SupRip Support Thread

[nautilus7]

Join the club!

And if you look more carefully, you 'll see missing letters/characters in the end/beginning of italic lines.

But none seems to care.

[menlvd]

last line in the sup look like this, ending by 0 :(
1043
02:00:43,277 --> 00:00:00,000
Line 1043

[quantum]

For Taktaal regarding sub/idx:

Good overview of the vobsub sub/idx format:
http://trac.annodex.net/wiki/VOBSub

Doom9 guide for ripping from DVD to sub/idx:
http://www.doom9.org/index.html?/dvobsub.htm

Misc:
http://forum.doom9.org/showthread.php?t=99815

Sourceforge links for vobsub:
http://sourceforge.net/project/showfiles.php?group_id=205650&package_id=246121
http://sourceforge.net/project/showfiles.php?group_id=82303

Manusse's (SubtitleCreator author) explanation of Standard Def subtitle format:
http://forum.doom9.org/showpost.php?p=1123472&postcount=195

A comment by Manusse on another forum:
"VobSub format (sub/idx) is a SD format. The only way to convert a HD-DVD subtitle file to VobSub is using SubtitleCreator. It will downsample it to SD resolution and decrease its number of colors to 4 and allow to export it as sub/idx. However it can't currently open Blu-Ray Sup files."

My understanding of sub/idx:
The standard definition (SD) sub/idx is a graphical subtitle format viewable on PC's using directvobsub. Zoom player, Media Player Classic, VLC, and maybe others can display subtitles in this format, even when playing HD files such as those demuxed from Bluray.
Pros: Subs are graphical, can display any character, and are not subject to OCR errors. Subs can be displayed anywhere on screen.
Cons: Subs are SD which looks somewhat blocky on a HD screen.
Challenge: The HD subs must be downconverted from HD sup to SD sub/idx format.

Some hardware players support the SD sub/idx format when playing HD files.
- DVICO TVIX HD M-6500A - supported
- Popcorn Hour A-110 - supposedly will be supported in a future firmware

SubtitleCreator can convert HD-DVD SUP files to SD sub/idx. Maybe Taktaal is willing to do the same thing with Suprip, and include Bluray support. :)

[nautilus7]

last line in the sup look like this, ending by 0 :(
1043
02:00:43,277 --> 00:00:00,000
Line 1043
Also noticed this some time ago. Another Suprip bug...

[Taktaal]

The problem with recognizing the first character in italic lines and the one with the 0 end time on the last subtitle will be fixed in the next version. I'm currently working on the sub/idx stuff and will release it when that's done.

[rica]

Thx in advance.

[quantum]

... I'm currently working on the sub/idx stuff and will release it when that's done.

I just noticed this on the Popcorn Hour beta firmware thread (http://www.networkedmediatank.com/showthread.php?tid=8159&pid=70110#pid70110) update for Dec 26:
IDX/SUB supported (first implementation, only external IDX/SUB)

Once this is released, there will be two (that I'm aware) hardware players supporting sub/idx. I have the Popcorn Hour, so I'll be glad to test ;)

[turbojet]

quantum, while it takes a few extra steps you can get vobsub from bluray subs by using supread to convert bluray to hd-dvd and then into subtitlecreator. I agree it would be nice if suprip could support it with one click.

[quantum]

quantum, while it takes a few extra steps you can get vobsub from bluray subs by using supread to convert bluray to hd-dvd and then into subtitlecreator. I agree it would be nice if suprip could support it with one click.

I have tried this path a few times. I got it to "sort of" work once. The other two times it failed either in supread or in subtitlecreator. The one time it did work, there were glitches in the subs, such that one of them would stay on until I switched subs off in the player.

However, seeing this technique in action is what got me interested in sub/idx in the first place ;) But I found the supread/subtitlecreator path to be awkward, slow, and sometimes would not work at all. But I agree that it is a path that is workable, and maybe others could have better luck with it than I had.

With that said, Suprip already supports HD DVD and Blu-ray sups, and Taktaal has been great in tweaking it to support the different variations that have been popping up. If he can get reliable sub/idx output, then it would be close to a one-click answer, maybe even command line we can put in a batch file ;)

[saint-francis]

The one time it did work, there were glitches in the subs, such that one of them would stay on until I switched subs off in the player.

I too have noticed this, but only once. The supread/subtitlecreator method has worked for me with all but 3 disks. I think the issue was with supread since Taktaal needed to spruce up suprip to make the original sup work with it.

[sshd]

There are some issues with 1.12:

- It is often impossible to get "1" correct. This number has a lot of space either side, which suprip translates into " 1 ". Not possible to adjust it under options, as it breaks everything else.
- "l" (lower capital L) is detected as "I".
- Replace high , with ' -- default setting is often wrong
- Crashes (where can I send more info)


It is possible and easy to fix the first three in postprocessing:

- Start by counting number "1" and " 1 ". If they are the same, then all 1s were detected as " 1 ".
- Replace all "I"s inside words with "l"s unless the word is a Roman number (Roman numbers contain of the letters I, V, L, C, D and M).
- Look up all words starting with "I" in a dictionary and convert to "l" if found. I can and will help you make such a dictionary, at least for English.
- Count the total number of , and '. If there are no commas at all or very few, then the "Replace high , with '" is most likely wrong.

[nautilus7]

All these are common problems... Plus the missing letters in the start/end of many subs.
To put it simple the program is useless in its current version, but the developer is busy with adding idx/sub support... No comment.

[rack04]

What is idx/sub and how does that relate to sup files demuxed from HD DVD and Blu-ray using eac3to?

[nautilus7]

It's another subtitle format. I don't know exactly, but you can easily convert sup to idx/sub.
It has nothing to do with eac3to.

[Taktaal]

idx/sub is basically the old DVD subtitle format. It's limited to low resolution, 4 colors and only one degree of transparency. It's also very badly documented. I checked a few open source utilities but even they don't agree what to set certain header bytes to. So Sub/idx format has been kinda postponed, maybe canceled. The documentation on the format is very sparse. It's possible to write a file but they generally just crash players as they get some kind of values they don't expect. As I've said before I really hate writing files in a format I don't have the official documentation from because it's just about impossible to find out whether a certain bug is your fault or not...

Anyway, I uploaded a new version that fixed a few earlier bugs.

I'll try to get ahold of some MPC-HC or VLC developers and talk with them about direct .sup display support. That's probably the better way to move bitmap subtitles forward.

[saint-francis]

I'll try to get ahold of some MPC-HC or VLC developers and talk with them about direct .sup display support. That's probably the better way to move bitmap subtitles forward.

MPC HC supports .sup in m2ts. The problem is that mkvmerge doesn't support .sup.

[rica]

MPC HC supports .sup in m2ts. The problem is that mkvmerge doesn't support .sup.

It doesn't matter; TSMuxer makes m2ts with sups.

[nautilus7]

Anyway, I uploaded a new version that fixed a few earlier bugs.Thank you!

I was able to successfully OCR the Greek sup that i 've uploaded earlier. No problems at all. :)

But this (http://www.sendspace.com/file/xkihy6) sup file is (also gave trouble before) still missing the last letter from some lines. Though now it's only missing the last. And only in case it's a "." (dot).

[nautilus7]

OCR'ed a few more subs...

The problem with last character missing (only . ) happens with italic lines only.
Also noticed one case where a comma was missing from the end of the 1st line (italic also).

Both bugs can be reproduced with this (http://www.sendspace.com/file/xu6rtq) sup file. Subtitles #583, 1180, 1331, etc.

Also subtitle #792 is messed in the above sup file. SupRip thinks it's italic where it's not.

[BZeeme]

Just tried v1.13. It will not run on Vista 64 ->
[critical error:]

Index was outside the bounds of the array.

Double BorderWidth(Byte[,], Int32, Int32)

at SupRip.SubtitleLetter.BorderWidth(Byte[,] array, Int32 side, Int32 position)
at SupRip.SubtitleLetter.AveragedBorderWidth(Byte[,] array, Int32 side)
at SupRip.SubtitleLetter.ReduceImage()
at SupRip.SubtitleLetter..ctor(Byte[,] i, String s)
at SupRip.SubtitleFont..ctor(FontType t, String fn)
at SupRip.SubtitleFonts..ctor()
at SupRip.MainForm..ctor()
at SupRip.Program.Main(String[] args)

Version 1.13 does run on Vista 32. I have been using v1.01 on Vista 64.

[FreaQ]

I tried to convert a .sup which was demuxed by EVODemux.
But i ran into this error, when I opened the file.

timeOfControl is not 0 on a start

Void LoadHddvdSup(System.IO.FileStream)

bei SupRip.SubtitleFile.LoadHddvdSup(FileStream fs)
bei SupRip.SubtitleFile.LoadFromSup(String supfile)
bei SupRip.MainForm.LoadSubtitleFile(String fileName)

So I tried to demux the .sup again with eac3to, but the error remains.

Programs: all latest versions

Any suggestions?

[legoholic]

:thanks: with the 1.13 version I was finally able to convert the Micheal Clayton sup (the 1.5Gb every frame duplicated one)

I still have a problem though with this one (members1.chello.nl/~e.abbink/00011.track_4609.zip) though from Asterix and the Vikings.

It has a cartoony font which totally screws up character recognition (especially separation) but even if I persist and keep filling in by hand the two and three letter combos it just crashes after a while. (Not with any crash log, but with the windows "suprip has encountered a problem and needs to close")

Also often it does not recognize the last character of the line.

I know its only one weird font, but maybe levelling the images before ocr-ing would help? remove the outline and leave only the white inner letter? that would problably help a lot with the character splitting.

[nautilus7]

Also often it does not recognize the last character of the line.
That's i believe the most significant remaining bug. It happens only with italic and if the last character of a line (either the 1st of the 2nd) is a dot(.) or a comma (,).

[legoholic]

no that's not what I mean. I don't have that problem (at least not anywhere near to the same degree) with other sups.

I think it is specifically caused by this font/sup. Maybe it has something to do with many of the images having the rightmost character right up to the right border of the (displayed) image. ie there is no black at all to the right of it.

[rica]

I tried to convert a .sup which was demuxed by EVODemux.
But i ran into this error, when I opened the file.



So I tried to demux the .sup again with eac3to, but the error remains.

Programs: all latest versions

Any suggestions?

The sups you get are HD-DVD compatible sups. You have to convert them to BD compatible sups first using SupRead.

[quantum]

The sups you get are HD-DVD compatible sups. You have to convert them to BD compatible sups first using SupRead.
:confused:
Suprip reads HD-DVD sup files.

[gkar]

I have completed my first rip using your programme: thanks for your efforts.

I demuxed the Ironman movie m2ts file using TSRemux to get the *.sup files.

I missed the 'only forced subtitles' on the SRT tab first time around & gave up on the manual OCR because I couldn't figure out how to get italics working.

After I came back to it, there were 4 four *.sup files to check, the first two had no forced subs but the third one did. The last one strangely enough, was in another language, even though I only had checked the English presentation files in TSRemux.

I did a manual OCR. The following issues presented themselves: with in the occasional word 'p' was replaced with 'c' even though it appeared correct in the area where the recognised words are surrounded by green box (not sure what your term is for this area) (red boxes). Also, a number of words were joined with no ability to separate them as there is no green box between the words (orange boxes). I had to manually adjust these inside the saved *.srt file.

http://img384.imageshack.us/img384/6821/ocr1fu9.jpg

Also, how do you use italics, bold text, etc. inside your programme? Thanks.

Finally, why is it not feasible to have a programme that just uses the named *.sup file directly inside your favourite player -eg. Zoomplayer? Why the need for this extra step of OCRing it to another format?

Please, no technical dissertations; just a plain-speak reason if it is at all possible. Cheers.

PS. any idea why Zoomplayer adjusted the aspect ratio when I am using the Ironman.srt subs file? Will post at their forum. Had to change AR from "source" to "disabled: (fit to window)".

[Taktaal]

The k and t are too close together to be seen as a space with your setting of Space Width = 12. Change that to 10 and it'll work.
And the p seems to have been wrongly entered by you. Select it, correct the letter in the text box and press "OK"
The .sup format is a closed source format by Sony and only official licensees get the format description. That's why no other players can have official support.

I tried to convert a .sup which was demuxed by EVODemux.
But i ran into this error, when I opened the file.



So I tried to demux the .sup again with eac3to, but the error remains.

Programs: all latest versions

Any suggestions?

This isn't a problem with demuxing. It looks like that subtitle is using some weird form of subtitle display time correction that I wasn't sure how to handle. Can you upload the .sup somewhere so I can have a look at how to do that?

I still have a problem though with this one (members1.chello.nl/~e.abbink/00011.track_4609.zip) though from Asterix and the Vikings.

It has a cartoony font which totally screws up character recognition (especially separation) but even if I persist and keep filling in by hand the two and three letter combos it just crashes after a while. (Not with any crash log, but with the windows "suprip has encountered a problem and needs to close")

Also often it does not recognize the last character of the line.

I know its only one weird font, but maybe levelling the images before ocr-ing would help? remove the outline and leave only the white inner letter? that would problably help a lot with the character splitting.

The next version will have a contrast setting that can be used to better split fancy outlined letters like in this subtitle.

:confused:
Suprip reads HD-DVD sup files.

I'd like to keep the program fully compatible with HD-DVD. A lot of the functionality is done anyway and there's still a bunch of legacy disks out there that people will have an interest in ripping so they can keep the movies they've paid for even after they sell off their drives.

[quantum]

I'd like to keep the program fully compatible with HD-DVD...
Which is a great thing. My post may have accidentally looked like a question. It could have been worded better this way:
Suprip does read HD-DVD sup files. There is no need to first convert to Blu-ray with Supread.

[gkar]

The k and t are too close together to be seen as a space with your setting of Space Width = 12. Change that to 10 and it'll work.
And the p seems to have been wrongly entered by you. Select it, correct the letter in the text box and press "OK"
The .sup format is a closed source format by Sony and only official licensees get the format description. That's why no other players can have official support.Thanks for the reply, especially on the Sony clarification, much appreciated.

Regarding the incorrect entering of the 'c', does not the image show the word 'weapon' correctly spelt in the main box? I don't think I had to manually enter that particular letter. Can the lower, grey area, where it is spelt incorrectly be different to the main area if one manually types the letter incorrectly?

Finally, for future reference, does the 'space width' box require adjustment depending on the original font used in the Blu-ray movies? Or could '10' be taken as an acceptable dimension for most *.sup files? Thanks.

[gkar]

Having just played around with the settings, adjustment to '9' for 'Space Width' worked perfectly. At 10, words ending & beginning in 't's were joined as well as 'rt's & 'rJ'.

Am right in assuming the following:

'Character Split Tolerance' refers to the separation of letters within any given word?

'Character Similarity Tolerance' refers to the separation of letters that appear similar in type to the Suprip programme? Cheers.

[laserfan]

Am right in assuming the following:

'Character Split Tolerance' refers to the separation of letters within any given word?

'Character Similarity Tolerance' refers to the separation of letters that appear similar in type to the Suprip programme? Cheers.Aren't the pop-up descriptions working for you? Sorry I don't have suprip on this PC or I'd confirm.

I do like that SupRip allows instant change/apply/try for these settings. Some sups are slow-going at first as one tries different settings and trains the program (man all the PGS/fonts from movie to movie are different!), but I've yet to run-across a sub for which it didn't work at all. Thanks to Taktaal!

[nautilus7]

@Taktaal,

Any progress on these (http://forum.doom9.org/showthread.php?p=1229237#post1229237) bugs reported here? They're quite annoying... Thank you.

[FreaQ]

I tried to convert a .sup which was demuxed by EVODemux.
But i ran into this error, when I opened the file.



So I tried to demux the .sup again with eac3to, but the error remains.

Programs: all latest versions

Any suggestions?


This isn't a problem with demuxing. It looks like that subtitle is using some weird form of subtitle display time correction that I wasn't sure how to handle. Can you upload the .sup somewhere so I can have a look at how to do that?

Sorry for the late answer, but here they are.
It is from two titles. Total Recall and Tomb Raider. Both HD-DVDs. One is demuxed with EVODemux and one with eac3to.
http://rapidshare.com/files/181486976/SUPS.7z

[Taktaal]

I just posted a new version.

FreaQ, I think there's something weird with your subtitles. They seem to be intended to display for only 0.07 seconds and don't have any transparency values set at all. I think a friend of mine has the Total Recall disc, so I'll have to check out the complete movie to see what's happening here.

[faza-ron]

I have a problem with the fonts.
When I'm loading a sup file (that was made from original blu-ray with tsMuxeR_1.8.8b) it is show me the requested fonts but when I try to save the file it is not save correctly.
As you can see in the black space, the fonts are correct.
But in the translation there are no fonts, just signs.
What to do?

http://img248.imageshack.us/img248/1488/78742902hp9.jpg (http://imageshack.us)

[nautilus7]

Thanks Taktaal for the update... I'll give feedback soon.

[legoholic]

Thanks for the update! (contrast setting), I was finally able to ocr Asterix and the Vikings.

[nautilus7]

I played around with v1.14 and most problems are now fixed. Thanks!

The remaining bugs i found have to do with lines with mixed normal and italic letters. Is there anything than can be done?

For example subtitles #137, 143, 145, 152, 156 are such.

Also, another similar bug has to do with very sort italic lines which SupRip mistakes them with normal.

For example subtiles #705, 917.

All subtitle numbers refer to this (http://www.sendspace.com/file/iq2700) .sup file.

[Thunderbolt8]

im not sure whether I understand suprip correctly, but Ive got 2 main problems: first is I seem to have to make each subtitle from the start, there no saving option from characters already know form former subs, right? takes just too long that way having to start each sub from the scratch ;/

the other thing is that sometimes (not too often though) letters, which are already given as recognized, are in fact presented wrong and I then cannot correct them, I can only edit those I was asked to enter myself (just checked, might be corrected meanwhile). another thing is then that it also doesnt seem possible to edit the lines in .srt view window seperately, I can delete stuff there, but not insert new letters :/

[magic144]

hi Taktaal, thanks for your continued efforts here

one thing I will say from using 1.14 on 2 .sup files today, the 00:00:00 end-time error seems to be back on the very last subtitle
I have seen this in the 2 files I processed...
00:00:00,-001

thanks again,

m

fyi, here is one of those files
here (http://www.sendspace.com/file/9xozbe)

[nautilus7]

It's not back. It was always there. It just doesn't happen with all subtitles. I was about to report this again with 1.14, but i didn't have a sup that produces this bug.

[magic144]

oh OK - I'd seen it fixed on a sup before so I assumed all was good

thanks for the supporting comments nautilus7 :-)

[faza-ron]

I have a problem with the fonts.
When I'm loading a sup file (that was made from original blu-ray with tsMuxeR_1.8.8b) it is show me the requested fonts but when I try to save the file it is not save correctly.
As you can see in the black space, the fonts are correct.
But in the translation there are no fonts, just signs.
What to do?
]

Any news?
Can someone help me to solve the fonts issue?

[Thunderbolt8]

im not sure whether I understand suprip correctly, but Ive got 2 main problems: first is I seem to have to make each subtitle from the start, there no saving option from characters already known form former subs, right? takes just too long that way having to start each sub from the scratch ;/

the other thing is that sometimes (not too often though) letters, which are already given as recognized, are in fact presented wrong and I then cannot correct them, I can only edit those I was asked to enter myself (just checked, might be corrected meanwhile). another thing is then that it also doesnt seem possible to edit the lines in .srt view window seperately, I can delete stuff there, but not insert new letters :/
anyone able to comment on that? if those things could be handled somehow I would use suprip more often. but either that doesnt seem possible or I just dont know how :S

[MounaLafi]

I was wondering about the possibility of adding support to ARABIC Language subtitles in SupRip.

The problem with Arabic Language that it is difficult for OCR due to the conjunction of its letters unlike Latin Languages. Moreover, the direction of the language is from Right to Left.

Appreciating your feedback.

[jj666]

A couple of mistakes on these files by default:
(all extracted via EAC3TO)

Lowercase 'l' interpreted as 'ι'
.SUP File: http://www.gigafiles.co.uk/files/7846/cars%20%281080p%20bluray%20h264%29.zip

Uppercase 'I' interpreted as 'l'
.SUP File: http://www.gigafiles.co.uk/files/7846/dumb%20and%20dumber%20unrated%20edition%20%281080p%20bluray%20vc1%29.zip

Uppercase 'I' interpreted as 'l'
.SUP File: http://www.gigafiles.co.uk/files/7846/kung%20fu%20hustle%20%281080p%20bluray%20mpeg2%29.zip

Cheers,

-jj-

[alc0re]

A couple of mistakes on these files by default:
(all extracted via EAC3TO)

Lowercase 'l' interpreted as 'ι'
.SUP File: http://www.gigafiles.co.uk/files/7846/cars%20%281080p%20bluray%20h264%29.zip

Uppercase 'I' interpreted as 'l'
.SUP File: http://www.gigafiles.co.uk/files/7846/dumb%20and%20dumber%20unrated%20edition%20%281080p%20bluray%20vc1%29.zip

Uppercase 'I' interpreted as 'l'
.SUP File: http://www.gigafiles.co.uk/files/7846/kung%20fu%20hustle%20%281080p%20bluray%20mpeg2%29.zip

Cheers,

-jj-

For the ocr to correctly determine the difference between lower case L and upper case I on the Arial font is not possible without a dictionary check from what I understand. On the arial font they look exactly the same. Personally I don't care anymore since I'm adding the srt to tsmuxer and outputting bluray structure, and I set my font to arial, so the ocr mistakes between those 2 are non aparent since Im changing the font back to arial when muxing. I believe the author said he's going to eventually try adding some sort of dictionary checking to the orc functionality to correct those mistakes.

[mbcd]

BugReport:

1.14 produces a wrong endtime at the last subtitle in all cases.

Endtime is before Starttime.

01:45:09,094 --> 00:00:00,-001

That makes Problems with MKVMerge, because it skips last Entry

[nurbs]

B1.14 produces a wrong endtime at the last subtitle in all cases.

Not in all cases. I only have this issue sometimes with blu-ray sups. The subtitles of the TV series I am encoding right now have the right endtime.