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

[soloporhoy666]

solo666 can i ask what cpu do you using ?

Hi, you don't need a super computer, because of the way I'm using, I only have a good internet connection, the equipment is provided by the Collab platform, I think Drive would be equipment in the cloud.
I leave the link for you to use, is it free? Yes, but it has limits so far 5 or 6 mp3 files of approximately 100 to 200 mb. So you decide to turn to the dark side...

Google Colab

colab.research.google.com

 

[soloporhoy666]

Here I leave the subtitle obtained with Whisper+VAD, I hope you enjoy it, it's not my type of film, but since it was a test, I leave it here, later I'll upload the others that I have already subtitled with whisper, but for now I'm enjoying them for me.​

RCTD-459 BUSTY MOTHER AND SISTERS DON'T NOTICE EVEN IF THEY'RE INSERTED BY ME​

 

[maload]

Hi, you don't need a super computer, because of the way I'm using, I only have a good internet connection, the equipment is provided by the Collab platform, I think Drive would be equipment in the cloud.
I leave the link for you to use, is it free? Yes, but it has limits so far 5 or 6 mp3 files of approximately 100 to 200 mb. So you decide to turn to the dark side...

Google Colab

colab.research.google.com

than you so much

 

[Imscully]

URE-073 Exclusive Ririko Kinoshita Writhes In Incest With Immoral Mother And Child! !! Original: Gonza The Day When I Straddle My Son-Gonzo Record Of Mother And Son-​

I came across a translation for URE-073. The translation was quite good (I'm not even sure it was a "machine" translation}. I cleaned it up a little bit (mainly correction of pronouns depending on who was speaking in the scene} and married it with a reduced mosiac version of the movie to create a DVD. I'm not usually a fan of reduced mosiac movies but this one was pretty good. Anyway enjoy and let me know what you think.

Thanks for this, Chuckie100. I have several different versions of this that I've been working on myself because it's a seriously hot movie. The scene when the mom pulls out the condoms that she bought in preparation for the day she knew her son would be ready to fuck her always pops my top.

 

[DeSade]

...

 

[soloporhoy666]

Okay, so I've had success with a couple videos like this, and the results were amazing (way better than any of the alternatives, by a long shot), but I find that a lot of the time I get a "connection errored out" message. Also, the times that I got it to work, the "run WebUI" page on Drive would have the status under the links so I would know how far everything's coming along, the same way it does for the "installing dependencies section" or the other sections, but sometimes it doesn't show up, and I have no way of knowing if it's working despite waiting for hours.

BTW, not sure if it matters, but I do the "mount drive" version of WebUI that saves the .srt for you. For whatever reason, I had a little more luck with that.

Any ideas? Suggestions?

Very well I will mention the following: When I am creating the subtitle with Whisper+Vad I am constantly returning to the page and clicking it inside it so that the connection does not disconnect or think that you are a robot, in my case the forward progress, most importantly unless you have a great internet connection don't watch YouTube or download videos from the internet when using Whisper.

 

[Chuckie100]

Thanks for this, Chuckie100. I have several different versions of this that I've been working on myself because it's a seriously hot movie. The scene when the mom pulls out the condoms that she bought in preparation for the day she knew her son would be ready to fuck her always pops my top.

Yea, I liked the part when the kid used up all the condoms! Only in a AV! lol

 

[Imscully]

Yea, I liked the part when the kid used up all the condoms! Only in a AV! lol

Yeah, I loved that part. I'm sure we all remember the days of our youth when we could go through a box of condoms in a night with the right girl.
Fuck-Cum-Relax-Repeat!
Upcoming Spoilers for URE-073 If you don't want the movie spoiled, don't read any further:

My versions (I have multiple that go in a few different directions) of that particular subtitle change a few key elements that make it a lot hotter for me, including;
- They don't pretend she is his "girlfriend" because she is too old for that. In my version I change it so that she is his "MILF" friend aka a single mother from his area who he met when her son was visiting during the summer (variations on that)
- She ends up getting pregnant by letting him cum inside her. I made that change because he allows his friends to fuck her bareback and cum in her at the end. This is a pet peeve of mine as any guy who had to wear condoms at first because his mother is still fertile sure as hell isn't going to risk letting his friends knock her up by letting the cum inside her. That just doesn't make sense from a basic biological need-to-breed vantage point. So, in my versions, when the mom comes in to meet his friends, she shares the news that she just found out she is pregnant. That small change allows letting them to cum inside her make sense, as his seed as already claimed her.
- At the end of the movie, I push the timeline forward a year and have her mention that the baby is sleeping, so Mommy wants to play.
(Yes, I am a sick, sick pervert.) ;-)

 

[Aikayikes!]

I have achieved a new record for lines of text, before it was 1700 very good lines, now I have achieved more than 2600 lines, Whisper is incredible, at least for me... the movie in question was to test the reading ability of whisper, that a colleague wanted to know.
RCTD-459 Big Tits Mothers And Sisters Don_t Notice When Inserted Into Me (WHISPER+VAD)
View attachment 3119828View attachment 3119829

So impressive! I just can't seem to get it to work more than a couple hundred lines.. then some odd statement over and over. I've tried breaking the files up into smaller sound files and still the same. I run it through Subtitle Edit.. I might be missing a required program on my PC.. Hmm..

 

[soloporhoy666]

So impressive! I just can't seem to get it to work more than a couple hundred lines.. then some odd statement over and over. I've tried breaking the files up into smaller sound files and still the same. I run it through Subtitle Edit.. I might be missing a required program on my PC.. Hmm..

Hello, possibly you are not doing something, or using a very large file, my 100% recommendation is to use mp3 files, I use the audio converter from the MOVAVI suite, I apply an audio increase to all my audios with this same program. It is also important to activate the execution environment of Collab, after saving it and you should already be working with virtual GPUs, another important step is to download the repositories that come with Colla (I leave images), they are two files that must be activated every time you use the Collab, put the address of the file and hit play. Another recommendation from the developers is to have Python 3.10 installed on our computer, I don't know why, I'm not a programmer, I just followed the recommendations they told me.

 

[soloporhoy666]

So impressive! I just can't seem to get it to work more than a couple hundred lines.. then some odd statement over and over. I've tried breaking the files up into smaller sound files and still the same. I run it through Subtitle Edit.. I might be missing a required program on my PC.. Hmm..

They also recommend installing FFmpeg and GIT, as I'm not a programmer it took me a while to figure out how to install each of these files and the easiest option was through Chocolatey's code page, the only problem is that you have to install Chocolatey first on your machine through window powersell, which frustrated me a lot, because I had no knowledge of any of this, but the result was worth every hour I spent researching how to install each of these files.

 

[okconde12355]

They also recommend installing FFmpeg and GIT, as I'm not a programmer it took me a while to figure out how to install each of these files and the easiest option was through Chocolatey's code page, the only problem is that you have to install Chocolatey first on your machine through window powersell, which frustrated me a lot, because I had no knowledge of any of this, but the result was worth every hour I spent researching how to install each of these files.

I have upload video just 150MB took me 30 minutes . It so long, can you guide me how to add video from drive ( i have upload video to Google Drive and mount it but dont know how to set audio_path to the absolute path ).
Hope you can teach me. Except that part is a bit long, everything is fine, I think the subtitles are pretty good .

 

[Aikayikes!]

Hello, possibly you are not doing something, or using a very large file, my 100% recommendation is to use mp3 files, I use the audio converter from the MOVAVI suite, I apply an audio increase to all my audios with this same program. It is also important to activate the execution environment of Collab, after saving it and you should already be working with virtual GPUs, another important step is to download the repositories that come with Colla (I leave images), they are two files that must be activated every time you use the Collab, put the address of the file and hit play. Another recommendation from the developers is to have Python 3.10 installed on our computer, I don't know why, I'm not a programmer, I just followed the recommendations they told me.
View attachment 3121045
View attachment 3121046
View attachment 3121047
View attachment 3121048
View attachment 3121049

View attachment 3121050I

I appreciate the time you took to create such a clear reply. Thank you!

 

[soloporhoy666]

I have upload video just 150MB took me 30 minutes . It so long, can you guide me how to add video from drive ( i have upload video to Google Drive and mount it but dont know how to set audio_path to the absolute path ).
Hope you can teach me. Except that part is a bit long, everything is fine, I think the subtitles are pretty good .

Well, in Collab, there is a folder icon, there you go drag your file, it will take 2 or 3 minutes to upload, then it should appear in a list, you just right click and copy path and paste it and that's it, hit play and wait between 15 to 60 minutes.

 

[KingofBugs]

Wisper+VAD is really incredible. I have been messing around with it for the past week with a couple of different movies without doing any manipulation of the audio files and it appears to definitely work great in ideal circumstances where its primarily only one person talking at a decent volume. Had my best results when translating SDJS-163 as at appeared to catch most of the dialogue and most of what was written down made sense. It wasn't able to catch when someone was whispering very quietly or when the background music was playing alongside 2 ppl talking but I pretty sure if I manipulated the audio a bit in Audacity I would get better results. At the current level I and the program are at I think I can make sub files for the different news anchor series that are pretty good.

 

[Chuckie100]

NACR-600 Estrus Mother Yuri Oshikawa​

I found a fairly good machine translation for NACR-600. I cleaned it up a bit and re-interpreted some of the meaningless 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. Enjoy and let me know what you think.

 

[ericf]

Whisper Colab seems to limit usage variably. Today, I could only process two 1 hr files before getting told I didn't have more GPU resources available.
Trying to get as complete a result from the software for one specific video.

 

[soloporhoy666]

Whisper Colab seems to limit usage variably. Today, I could only process two 1 hr files before getting told I didn't have more GPU resources available.
Trying to get as complete a result from the software for one specific video.

You can compensate this situation with another account, remember that the use of Whisper is free, so it has some limitations, it also has a payment option, this is only to be able to use the gpus for a month. You can alternate one account and another, with that I can do between 8 and 12 subtitles in a day, no more is needed.

 

[Makkdom]

Wisper+VAD is really incredible. I have been messing around with it for the past week with a couple of different movies without doing any manipulation of the audio files and it appears to definitely work great in ideal circumstances where its primarily only one person talking at a decent volume. Had my best results when translating SDJS-163 as at appeared to catch most of the dialogue and most of what was written down made sense. It wasn't able to catch when someone was whispering very quietly or when the background music was playing alongside 2 ppl talking but I pretty sure if I manipulated the audio a bit in Audacity I would get better results. At the current level I and the program are at I think I can make sub files for the different news anchor series that are pretty good.

So Wisper can't catch whispers? Anyway, keep up the good work. I am looking forward to the results.

 

[soloporhoy666]

So Wisper can't catch whispers? Anyway, keep up the good work. I am looking forward to the results.

I think there are many variables with whisper, in my case I got good results from bad movies, in terms of audio quality and noise or music, most of the subtitles that I have been able to do have 90% excellent results, as already I explained before I modify my audios with movavi, I increase the audio and balance it to help the VAD system and consequently Whisper, I have obtained 3 subtitles with more than 1000 lines in old and poor quality movies, which have the particularity of being recorded in the interperie, with several actors talking and the result is INCREDIBLE (DVUMA) this series is recorded in noisy areas (sex with homeless people)
obviously it is not perfect, but it is much better than other options that have been around for a long time, remember that this AI is open source so it can be improved by people or developers who can, regardless of this, the AI is still in training , so we can expect it to improve even more.