 |
Is anyone already working on this, or planning to? If so, I'm willing to help. If not, you can assign this to me. |
 |
Please proceed. You aren't currently listed as a developer so I cannot assign the ticket to you but I can log your submission and review it. Taaaki or Greebo are the ones who could grant you ticket ownership |
|
|
Thanks. Submit & review would probably be fine. I'll be starting with a .subs file for The Thug character. I think I understand how to prep a .subs file. How I would comprehensively test all its phrases is less clear. Maybe I need to create another ticket for DR, to show a selected soundfile's subtitle (using the specified TDM path). Unless there's some existing tool or method you know of. |
|
|
I am unsure how to test all the subtitles. Probably needs a test map with a script to provoke each bark? |
|
|
That's OK, I'll research it. The Thug, as a representative character, has over 400 sound files, so clearly making this process easy is important. Maybe the script can express sounds as the player rather than the AI. If I can avoid manually looking up the names of a 100+ existing soundshaders, that would be great. I'll also work up a suitable DR subtitles request. |
 |
Thanks for taking this up. I think the most economical approach would be to obtain a list of all .ogg files in the voices sound folder (i.e. via a command line tool), convert that list into a series of new, single-sound soundshaders (i.e. via Excel, I can assist with this) (can't really use existing soundshaders because they often randomly pick from several .ogg's), also convert that list into a series of spawnargs (i.e. "snd_test173" "test_thug_alert_to_idle", where the latter is the name of the test soundshader) and paste those into an entity in a .map file. A script can systematically look up those sound spawnargs and play their sounds. You can find a working example of such a script in the subs3.map that I uploaded to the A New Job subtitles ticket. |
|
|
I pursued a similar approach, but using a standardized soundshader naming with embedded index, so didn't need to involve spawnargs. See more about the resulting test FM at: https://forums.thedarkmod.com/index.php?/topic/21740-english-subtitles-for-ai-barks/&do=findComment&comment=482682 As for making the AI-specific file from a directory dump, there are various ways to generate the incrementing line numbers needed, and combine them with the .ogg directory dump. Your idea to use Excel sounds promising in a Windows context. (Otherwise, Linux command tools, or - going the fully automatic route - a small C++ program.) This is my current file format (subject to change), showing here the first few voice samples from TDM's .ogg dump (with alphabetic ordering from Windows "dir /b" > dump.txt"). --- start of file "fm_ai_thug_shaders.sndshd" // enumerate starting with zero; special sound for start of list, used with all AI voice samples. It shows subtitle: [start of subtitles list being tested] fm_ai_thug_shader0 { sound/subtitle_list_start_hit_high01.ogg } fm_ai_thug_shader1 { sound/voices/thug/alarm01.ogg } fm_ai_thug_shader2 { sound/voices/thug/alarm02.ogg } fm_ai_thug_shader3 { sound/voices/thug/bugger01.ogg } fm_ai_thug_shader4 { sound/voices/thug/get_over_here.ogg } fm_ai_thug_shader5 { sound/voices/thug/give_order01.ogg } fm_ai_thug_shader6 { sound/voices/thug/give_order02.ogg } fm_ai_thug_shader7 { sound/voices/thug/give_order03.ogg } fm_ai_thug_shader8 { sound/voices/thug/help_search01.ogg } ... } --- end of file If you'd like to work out an Excel recipe for getting that results for the repetitive part (starting with "1" for the first .ogg file), that would be appreciated. I noticed that the "alphabetic" order created by "dir /b" is different than what win11 shows in File Explorer, regarding ordering of underlines. Code-point versus "natural" alphabetizing? I haven't decided yet whether it's a good or bad idea to use character specific names for the shaders (fm_ai_thug_shader<n>) as opposed to generic (fm_subtitle_test_shader<n>). But easy to change this, since they are just fixed prefix strings. |
|
|
nbohr1more, please log and review this submission of subtitles for The Thug vocals. It is embedded in the FM "testSubtitlesThug": https://drive.google.com/file/d/1VNP2fNge-2Ff3RkdUvCLagRENUESVW40/view?usp=sharing Minor changes you can make directly (and let me know); if there's a need for sweeping changes, we can discuss. After review/revise, I will leave it to you to extract the subtitle list into its destination .subs file, rename things from "fm_..." to "tdm_..." and place it where you see fit. There are no srt files involved for The Thug, and the special .ogg file is only used for the test program,. This link is also being posted on the "English Subtitles for AI Barks" forum. |
|
|
Main Progress Thread: https://forums.thedarkmod.com/index.php?/topic/21740-english-subtitles-for-ai-barks/ Related discussion: https://forums.thedarkmod.com/index.php?/topic/21741-subtitles-possibilities-beyond-211/ |
|
|
Committed the bulk of the completed sets between Rev 16904 and 16921 Small fix-up in Rev 16922 SVN core structure: /darkmod/subtitles /darkmod/subtitles/tdm_root.subs ( has includes for all sub-definitions ) /darkmod/subtitles/tdm_ai_xxxx.subs ( naming convention for the subtitles sub-def for each AI ) /darkmod/subtitles/<ai_folder> ( example /darkmod/subtitles/simpleton/ - has all the srt files associates to each AI sub-def ) |
|
|
Rev 16923 "The Cynic" subtitles completed |
|
|
Final revision of the The Wench released, ready to integrate. With that, done: no further barks foreseen for 2.12 |
|
|
Rev 16924 "The Wench" subtitles added |
- Anonymous
