Support for proper release naming

[js37173]

It's not very user friendly to delete subs because of duplicate content, while the release naming is ignored.

How is a user holding
This.One.Movie.1999.1080p.BluRay.x264-THATGROUP
supposed to know that
Thismovie English.srt
subs for this release are wrongly timed (1000ms+ wrong), while
Thismovie [Sub] [Eng].srt
are proper but full of OCR typos, but if you did connect telepathically to OpenSubtitles.org admin's brain then you'll actually know the proper ones to download are
This.One.Movie.1999.720p.BRRip.DTS.H264-OTHERGROUP
or some other completely unguessable title.

It's not possible to know. So when a user uploads the best subs for This.One.Movie.1999.1080p.BluRay.x264-THATGROUP and they get deleted because "they are duplicates" you are forgetting one important aspect:

The user doesn't know this. Sure it's easy for you as admins to run a diff/whatever script to help detect it, but it does not help a user to find the proper subs at all. If you're unsure about this just make a poll, do you prefer to find your release name immediately, or spend 30 minutes hunting for proper subtitles.

A working solution would be as easy as just sym/hardlinking the release name to the now-detected dupe, costing zero disk space, but helping immensely everyone in possession of this exact release.

Comments etc are not a solution because they won't show in Kodi.

I'm not going to whine about "muh deleted uploads" because it's not about me but everyone finding their proper subs in the shortest possible time the easiest possible way.

[SmallBrother]

Imagine this, reality being simplified:
- We have three 'families' of video files: Bluray, WEB-DL and HDTV, with each family existing of 20 video files with SAME sync, but DIFFERENT names. The three families (BR/WEBDL/HDTV) differ from each other in sync.

By deleting subs we are not forgetting the aspect you mention, beacuse:
- It is enough to host three subtitles.
- The downloader has to try maximum three subs.

By NOT deleting duplicates, the reality would be this, for example:
- We have 10 subs for bluray video files, 10 for WEB-DL, and 10 for HDTV versions.
- These subs are still not covering ALL available video files, so in this case the downloader has a 50% chance that his desired version is not listed. If release names are not original, this percentage is even higher.
- Now the downloader has to try not 3, but possibly up to 21 subs, to find a match.

Or even worse:
- We have only subs for Bluray and WEB-DL videos, 10 each.
- The downloader wants subs for HDTV.
- The downloader has to try 20 subs and finds out NONE of the subs are matching.

Maybe, if literally ALL release names have their own subtitle, it might be easier to find a match. But in reality that situation will never be reached. So I think it's not a good idea to allow duplicates for the reason you mention.

Ideally, releasenames are linked to some subtitle by hash, so they are visible for API-driven software. This happens when uploading using software, like the OpenSubtitles Uploader of Vankasteelj. Over time, additional release names are automatically added to the subtitles, by users using API-driven software/players or actively/deliberately added.

That being said, it would be a good idea if a hash for some release could be 'transferred' to an existing subtitle, before deleting the duplicate.

[fun7om]

and im just dropping the video file in the serch box,
and get in 1 second the match subtitleS in high accuracy.

i guess some of the softwares(API) uses the same trick.

[js37173]

Imagine this, reality being simplified:
- We have three 'families' of video files: Bluray, WEB-DL and HDTV, with each family existing of 20 video files with SAME sync, but DIFFERENT names. The three families (BR/WEBDL/HDTV) differ from each other in sync.

If only it was this simple, but we have groups that decided to cut something from the start of the movie or a movie released in another region had differing length pre-credits/copyright warning/whatever, and the timing is off. And where did any analog ripper start recording.

And you juggle with the timing adjustment, only to later find out that it keeps getting out of sync, because your movie and the subs are not matching 23.976/24/25 fps... (+ 29.97 for ntsc dvd).

So in reality it's a mess of rip lengths and framerates and sometimes takes a lot of time to find the proper sub. Frame rate issues are the worst because they're not immediately obvious, but do ruin the fun of just watching.

The only way this can be avoided is to serve subs for each specific release. Which only takes minimum resources because they can be linked as duplicates of the same base sub.
This will decrease your bandwidth utilization because instead of trying 3-10 random subs they get the right one immediately.

I'm really happy for you guys who watch on pc and have some fancy software that does magically find the right subs, but please don't forget about the millions just using Kodi etc with a dumb sub searcher, they could definitely use some improvement.

[SmallBrother]

If only it was this simple, but we have groups that decided to cut something from the start of the movie or a movie released in another region had differing length pre-credits/copyright warning/whatever, and the timing is off. And where did any analog ripper start recording.

This actually supports my argumentation. The idea is to have only subs which are significantly different. If the 'bluray family' is divided into two different versions due to other cuts or intros, there would be four subs to try. That's more than three, but still less than FOUR times 10 or 20.

The only way this can be avoided is to serve subs for each specific release. Which only takes minimum resources because they can be linked as duplicates of the same base sub.

Theoretically, this might be right, but only if literally ALL versions are covered. But realistically, this is not possible. Who is going to find/check/link all those versions? And when? If only ONE version is missing, and I want subs for that specific version, I would have to dig into hundreds of subs, not just a few.

[oss]

thats why we want all subtitles are uploaded with video files, so we can use video hash matching.