Post your JAV subtitle files here - JAV Subtitle Repository (JSP)★NOT A SUB REQUEST THREAD★

If you look at the duration, every single sub seem to be at 1:29 .09x length. You can also see that the url pointers are incremented by constant values. I think it indicates that this vtt is a bogus or a dummy filler than a viable subtitle file.
Which makes me wonder how the streaming server picks up the correct subs at time of playback.
It is a type of html subtitle, the coordinates will point to the exact position of the text on the existing image.
 
  • Like
Reactions: HaesGwi and mei2
I hadn't considered GPTSubtitler.com. I just checked them out --it does look quite decent and transparent on the surface.
I have adopted this repo for for my script: chatgpt-subtitle-translator.

Getting the names and the pronouns right is a pain. Diarization is a good thinking. In case you're interested (and planning to use pyannote) there is a Japanese fine tuned version that imo is more decent than the stock version: transcribe_japanese_with_diarization.py
try this https://ai-subtitle-translator.netlify.app or this https://translator.mitsuko.web.id
 
  • Like
Reactions: mei2
Speaking of Manga inspired JAVs, have you seen the new URE-121 film coming soon with Jav Milf Ririko Kinoshita? Its also based on a Manga where an office wife is having an affair with a younger guy behind her husband's back. It's called "My relationship with Mrs. Fujita". I have not been able to find the Manga translated in English, but I did find a Spanish translation and it reads HOT! :p Release date is March 7. I love JAV MILFS! ;)

1741041982921.png1741042047459.png
 
  • Like
Reactions: Chuckie100

MTALL-144 Kana Yura - A college student, treated like a sex doll, is defiled in broad daylight when a man invades through her window and inserts his penis while she sleeps.​

MTALL-144-cover.jpg

My "100" lines test I expected to be fairly quick turned out to be way harder than expected. I figured there wouldn't be much dialog with a girl sleeping basically the whole movie but I should have figured the guy would be whispering the whole time and whisper is not great at catching low volume speech so I ended up having to cut 133 audio files with boosted audio I had to feed to whisper one by one to try and figure out what it first missed.

I used the Pro version of the whisperwithVAD colab with identical settings(the default) for both whisper versions and then used normal python whisper with the large-v2 model for the audio files I then manually extracted and amplified the audio of(10-20db).
The fully translated by whisper version is 109 lines.
The transcription only(Japanese text) version of whisper is 168 lines(I expected it would contain duplicates and hallucinations, but nope, just a handful of sound translated as speech and that's it).
The manually processed(aka me translating with AI and adding lines) whisper transcription ended up being 279 lines after removing the commented ones, so about triple what I expected, lol.

What I ended up doing is translating line by line with Gemini(the google AI) and the result was surprisingly good. The biggest problem by far was whisper either not picking up lines or transcribing them wrong. I had to sound it out in romaji to the AI myself for many of them and ask it to find me something close that made sense, which worked well most of the time.
There's a few lines that I had to guess more than I would have liked and some that were translated a bit awkwardly, but overall I'm very pleased with the result, I consider it good enough to be worth the effort even with my extremely limited Japanese knowledge.
The AI helps a lot for translating and speeds things up quite a bit compared to fully doing this manually with a low Japanese understanding.

Translating the whole movie ended up taking roughly 36 hours(not counting the initial whisper transcription) and another 4 hours to retime the whole subtitle file, do some light editing and quality check it.

If anyone is curious about comparing the results, MTALL-144_comparison_subs.zip contains the full whisper translation(MTALL-144_whisper.en.srt), the whisper transcription I used as the base(MTALL-144_whisper.jp.srt) and a .ass that contains the full whisper translation displayed on top of the video and the final manual translation displayed at the bottom(MTALL-144_comparison.ass).

For those who don't care and just want the sub, skip to here and read below:​

For the proper version, there's both an .srt and .ass version in MTALL-144_manual_subs.zip . The only difference is that the .ass contains some commented line I either decided not to display or couldn't figure out if they were dialog or gibberish and it also makes use of the actor function to display who is saying which lines(which doesn't matter at all for playback, only help keep track easier when editing).
 

Attachments

Last edited:
  • Like
Reactions: Scrapper and mei2
I used the Pro version of the whisperwithVAD colab with identical settings for both whisper versions and then used normal python whisper with the large-v2 model for the audio files I then manually extracted and amplified the audio of(10-20db).
The fully translated by whisper version is 109 lines.
The transcription only(Japanese text) version of whisper is 168 lines(I expected it would contain duplicates and hallucinations, but nope, just a handful of sound translated as speech and that's it).
The manually processed(aka me translation and adding lines) whisper transcription ended up being 279 lines minus the commented ones, so about triple what I expected, lol.

Thanks @SamKook , very good test report. I plan to look into it during the weekend.
Quick question: is your source from streaming sites, or sukebei? I plan to use the same source.
 
Quick question: is your source from streaming sites, or sukebei? I plan to use the same source.
I never use streaming, I'm already complaining about the quality of 6000 bitrate web releases so my eyes wouldn't be able to tolerate it.

It's the 1080p 4k2 version to be specific, not the hhd800 one(although they should be near identical since they're direct rips of the original), which both can be found on sukebei.
 
  • Like
Reactions: mei2 and Chuckie100
Question for all translators: do the subtitles respect the original dialogues and plot, or are they just invented dialogues? The thing is, the subtitles from a certain JAV site that ends in 'guru' seemed decent at first since they published two subtitles a week, but over time I realized that the translated dialogues were different from the original story and rather seemed like invented dialogues, which is sad because those subtitles told an invented story.
 
Question for all translators: do the subtitles respect the original dialogues and plot, or are they just invented dialogues?

About the invented stories, the one that you came across was not a fan-fiction by any chance, was it?

You are right, in general most of the sites are going for fast AI subbing which means a total miss, or, at best, a 50% miss in quality :)

In my experience, better quality subs usually show up three to four weeks after the movie release. I think the race to produce subs fast for sites to get more viewers has ruined the subs quality from sources that we used to enjoy. Even the Chinese subs are getting affected by that trend. I don't say "fast-subbing" is completely a bad thing; back in the day when I was coding WhisperJAV, almost all comments I received from "a" community were that people prefer faster production of subs over slow but higher quality subs.

Having said that, I think the quality continues to improve and there will be an acceptable level of base quality in some months; LLMs are helping that quite a bit. I am now testing a workflow (that I am trying to automate) which makes reasonably decent subs, but it takes about 3-4 hours to produce and costs around $2 per sub. There are so many good development works happening in the community. My guess is that in 6 months, we will have quality subs that are also fast produced.

At any rate the best subs continue to be AI + human. And hopefully not as high as 36 hours of human correcting and polishing the final output :)
 
  • Like
Reactions: Imscully
Question for all translators: do the subtitles respect the original dialogues and plot, or are they just invented dialogues? The thing is, the subtitles from a certain JAV site that ends in 'guru' seemed decent at first since they published two subtitles a week, but over time I realized that the translated dialogues were different from the original story and rather seemed like invented dialogues, which is sad because those subtitles told an invented story.
Imo they're good enough, and the highest quality out there from all the popular streaming sites.
But I do question sometimes some translations as well yeah, pretty sure the person does not speak Japanse either and I wonder what the source of his translations are? Is it just your default subs you find on subtitlecat (mostly done by whisperJAV I think) and he just improves upon them? SexTB does this as well and it's very noticeable. I feel like they put in 10-20 minutes of work, like fixing pronouns, and call it a day. While guru puts in a lot more work.

I did a quick test of NIMA-049, a recent one guru did. Just compare it yourself. This is a deepseek translation. Zero editing besides format. While there's some obvious mistakes and missing lines, the translation is just better. I absolutely love the "Of course, this vanilla stuff could never satisfy me." It clearly understands JAV lingo.
1741177400821.png

Now I don't speak Japanese either, like most of us we do understand some words and intonation.

Different topic now,
I've been testing out Gemini 2.0, since it's crazy cheap, way cheaper than deepseek even and waaaay faster. But the result are lackluster to be fair. I find that the translations are the same, just worded differently and they are harder to read and not that natural. I'm wondering if others have noticed the same thing ?
 

Attachments

"Guru" used to be great because they would rip off ERO-Japanese subs and repost them. But since ERO-Japanese's Red catalog basically shut down their source of mother-son themed Subs dried up. At least that's my opinion.
ooh that makes a lot of sense. I did notice there was a lot of quality difference in between. Sadly I was never able to find out about this Red catalog early enough. Missed out, huh ? :terharu:
 
About the invented stories, the one that you came across was not a fan-fiction by any chance, was it?

You are right, in general most of the sites are going for fast AI subbing which means a total miss, or, at best, a 50% miss in quality :)

In my experience, better quality subs usually show up three to four weeks after the movie release. I think the race to produce subs fast for sites to get more viewers has ruined the subs quality from sources that we used to enjoy. Even the Chinese subs are getting affected by that trend. I don't say "fast-subbing" is completely a bad thing; back in the day when I was coding WhisperJAV, almost all comments I received from "a" community were that people prefer faster production of subs over slow but higher quality subs.

Having said that, I think the quality continues to improve and there will be an acceptable level of base quality in some months; LLMs are helping that quite a bit. I am now testing a workflow (that I am trying to automate) which makes reasonably decent subs, but it takes about 3-4 hours to produce and costs around $2 per sub. There are so many good development works happening in the community. My guess is that in 6 months, we will have quality subs that are also fast produced.

At any rate the best subs continue to be AI + human. And hopefully not as high as 36 hours of human correcting and polishing the final output :)
I see. My problem was that on that website, I watched the movie NSFS-081 with the beautiful Mina Kitano, and for example, in one part, she told her husband something like 'It's crazy that we're doing this,' but if you use a translator, the translation says something like 'You’re going to pick it up,' referring to the part in the story where her husband wanted to bring a man to do NTR. It’s good to know there are people who dedicate several hours to offer decent translations. It doesn't matter if they don't speak Japanese or if there are small errors, but the important thing is to maintain the original story. We all read subtitles for the story, and if the story is changed in translations, then that movie would be like any other movie.
 
Imo they're good enough, and the highest quality out there from all the popular streaming sites.
But I do question sometimes some translations as well yeah, pretty sure the person does not speak Japanse either and I wonder what the source of his translations are? Is it just your default subs you find on subtitlecat (mostly done by whisperJAV I think) and he just improves upon them? SexTB does this as well and it's very noticeable. I feel like they put in 10-20 minutes of work, like fixing pronouns, and call it a day. While guru puts in a lot more work.

I did a quick test of NIMA-049, a recent one guru did. Just compare it yourself. This is a deepseek translation. Zero editing besides format. While there's some obvious mistakes and missing lines, the translation is just better. I absolutely love the "Of course, this vanilla stuff could never satisfy me." It clearly understands JAV lingo.
View attachment 3634683

Now I don't speak Japanese either, like most of us we do understand some words and intonation.

Different topic now,
I've been testing out Gemini 2.0, since it's crazy cheap, way cheaper than deepseek even and waaaay faster. But the result are lackluster to be fair. I find that the translations are the same, just worded differently and they are harder to read and not that natural. I'm wondering if others have noticed the same thing ?
As long as the translations are similar to the original, there’s no problem, but I’ve seen on that website translations that have nothing to do with the original story, as if they were just following what they see and not what is actually being said. That’s why I wanted to know if the subtitles here respect the original story. I’m not saying the translations have to be perfect, just that they should be decent and maintain the essence of the story.
 
It's not about subtitle here or somewhere else, it's about how much effort was put into it by whoever worked on it. The bigger difference is that here people will usually tell you how it was translated.

AIs are always only guessing. It's educated guesses, but there is no way to know if their output is correct or not other than a human checking it themselves, which is very time consuming.

You'll get a varying degree of quality here ranging from extremely good to pretty bad. Often the uploader will mention how they made the subtitle which is what you want to pay attention to know how good or potentially bad it will be.

Low - If it ways it was done all by some version of whisper or other AI transcription service and no editing was done, it'll have the same issues you noticed, possibly some missing lines and other potential issues associated with AIs.
Processes for this are constantly being improved on so it gets better with time, you just have to pay attention to the feedback to see how good a specifically used process is and see if that's good enough for you.

Mid - If it says some editing was done, that means someone at least took the time to look at the dialog and made it into more natural English and hopefully fixed some lines that made no sense, but didn't necessarily verify the accuracy of every lines.

High - If it says they did the translation themselves in some way, then that means most if not all lines were manually checked so a best effort(depending on their Japanese skill level) was made and this is pretty much the best you'll get short of having it professionally translated.


Mid is realistically the best you'll be able to find anywhere unless it's someone passion project or you paid for it(but even then, you might end up with AI garbage) since it can be done fairly quickly and doesn't require Japanese knowledge. You're likely to find some lines that are completely different from what is actually being said, but it should make enough sense that you may not even notice it's wrong unless you understand the Japanese.

You have to understand people here do this for fun because they love those movies and then they share it for free. If you're not happy with the quality, you can double-check the translations of the movies you're interested in, correct them and then share revised versions here, making the overall quality better for everyone.

Some people will also modify some aspect of a translation to suit their taste or even completely invent something for fun because they think it's better in some way. It'll be mentioned in their post here but somewhere else you'd likely have no idea.
 
Last edited:
  • Like
Reactions: Chuckie100

DV-1690 Tsukasa Aoi Whorehouse Rapid-fire Campaign: free if you can cum 3 times in an hour

1741225878184.png

Another experiment with 2 pass WhipserPRO Transcription + Deepseek Translation.

In this experiment I kept the disfluencies during the sanitization. Many people don't like it but I wanted to test a new sanitization script that hopefully balances the caption vs. viewing enjoyment.

I've always been a fan of Alice Japan productions with Tsukasa Aoi. I plan to continue subbing the old ones.

 

Attachments

  • Like
Reactions: ericf
1dvdes825pl.jpg

DVDES-825 [2015] Aunt-Nephew
I got a video that came with Thai subtitles in the mkv. I just extracted the file ass. and converted it to srt.
I don't have api to translate it, I tried with Mitsuko but I got an error:
[An error occurred: Request failed (400), [{"code":"invalid_type","expected":"string","received":"undefined","path":["subtitles",1,"content"],"message":"Required"}]]
I don't know how it works, sorry.
Credits, according to the file, to Fuckingday from SiamBit.me [It seems that this page has been taken down].
I would appreciate if someone would be willing to translate it by IA to English and/or Spanish. Thank you.
 

Attachments

Last edited:
1dvdes825pl.jpg

DVDES-825 [2015] Aunt-Nephew
I got a video that came with Thai subtitles in the mkv. I just extracted the file ass. and converted it to srt.
I don't have api to translate it, I tried with Mitsuko but I got an error:
[An error occurred: Request failed (400), [{"code":"invalid_type","expected":"string","received":"undefined","path":["subtitles",1,"content"],"message":"Required"}]]
I don't know how it works, sorry.
Credits, according to the file, to Fuckingday from SiamBit.me [It seems that this page has been taken down].
I would appreciate if someone would be willing to translate it by IA to English and/or Spanish. Thank you.
Thamks Michifuz. I translated your Sub to English using SubtitleEdit and have posted it.. I have a Sub from EroJapanese that I wanted to compare it to and maybe get some ideas for "enhancement" if possible. ;)
 

Attachments