I've fixed my docker image to get better results and run faster. Won't bore with the details but there was some audio that only got picked up during a medium pass, while large large pass was better quality. I just hated the missing audio but still wanted the better quality. I've put together a solution that defaults to the large model for quality, but also takes the medium audio for segments where large missed it entirely. Seems to be the best option, but scanning takes multiple passes. I'm rescaning all off BBAN and plan to post in chunks of 50 at a time soon. (I'm also trying to find my missing titles to try and be somewhat complete).
Post your JAV subtitle files here - JAV Subtitle Repository (JSP)★NOT A SUB REQUEST THREAD★
- Thread starter Eastboyza
- Start date
-
Akiba-Online is sponsored by FileJoker.
FileJoker is a required filehost for all new posts and content replies in the Direct Downloads subforums.
Failure to include FileJoker links for Direct Download posts will result in deletion of your posts or worse.
For more information see this thread.
Thanks for all your efforts, mate. I commend you on your efforts with the first 200 or so in the BBAN series.
As a fellow lesbian JAV afficionado, the BBAN & AUKG series are my go-to staple for good quality content. However, as someone who doesn't know a lick of Japanese, my enjoyment of this series has been somewhat impaired by the lack of availability of comprehensive English subtitles across these two series.
I've had a bit of a crack at trying Whisper to try to make my own subtitles, but what invariably results is the misappropriation of pretty much all of the pronouns whenever anyone speaks, eg. 'you' for 'me' or I and 'her' for 'him', and vice versa. Makes it soo confusing to decipher.
So I look forward with bated breath until when you publish the next lot.
Cheers!
Last edited:
Thanks. Glad I'm not the only one that want's to be able to understand the the dialog. I'm just trying to tweak solutions that were already in place and automate it. I definitely ran into alot of the she vs he. DASS-122 is an example of one I couldn't get right. Medium would mess up the pronouns and be less accurate and have HE/HIM all over the place. Large seem's to get the pronouns right, but would miss audio for large chunks on some titles.
I think mixing the two gives the best results I could automate...
Attachments
Yeah, the mixed (opposite) use of pronouns when spoken by characters, both about themselves as well as about others, is pretty much stock standard in every effort that I've made, hence my reluctance to publish them.
I never realised that changing the model_size would yield those kinds of variations. But my preference would be to have more dialogue picked up (albeit possibly inaccurate) than to have them be missed out altogether.
However, in addition to the issue of mixed-up pronouns, occasionally (about half the time) my efforts would result in subtitles which would contain certain artifacts which would fill up the whole screen and make the movie unwatchable, which means that I would have to scan all my results and then personally edit those which have them.
Notwithstanding these issues, and as brilliant a tool as Whisper is at the moment, artificial intelligence still cannot translate idioms across languages comfortably.
As a non-Japanese speaker, hearing them say, "I want to sleep" in the throes of passion threw me off at first, but i kinda get the gist of it now. However, I'm still grasping at straws to try to replace it with an English alternative which sounds like a normal expression, like maybe "I'm in bliss" or something. Maybe others (who know Japanese) can help here.
that's just a problem with translations. Most people prefer to have 10x or more subtitles for JAV out there than having that 10% but with proper pronouns.
There is probably potential for a Whisper model that weights under the hood to get better sex talk or pronouns but that would be very difficult.
This can be daunting. Characters often in JAV refer to themselves by their names.
Like yesterday sub, a character, "Hmm.. Yes, Hikaru wasn't born yet". She was referring to herself.
It's cute when the actresses are saying but it throws me off.
Sometimes I don't change it. Then sometimes I will "fixed" it by adding "I" or "I'm" or etc.
*I've come to realize if the subs are for me, such changes aren't really needed.
I fully agree but recently I have made myself starting to hold back lots of dialogue.
Very difficult for me since I want to know everything...
Another area where I recently started to hold back. I would hear not familiar words and research and found out what it mean. Then I would add that artifact in my subs to help those to under what was jus stated. For example, jengis, Jingisukan (ジンギスカン).
It was too much dialogue. So I began to go back and remove them.
I had to remind me. The subs are in general for me, so all of that is not necessary.
I don't use whisper but I am not surprise. Idiom in any language is difficult.
I want to sleep is usually, "I want to make love" or "I want to have sex"
I used Whisper to create these Subs for VENU-223 and VENU-237. Both are of the mother/aunt-son/nephew swap storyline variety. I attempted to clean them up a bit, and re-interpreted some of the meaningless/ "lewd-less" dialog. Again, I don't understand Japanese or Chinese so my re-interpretations might not be totally accurate but I try to match what is happening in the scene. Anyway, enjoy and let me know what you think.
It was a little difficult to remember which "son/nephew" was in the batter's box however!
It was a little difficult to remember which "son/nephew" was in the batter's box however!
BBAN Subs[001-050]
(Whisper Large model based should be a better then my earlier post)
(Whisper Large model based should be a better then my earlier post)
Heres some from my back pile of subs I've wanted to do out of the 28000+ subdump.
All Machine Translated and unedited, Most turned out pretty good from the quick look over I did when I made them.
apologies if some are reposts
Im also taking requests here since they aren't allowed in this thread.
All Machine Translated and unedited, Most turned out pretty good from the quick look over I did when I made them.
apologies if some are reposts
Im also taking requests here since they aren't allowed in this thread.
Last edited:
@Chuckie100 I'm wondering what are your whisper settings?ROE-152 To My Mother Who Is Getting Remarried... I'm Jealous Of My Unstoppable Creampie Incest Seika Toyosaki
I used Whisper to create this Sub for ROE-152 because I found a "Reduced Mosaic" version of the video that I though was pretty erotic. I attempted to clean it up a bit, and re-interpreted some of the meaningless/ "lewd-less" dialog. Again, I don't understand Japanese or Chinese so my re-interpretations might not be totally accurate but I try to match what is happening in the scene. Anyway, enjoy and let me know what you think.
model_size?
language?
translation_mode?
vad_threshold?
chunk_threshold?
max_attemps?
And what audio code do you use? e.g. aac / mp3
Your translations are superior
I use Whisper AI's default setting and I use the mp3 ripped from the video (using VideProc Converter). On occasion I will up the volume on the video, save the new video and then rip the audio from the hi-volume video into a MP3 file. I'll sometimes do this when it appears that the video is low volume where there is missing dialog. I've also tried adjusting the default Whisper setting but haven't had any luck with that.
I will also try to find a "machine translation" from SubtitleCat. Then I decide which sub I think is better and use it for the basis of the new sub and use the other sub to provide clarity and fill in missing dialog. The real effort is in editing the subs post Whisper AI and SubtitleCat download.
The general procedure I use for editing is to first open the srt file in Subtitle Edit and run the "fix Common Errors" in the Tool dropdown. Then run Spell check and thn save the sub. Next I open the video on the left-hand side of the monitor and then open the srt file using notepad on the right hand side of the monitor and then attempt to edit the sub to match what is happening and who is talking. Correct pronoun usage and correcting for two many "I'm sorry", "I'm sleepy" and "comfortable" are examples where I feel need editing. It's by no means perfect, but what do you want for free?
Last edited:
AED-194 Incest Chisato Shoda Creampie Vagina For Fifty Mom
I found another mother-son incest video that I couldn't pass up. I used Whisper to create this Sub for AED-194 starring one of my favorite MILFs, Chisato Shoda. I attempted to clean it up a bit, and re-interpreted some of the meaningless/ "lewd-less" dialog. Again, I don't understand Japanese or Chinese so my re-interpretations might not be totally accurate but I try to match what is happening in the scene. Anyway, enjoy and let me know what you think.
Good morning fella pervs!!!
I have some good news.
This week my new(ish) girlfriend has started help me with some new subtitles.
I'll have to come up with a screen name for her as I should give her co-writer or writer-assistant credit for the kinky subs we end up sharing.
Creating subtitles for mom-son incest movies while getting head from your insatiable, just as kinky girlfriend is a dream come true!
I feel like I won the pervert lottery!
She is into BDSM, D/s, rough sex, cuckold themes and of course, Mom-son incest.
She's a kink lovers dream. (Except her tits are only a 36-C cup. I have always consider a D-cup to be a Titty Starter Kit.
Since I was 20, I've dated DD or bigger (up to an M-cup), but a guy can't have everything.
If you have any requests that you'd like me to get her input on, let me know.
Cheers
This is a great film. From what I can gather, Yu and her husband are at his families' reunion of some sort. Yu, does not care for her in-laws except one, a brother-in-law perhaps. Yu and the "Brother-in-Law" seem to energetically rekindle an affair they were having with an almost reckless abandon. An "Uncle" sees them and uses the information to force himself upon her as the "Brother-in-Law" runs away. Later at the last night of the reunion, Yu and her husband get very drunk. The husband passes out while a very inebriated Yu is led into a back bedroom by the "Uncle." He takes advantage of her in a well choreographed encounter. Then all the other male relatives take their turn with her, including the "Brother-in-Law." The ending shocks Yu and her husband. An english subtitle for this film would be great. Thanks for listening.
quantity over quality. AI stuff is wrong even on easy lines. As mentioned, even pronouns are wrong. In japanese, they are often omitted. In those cases, only a human can get who is speaking.
Let's not forget parts when there is some background noise and/or more than one person talking, AI gets lost ans miss a lot of lines.
VENX-166 After Taking An Aphrodisiac, Her Sensitivity Is 100 Times Higher. Mother And Son Go Crazy For Rich Creampie Sex Chisato Shoda
I found another mother-son incest video that I couldn't pass up. I used Whisper to create this Sub for VENX-166 starring one of my favorite MILFs, Chisato Shoda. I seem to be fixated on her lately...arguably not the most beautiful JAV actress or the best body out there, but she rocks my boat! I attempted to clean it up a bit, and re-interpreted some of the meaningless/ "lewd-less" dialog. Again, I don't understand Japanese or Chinese so my re-interpretations might not be totally accurate but I try to match what is happening in the scene. Anyway, enjoy and let me know what you think.
Yeah. The background voice thing is especially a problem due to how long these videos can be. The Whisper settings for some parts of the video may be bad at other parts.
I hope Whisper 2 has some better VAD implementation on this kind of thing. The ctranslate/Faster Whisper implementations are better but still not good for some 3 hour piece of cinema that has taunting in the background while some woman closer to the mic moans loudly.
