hahaha this is a response from Imscully, I wasn't expecting
but it's a response that's making me laugh uncontrollably lol
Honestly, it's a good question.
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.
but it's a response that's making me laugh uncontrollably lol
Honestly, it's a good question.
Hmm... I'm not very impressed.... but it's not unfair for @creeplike to ask, I mean I don't have to accept the trade. It's not unfair to propose any kind of unfair trade...
So @creeplike what do you want for PPPD-120?
I made a machine sub for DASS-020 Yuu Shinoda:
This uses direct audio transcription --for this version I experimented with the VOSK model. It seems to do a rather acceptable job --better than PyTranscribe, Autosub, and facebook Wav2vec, specially for multi-speaker and cross talk situations. Still not the best for picking up the sex secenes --heavy breathing and mumbled. But still reasonably well, I thought. Also picks up some funny words that gets quite amusing in Englidh translation: "rice field", "mosquito", "it's going to be cheap", etc.
I've kept those untouched for the fun of it
PS. The credit goes to this forum (Akiba) and Scanlover. I hope and I'd suggest to whoever copies and versions these subs into own sites would keep due credits.
This uses direct audio transcription --for this version I experimented with the VOSK model. It seems to do a rather acceptable job --better than PyTranscribe, Autosub, and facebook Wav2vec, specially for multi-speaker and cross talk situations. Still not the best for picking up the sex secenes --heavy breathing and mumbled. But still reasonably well, I thought. Also picks up some funny words that gets quite amusing in Englidh translation: "rice field", "mosquito", "it's going to be cheap", etc.
I've kept those untouched for the fun of it PS. The credit goes to this forum (Akiba) and Scanlover. I hope and I'd suggest to whoever copies and versions these subs into own sites would keep due credits.
Attachments
Last edited:
I think that is the nickname of the son --her mother calls him that: Chunshu, we are mother and son [ 純樹, 我們是母子哦 ]
The Japanese translation can be: pure wood, it seems.
I'm not sure if that can give more clarity to the context. Or it is a pure error in transcription/caption.
Last edited:
Were you using the Big Japanese Model which can be used with Subtitle Edit? Because I've tried that one a week ago and it didn't seem much better than the small one (which was aweful)I made a machine sub for DASS-020 Yuu Shinoda:
This uses direct audio transcription --for this version I experimented with the VOSK model. It seems to do a rather acceptable job --better than PyTranscribe, Autosub, and facebook Wav2vec, specially for multi-speaker and cross talk situations. Still not the best for picking up the sex secenes --heavy breathing and mumbled. But still reasonably well, I thought. Also picks up some funny words that gets quite amusing in Englidh translation: "rice field", "mosquito", "it's going to be cheap", etc.
PS. The credit goes to this forum (Akiba) and Scanlover. I hope and I'd suggest to whoever copies and versions these subs into own sites would keep due credits.
1. Without seeing the movie, I can only assume she's talking about his cock being hard or something?
2. Maybe she is calling his name?
Exactly!!! I used the new release of the big model -I'm sure you used the new release as well. I got slightly better results with the big model -- like 25% more lines. I ran the big model on 5 different audio versions: playing with volume, bitrate, and speed. Not much difference. The slow speed was worse among all versions
I have been looking around if one can fine tune the model through Kaldi. I haven't done model training before so it is a steep learning for me. It seems that almost all the models are trained for single speaker in "quiet" acoustics. Let me know if you come across any good workarounds
Side note, have you tried Google Visoion APIs for transcription? I haven't done any research on that but I was wondering if they use similar technique as the Youtube. Presumably more promising than Google Cloud Speech
My offer is to translate it from chinese to english using subtitle edit and manually edit it myself. I will try my best to edit it and not put "meat stick".
i think your confused. I have the english subtitle of this movie already. Im asking someone to help me download other subtitles from sehuatang and i will translate it into english and manually edit it to my capability and share it here.
and also its not a perfect translation since i dont know how to speak japanese but i put my time and effort to manually edit it, (removing meat stick and other common words used by subtitle edit )
If I am confused, then explain yourself more clearly. I want to have the original source version (Chinese) for PPPD-120, you posted the English version (semi-AI) without the source version.
Note that anyone who has SHT points is almost definitely Chinese-speaker, so they (and I) don't need English version for existing Chinese SRT. I have made a few full manual Chinese-to-English translations without asking for any returns. So what you offer has zero value to me but what you ask is worth a minimal value. Minimal is not much but it's more than zero. Right? Whatever I download from SHT I'm immediately sharing on iKOA Discord. Here in the English world, the sharing culture is just different.
If you want some specific Chinese SRTs, just ask and someone may help you, but don't expect to make trades or barters here.
In case it's not been mentioned to the English internet here, there is a major crisis in the Chinese soft sub world. New release soft subs have been very difficult and scarce now. Maybe someone in the English fan base can pick up the slack, but I'm not optimistic. Unlike the anime scene, for whatever reasons, the English JAV fansub community never materialized.
If you post a single phrase (two Chinese words) or dialog line it's very difficult to get help. Chinese verbal sentences are terse (not quite as bad as Japanese) so without context (a few lines before and after) and seeing the scenes/characters, it's very easy to misunderstand/mistranslate, equally for AI and human translator.
“純樹” itself makes no sense in Chinese, but "純木" "純水“ ”純金“ make some sense. Also looking at many times the phrase appeared in the full SRT, yes its definitely the son's proper name. Possibily : https://en.wikipedia.org/wiki/Junki
Note that unlike English, trees 樹 or wood 木 is never a slang for any sex organs, look for "meat stick" 肉棒 chicken/cock 鷄鷄 etc. Also you never see cat 貓 in Chinese to mean vagina (pussy).
Another question was ”rice field" I'm pretty sure it's 飯田, check the Chinese SRT for "飯田" and listen to the Japanese audio for Iida (yes it's a four letter words with two "i"s and one "a"). "飯田" in Chinese is non-sensical, the Chinese for "rice field" or "rice paddy" is "稻田", it's unlikely... but possible, sex in a rice paddy is pretty muddy. The only sensible interpretation is Japanese surname as in one of the characters in My Hero Academia.
EDIT: Ehh... never mind, if it's direct from Japanese audio to English, then it's impossible ”rice field" be the proper name, 飯田. I have no idea.
Last edited:
Someone here downloaded the chinese file from SHT. Im not sure if he knows chinese language or what but he was kind enough to download it from SHT and shared it to me. Then i tried my best to translate it into english and shared it here as well. Im just saying what i would do because someone might think that im a leecher only (keeps on asking subtitles and not doing anything with it or sharing it). I know a lot of people here are experts and they dont want computer translation. Im just asking people if they could help me download srt files from SHT that are not readily available on the web. A lot of JULIA subtitles can be found on SHT.
Attached is the original file downloaded from SHT.
Interesting. I've compared the Vosk model with Autosub and it seemed that Autosub did a much better Job than Vosk. May have to do some more testing.Exactly!!! I used the new release of the big model -I'm sure you used the new release as well. I got slightly better results with the big model -- like 25% more lines. I ran the big model on 5 different audio versions: playing with volume, bitrate, and speed. Not much difference. The slow speed was worse among all versions
I have been looking around if one can fine tune the model through Kaldi. I haven't done model training before so it is a steep learning for me. It seems that almost all the models are trained for single speaker in "quiet" acoustics. Let me know if you come across any good workarounds
Side note, have you tried Google Visoion APIs for transcription? I haven't done any research on that but I was wondering if they use similar technique as the Youtube. Presumably more promising than Google Cloud Speech
I've never tried the Google Vision API but isn't it just for pictures and not for transcription?
Thanks for the reply. That actually provides a lot of value. "Meat stick" and "Comfortable" are so annoying to see over and over. Cheers and good luck. What one title that you are looking for? I can hook you up.
You're right, Google Vision API doesnt have that function. I mistook caption APIs of video-transcoder as a transcriber. But not that is just adding caption.
Question: have you come across this service: https://www.lalal.ai ? I am toying around if vocals isolation might help transcription results. Among the services I tried so far, these guys produced the best results. I only have tried 2 min clips. Thought of asking if you might have gone that path already.
Haha, actually, I have tried that one, but the results weren't particularly better. Some parts got improved and other got worse, so not worth the effort.
Haha -- great mind thinks alike
Same observation here, I tested several providers for vocal isolation and the results are not satisfactory. Side note: I liked the promise of "Spleeter" as it is open-sourced, but the results are underwhelming like the others. I'll let you know if I learn something.
The journey continues