The Akiba-online English Sub Project★NOT A SUB REQUEST THREAD★

  • Throughout the month of April 2024, participate in the FileJoker Thread Contest OPEN TO EVERYONE!

    From 1st to 30th of April 2024, members can earn cash rewards by posting Filejoker-Exclusive threads in the Direct-Downloads subforums.

    There are $1000 in prizes, and the top prize is $450!

    For the full rules and how to enter, check out the thread
  • 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.

okconde12355

New Member
Nov 5, 2021
5
0
May I ask something, what software do they use to create file srt so fast . It seems that any movie that has just been released has been available for a few days
 

maload

Active Member
Jul 1, 2008
615
118
May I ask something, what software do they use to create file srt so fast . It seems that any movie that has just been released has been available for

May I ask something, what software do they use to create file srt so fast . It seems that any movie that has just been released has been available for a few days
can i ask you where is them ?
who is THEM ?

i want to check this subtitle too. i just want to download some subtitle .
 

avatarthe

Well-Known Member
Feb 1, 2008
184
277
May I ask something, what software do they use to create file srt so fast . It seems that any movie that has just been released has been available for a few days
I don’t know how the original Chinese had subs are made so fast, I think they must be getting prerelease copies of the films. The SRT files are made by people scrapping these hard sub copies of the films with Video Sub Finder And then passing the subs through an OCR program like JaidedAI/EasyOCR. Using these Two programs you can extract a srt file from a hard sub movie in like 20 mins. You can then take the Chinese srt file and translate it into English with Deeplv4.py in 2 minutes.
 

ding73ding

Akiba Citizen
Oct 25, 2009
2,332
2,070
I don’t know how the original Chinese had subs are made so fast, I think they must be getting prerelease copies of the films. The SRT files are made by people scrapping these hard sub copies of the films with Video Sub Finder And then passing the subs through an OCR program like JaidedAI/EasyOCR. Using these Two programs you can extract a srt file from a hard sub movie in like 20 mins. You can then take the Chinese srt file and translate it into English with Deeplv4.py in 2 minutes.
Well first you may ask how the pirates get the high quality video so quickly, often beating the scheduled release date. One (older) theory is that the DVD (now BluRay) are manufactured (like everything else) in China the factory are a source. Then it seems like (no evidence at all) there's a production line of translator and SRT creators.

It's absolutely certain that soft subs were created first. It's just how it works, hard subs are made from soft subs. But the original subbers choose not to release the soft subs, at least not immediately and not for free. So then a few people in the community went thru the silly steps of VideoSubFinder etc to create soft subs from hard subs. I did that myself a few times (very few, because it really is silly).

VideoSubFinder itself is far from fully accurate/automatic, and OCR is generally hit and miss. Not that I have tried those OCR programs you mentioned. But imperfect sub-extract (creates a stack of still images), then imperfect OCR, then imperfect Chinese to English translation. Pffft... I have plenty reasons to think we are still many years away from a reliable English sub work flow that doesn't require a Chinese-fluent human many hours of work.
 

Taako

Akiba Citizen
May 25, 2017
1,266
856
Well first you may ask how the pirates get the high quality video so quickly, often beating the scheduled release date. One (older) theory is that the DVD (now BluRay) are manufactured (like everything else) in China the factory are a source. Then it seems like (no evidence at all) there's a production line of translator and SRT creators.

It's absolutely certain that soft subs were created first. It's just how it works, hard subs are made from soft subs. But the original subbers choose not to release the soft subs, at least not immediately and not for free. So then a few people in the community went thru the silly steps of VideoSubFinder etc to create soft subs from hard subs. I did that myself a few times (very few, because it really is silly).

VideoSubFinder itself is far from fully accurate/automatic, and OCR is generally hit and miss. Not that I have tried those OCR programs you mentioned. But imperfect sub-extract (creates a stack of still images), then imperfect OCR, then imperfect Chinese to English translation. Pffft... I have plenty reasons to think we are still many years away from a reliable English sub work flow that doesn't require a Chinese-fluent human many hours of work.
What your wrote makes sense. Extracting is ok i supposed but seems complicated... to me :rolleyes:
I never tried extracting, I wonder if it's worth considering i will go and recheck anyways. I was looking a t SUN-037 recently and somebody made the worse English sub to date. They imprinted the subs over the Chinese subs. Its a mess. I'm thinking just manually writing the sub and not bother.
 

avatarthe

Well-Known Member
Feb 1, 2008
184
277
What your wrote makes sense. Extracting is ok i supposed but seems complicated... to me :rolleyes:
I never tried extracting, I wonder if it's worth considering i will go and recheck anyways. I was looking a t SUN-037 recently and somebody made the worse English sub to date. They imprinted the subs over the Chinese subs. Its a mess. I'm thinking just manually writing the sub and not bother.
Extracting is now not very difficult at all. Taking a hard sub film in Chinese and creating pretty good MTL English soft subs can be done in 20- 30 minutes also the ocr technology got a big jump about a year ago when a Chinese hacker created a program that uses bandu’s ocr, it’s neat 100% accurate but requires and bandu account, I have used EasyOCR and it’s almost as good as the Chinese hackers program (99% accurate), it’s just a little slower… maybe 20 mins to OCR a film @ding73ding it’s so much better, like night and day compared with the software I was using when you and I subs a few films together a couple of years ago. Finally @ironfevers deeplv5 software which translate the srt file using Deepl is amazing! There still are some pronoun issues my/ your, he/she issues from time to time but most of the sentences cooker out I’m pretty amazing English, I encourage people to try it for your selves. There is a tutorial on the tutorial section on this BBS that does you Joe to set up and use easy OCR ( installing it is a bit tricky but do try) and another tutorial, if you search my posts, that tells you hire to Install deeplv4


EasyOCR tutorial


Deeplv4 tutorial
https://www.akiba-online.com/thread...-request-thread.1466451/page-113#post-4234482

Also when I say the hard subs are made first I mean the Chinese pirate are creating subtitles and only releasing hard subbed copies of the films so you have to watch the copies with the ad on them, if they release the soft subs no on would watch the u ad ladled had sub copies of the films and they would loose money.

I will upload the srt file for latest film I extracted and MLT for you guys to check out, it’ss good film, but just check the quality of the unedited MLT subs, again yes there are some pronoun issues but look beyond that…

The Description is also MLT from the Japanese site..

poster.jpg
DASD-934 - Lubricated Slender Step-Sisters Rub Big Tits With Lotion Lovingly To Tempt Father-in-law Yu Shinoda Yuria Hakaz (2021)

Starring Yu Shinoda & Yuria Hafu

Yu and Yuria live with their new father after their mother remarries. One day, their father-in-law calls them for a threesome course without knowing it. One day, Yuu is fired from his job due to the recession. Yuu, who doesn't earn enough money, takes advantage of her father-in-law's love of sex and uses her erotic skills to beg him for sex. And Yuria, who witnessed them having sex, is also ...... trying to compete with her sister.
 

Attachments

  • poster.jpg
    poster.jpg
    46.8 KB · Views: 90
  • DASD-934.dl.en.rar
    27.1 KB · Views: 281
Last edited:

ding73ding

Akiba Citizen
Oct 25, 2009
2,332
2,070
Extracting is now not very difficult at all. Taking a hard sub film in Chinese and creating pretty good MTL English soft subs can be done in 20- 30 minutes also the ocr technology got a big jump about a year ago when a Chinese hacker created a program that uses bandu’s ocr, it’s neat 100% accurate but requires and bandu account, I have used EasyOCR and it’s almost as good as the Chinese hackers program (99% accurate), it’s just a little slower… maybe 20 mins to OCR a film @ding73ding it’s so much better, like night and day compared with the software I was using when you and I subs a few films together a couple of years ago. Finally @ironfevers deeplv5 software which translate the srt file using Deepl is amazing! There still are some pronoun issues my/ your, he/she issues from time to time but most of the sentences cooker out I’m pretty amazing English, I encourage people to try it for your selves. There is a tutorial on the tutorial section on this BBS that does you Joe to set up and use easy OCR ( installing it is a bit tricky but do try) and another tutorial, if you search my posts, that tells you hire to Install deeplv4


EasyOCR tutorial


Deeplv4 tutorial
https://www.akiba-online.com/thread...-request-thread.1466451/page-113#post-4234482

Also when I say the hard subs are made first I mean the Chinese pirate are creating subtitles and only releasing hard subbed copies of the films so you have to watch the copies with the ad on them, if they release the soft subs no on would watch the u ad ladled had sub copies of the films and they would loose money.

I will upload the srt file for latest film I extracted and MLT for you guys to check out, it’ss good film, but just check the quality of the unedited MLT subs, again yes there are some pronoun issues but look beyond that…

The Description is also MLT from the Japanese site..

View attachment 2765399
DASD-934 - Lubricated Slender Step-Sisters Rub Big Tits With Lotion Lovingly To Tempt Father-in-law Yu Shinoda Yuria Hakaz (2021)

Starring Yu Shinoda & Yuria Hafu

Yu and Yuria live with their new father after their mother remarries. One day, their father-in-law calls them for a threesome course without knowing it. One day, Yuu is fired from his job due to the recession. Yuu, who doesn't earn enough money, takes advantage of her father-in-law's love of sex and uses her erotic skills to beg him for sex. And Yuria, who witnessed them having sex, is also ...... trying to compete with her sister.
Ok ok I remember ... yeah you kinda know your thing. I have no reason to dispute you, but ... I think if you wait some time, runbkk eventually will get the soft subs out there. Every week or so I just search his new releases and download the .SRT folder (most of the movies I don't care to download). It's frustrating the soft subs come four and a half months after original release. But otherwise I'd rather wait that long (doing whatever I want with my time) rather than deal with multiple imperfect (no matter how good they are getting). So at least that eliminate the issues with OCR.

Leaving the final step, I think technically the hardest step, ZH-EN translation. I am skeptical but ok... maybe deepl is really that amazing... (main reason I'm skeptical is that even 100% accurate soft sub contains a lot of poorly formed Chinese sentences and imprecise slangs (that's how it is with common folk spoken Chinese) so even I am fully fluent in Chinese I have to take some guesses, and often have to check the video footage to confirm the meaning of many words (esp. pronouns and some slangs) (I mean, I often translate Chinese soft sub directly in the text file, not checking the video footage unless necessary)

One thing though... having said all that, there are a few vids I only managed to find hard sub and still not the soft sub (runbkk doesn't cover a lot of 2nd tier and genre studios, unfortunately). One being the excellent ghost story DASD-788.
 
Last edited:

Taako

Akiba Citizen
May 25, 2017
1,266
856
Extracting is now not very difficult at all. Taking a hard sub film in Chinese and creating pretty good MTL English soft subs can be done in 20- 30 minutes also the ocr technology got a big jump about a year ago when a Chinese hacker created a program that uses bandu’s ocr, it’s neat 100% accurate but requires and bandu account, I have used EasyOCR and it’s almost as good as the Chinese hackers program (99% accurate), it’s just a little slower… maybe 20 mins to OCR a film @ding73ding it’s so much better, like night and day compared with the software I was using when you and I subs a few films together a couple of years ago. Finally @ironfevers deeplv5 software which translate the srt file using Deepl is amazing! There still are some pronoun issues my/ your, he/she issues from time to time but most of the sentences cooker out I’m pretty amazing English, I encourage people to try it for your selves. There is a tutorial on the tutorial section on this BBS that does you Joe to set up and use easy OCR ( installing it is a bit tricky but do try) and another tutorial, if you search my posts, that tells you hire to Install deeplv4


EasyOCR tutorial


Deeplv4 tutorial
https://www.akiba-online.com/thread...-request-thread.1466451/page-113#post-4234482

Also when I say the hard subs are made first I mean the Chinese pirate are creating subtitles and only releasing hard subbed copies of the films so you have to watch the copies with the ad on them, if they release the soft subs no on would watch the u ad ladled had sub copies of the films and they would loose money.

I will upload the srt file for latest film I extracted and MLT for you guys to check out, it’ss good film, but just check the quality of the unedited MLT subs, again yes there are some pronoun issues but look beyond that…

The Description is also MLT from the Japanese site..

View attachment 2765399
DASD-934 - Lubricated Slender Step-Sisters Rub Big Tits With Lotion Lovingly To Tempt Father-in-law Yu Shinoda Yuria Hakaz (2021)

Starring Yu Shinoda & Yuria Hafu

Yu and Yuria live with their new father after their mother remarries. One day, their father-in-law calls them for a threesome course without knowing it. One day, Yuu is fired from his job due to the recession. Yuu, who doesn't earn enough money, takes advantage of her father-in-law's love of sex and uses her erotic skills to beg him for sex. And Yuria, who witnessed them having sex, is also ...... trying to compete with her sister.
I thank you for your insight. I will check it out for sure.:)
Do you use notepad++? Is there a way to edit .srt files with it... if you're using pytranscriber?
Anyways, thank you for your tip.
 

avatarthe

Well-Known Member
Feb 1, 2008
184
277
Ok ok I remember ... yeah you kinda know your thing. I have no reason to dispute you, but ... I think if you wait some time, runbkk eventually will get the soft subs out there. Every week or so I just search his new releases and download the .SRT folder (most of the movies I don't care to download). It's frustrating the soft subs come four and a half months after original release. But otherwise I'd rather wait that long (doing whatever I want with my time) rather than deal with multiple imperfect (no matter how good they are getting). So at least that eliminate the issues with OCR.

Leaving the final step, I think technically the hardest step, ZH-EN translation. I am skeptical but ok... maybe deepl is really that amazing... (main reason I'm skeptical is that even 100% accurate soft sub contains a lot of poorly formed Chinese sentences and imprecise slangs (that's how it is with common folk spoken Chinese) so even I am fully fluent in Chinese I have to take some guesses, and often have to check the video footage to confirm the meaning of many words (esp. pronouns and some slangs) (I mean, I often translate Chinese soft sub directly in the text file, not checking the video footage unless necessary)

One thing though... having said all that, there are a few vids I only managed to find hard sub and still not the soft sub (runbkk doesn't cover a lot of 2nd tier and genre studios, unfortunately). One being the excellent ghost story DASD-788.
@ding73ding I decided to rip your ghost story film as a test
Video Sub Finder took 12 minutes to extract the subs as jpg files
I looked over the subs and removed the blanks in 9 minutes (this is the only manual part of the project and this film has a lot (1418) of subtitles!
The Chinese hacker OCR program (RGB_Combine) took 12 minutes o convert the JPG subs to real text and crate a Chinese SRT file
Finally Deeplv4 took 2 minutes to translate the Chinese srt file into English.
Total time to extract and translate subtitles 37 minutes,amount of my time required about 10 minutes


I've included both the Chinese and English srt files for you to compare.


fanart.jpg
My Beloved Girlfriend Cant See Whats Happening To Me. This Ghost Has Been Attracted To My Cock, And Here We Are, The 3 Of Us, Living Together In This Curious Arrangement. Mahiro Ichiki Sumire Mizukawa [DASD-788]

For the first time, they live together, but they are introduced to an accidental property. For some reason, my girlfriend Sumire liked it and decided to live there. From that day on, I was frightened by the psychic phenomenon around me. One day, a ghost reveals itself to me. ...Unlike what I expected, it was a tiny and cute ghost named Mahoro. She seems to be invisible to Sumire. She loves to do naughty things, and every time she appears, she plays erotic pranks on him... [She can't see me, it's my own secret threesome].
 

Attachments

  • DASD-788.rar
    54 KB · Views: 254

Taako

Akiba Citizen
May 25, 2017
1,266
856
@ding73ding I decided to rip your ghost story film as a test
Video Sub Finder took 12 minutes to extract the subs as jpg files
I looked over the subs and removed the blanks in 9 minutes (this is the only manual part of the project and this film has a lot (1418) of subtitles!
The Chinese hacker OCR program (RGB_Combine) took 12 minutes o convert the JPG subs to real text and crate a Chinese SRT file
Finally Deeplv4 took 2 minutes to translate the Chinese srt file into English.
Total time to extract and translate subtitles 37 minutes,amount of my time required about 10 minutes


I've included both the Chinese and English srt files for you to compare.


View attachment 2766063
My Beloved Girlfriend Cant See Whats Happening To Me. This Ghost Has Been Attracted To My Cock, And Here We Are, The 3 Of Us, Living Together In This Curious Arrangement. Mahiro Ichiki Sumire Mizukawa [DASD-788]

For the first time, they live together, but they are introduced to an accidental property. For some reason, my girlfriend Sumire liked it and decided to live there. From that day on, I was frightened by the psychic phenomenon around me. One day, a ghost reveals itself to me. ...Unlike what I expected, it was a tiny and cute ghost named Mahoro. She seems to be invisible to Sumire. She loves to do naughty things, and every time she appears, she plays erotic pranks on him... [She can't see me, it's my own secret threesome].
Interesting, I try using the video sub finder and the subs looked blurry. My computer might not be able to handle it :p
 

avatarthe

Well-Known Member
Feb 1, 2008
184
277
Interesting, I try using the video sub finder and the subs looked blurry. My computer might not be able to handle it :p
Just use VSF to create the RBG Image folder of clipped subtitle jpgs, don’t use it for OCR. Use EasyOCR on the RBG Image folder to create a folder of txt files containing the Chinese subs (txt Results), finally use VSF to “create subtitles from txt results” to create an srt file from the txt result folder.
 

Taako

Akiba Citizen
May 25, 2017
1,266
856
Just use VSF to create the RBG Image folder of clipped subtitle jpgs, don’t use it for OCR. Use EasyOCR on the RBG Image folder to create a folder of txt files containing the Chinese subs (txt Results), finally use VSF to “create subtitles from txt results” to create an srt file from the txt result folder.
Thanks for your advice, I will try it :)
 

Taako

Akiba Citizen
May 25, 2017
1,266
856
So my next title is this, VENU-311.. when I get too it. ;)

A Family with No Sense of Fidelity, starring Ryoko Iori. i personally find Ryoko Iori to be a beautiful MILF. She didn't do many films but bless her she's great in this series. And so is all the old school actors.​

ven311 1.jpgven311 2.jpg