase: assuming you want to create lipsynced anims from audio or text for a redkit project then you have to export the strings from the redkit project db once you are finished with writing the dialog(s). a menu option exists for this (though I don't remember where exactly it was). it will create a LocalEditorStringDataBaseW3_UTF8_mod_export.csv file in the root of the redkit folder. then this files needs to be copied into the lipsync project template root. see HOWTO.txt in the project template for subsequent steps.
2) no. the second package on this page (radish.lipsync.4.REDkit-v2024-11-29b.zip) already contains everything that is required from radish tools. just unzip into a directory of your choice and edit the _settings_.bat in your lipsync project so it points that directory.
@abilitydraft: as already written this tool does not help with the actual speech generation but only with the lipsynced animations for existing audio. Recording/creating decent quality speech for voiced mods is not as easy as it may seem. Although voiced mods are much nicer than non-voiced I would not expect a sudden surge of voiced mods.
However, although the primary goal of this tool is to provide support for higher quality lip synced animation for voiced mods it's also *very* easy to add just lip animations to non-voiced text-only mods. At least this way the characters will seem more alive in the silent scenes. A small but also a very easy to add improvement.
Exactly as rmemr said, making quality audio takes time. I got around 350 lines in my mod and 25 hours of doing just audio work got me like 50% done. This tool is an amazing help though.
every language requires its own audio phoneme and language model to extract timings from audio. unfortunately these are available only for a couple of languages out of the box. creating a new model is not trivial as it requires a high quality dataset to be usable. so the priority was to support the most common use cases for existing voiced mods. so far I haven't seen a custom mod with brazilian portuguese voicefiles.
as for creating lip animations from text for brazilian portuguese (and a couple of other languages): it *does* work (with some minor tweaks to the project template) but unless you actually want to lip read the text it's probably easier to just copy the animations from one of the other supported languages.
however if you really want to create lip animations for brazilian portuguese I can provide you the necessary info.
I exposed the set of languages that are scanned (and potentially generate lip animations) in the project template settings file. see comments and the updated HOWTO for caveats.
18 comments
1) Where can I find the strings csv? Do I need to cook the strings in REDkit once I'm done writing the dialogue lines?
2) Do I need to have the other radish tools installed?
2) no. the second package on this page (radish.lipsync.4.REDkit-v2024-11-29b.zip) already contains everything that is required from radish tools. just unzip into a directory of your choice and edit the _settings_.bat in your lipsync project so it points that directory.
However, although the primary goal of this tool is to provide support for higher quality lip synced animation for voiced mods it's also *very* easy to add just lip animations to non-voiced text-only mods. At least this way the characters will seem more alive in the silent scenes. A small but also a very easy to add improvement.
as for creating lip animations from text for brazilian portuguese (and a couple of other languages): it *does* work (with some minor tweaks to the project template) but unless you actually want to lip read the text it's probably easier to just copy the animations from one of the other supported languages.
however if you really want to create lip animations for brazilian portuguese I can provide you the necessary info.
Huge thank you for you dedicated job and sharing amazing tool.