Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > General > Subtitles

Reply
 
Thread Tools Search this Thread Display Modes
Old 12th September 2005, 13:17   #201  |  Link
chkp45
Registered User
 
Join Date: Sep 2005
Posts: 6
Many thanks for new features, I'll try to test .sup tonight.
chkp45 is offline   Reply With Quote
Old 12th September 2005, 17:08   #202  |  Link
Saligia
Registered User
 
Saligia's Avatar
 
Join Date: Jun 2005
Location: Denmark
Posts: 11
@ai4spam: Thank you very much for the reply.

ABOUT THE 44 LIMIT.
The 44 limit was inded on of the first barrieres I meet when "developing" my method. I did spend some time wondering why it was there. Now I know.

ABOUT THE SCALING AND PROCESSING.
The scalingfactor can be set in the .idx file for the vertical/horizontal resolution, as a percentage. Like say 50% to reduce or 155% to enlarge, to suit your needs and likings.

The number is used by the Avisynth VSFilter.dll, when the file is frameserved. Thus no scaling will be performed when opend directly in Subrip.
And implementing that, would be a whole new world, or rather... -program!
So beeing just a little realistic, it's doable but in my opinion, way off the scope of Subrip, since there are ways of doing this, without rewriting Subrip

ABOUT THE BEST GUESS
I'm sorry to hear, that you do'nt like the idea of an best-guess-automode.
It's true, that it will introduce errors, when best guess is wrong or missing, but looking at pro's and con's I find it easyer to wipe out some errors using a search function, rather beeing tied up to the PC for hours.
But I do respect that this is'nt a matter of sceince, but a matter of opinion.
I still think it would be a big enhancement thoug.
But for now I will use your suggestiens, and do some more tweaking the parameters to optimize the ocr'ing

ANOTHER REQ
Would it be possible to allow editing the output file in Subrip?

LAST BUT NOT LEAST
I fully respect your work, and know that Subrip is'nt like your whole life, and thus there's natural limits to fullfilling the newer ending craving of feautures and options.
So I want to point out, that I find it more importent to appriciate what you HAVE done for us, rather than what will or could be done.

All those hours and hours and hours... -just for free.
That's the truely amazing thing here.

Best regards.
Saligia is offline   Reply With Quote
Old 12th September 2005, 19:23   #203  |  Link
chkp45
Registered User
 
Join Date: Sep 2005
Posts: 6
Well, 50 % success..

Of two programs one converted as it should (code name italiaa), the other (code name prisma) did not when given .txt as input. Both worked when given the first .bmp, but of course the timing was screwed

Samples (a few hundred bmps...) in the same place as last time.

Happy bug hunt (well, not a bug really, bad initial specification rather)
chkp45 is offline   Reply With Quote
Old 12th September 2005, 22:42   #204  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
@Saligia: Actually, the best guess request would be trivial to implement, compared to the resampling, wich can be done without "rewriting SubRip" (but with some substantial effort). So, how about I make the best guess an option, and you help me with the resampling? I need the original subs and a couple of screenshots from the .avi that show exactly what the rescaling does.
Editing the output is not easy, because the list is re-filled each time some operation is performed from the internal structure, so I'd have to allow editing that structure.

@chkp45: Did you check the file names? I mean, do the .bmp files in the directory actually coincide with the file names in the .txt file? I actully ran DVDSupDecode on an example and it worked on the output, so there must be something wrong with the file names.
ai4spam is offline   Reply With Quote
Old 13th September 2005, 00:19   #205  |  Link
Saligia
Registered User
 
Saligia's Avatar
 
Join Date: Jun 2005
Location: Denmark
Posts: 11
@ai4spam: Seems like a pretty good deal to me!

ABOUT DOKUMENTATION.
I'll be more than happy, to put together some dokumentation with screenshoots and all, of what I'm fiddling around with, and what i looks like.
I'll do that by tomorrow. PM or here?
But like I said, I know it will be a huuuge task, to implement - if possible at all. that's why I was'nt to pushy about it.
But I'll let you judge yourself, if it's any good, and worth the effort.

ABOUT BEST GUESS OPTION.
After I read your reply, I did some tweeaking here and there, and I did get some minor improvements. But not even close to what an best-guess-auto option would be. I don't think I was clear enough on this point in my first messesage: The fact is, that when ocr'ing this way the Best Guess is correct 99,99% of the time. So cleaning out a pair of errors afterwards, is a so little price to pay, for the ability to let Subrip do it's magic on its own overnight!

Therefore I'm absolutly exited, that you are willing to give it a try!
You´r to cool to be true.

ABOUT THE SAVE OPTION IN VIDEO FILE VIEWER WINDOW.
I've tryed ticking this option on. I see that i creates a subfolder with .pgm files. But i haven't been able, to find any doku about it. First i thought it would save some images for later processing, but it keeps on ocr'ing AND saving .pgm files. Sorry if it seems like I've been living under a rock, but how is it supposed to be used? And what is a .pgm file?

ABOUT EDITING OUTPUT FILE.
It's no biggie that it can't be done. But as you see, most of the time we have absolutly no idea of what we REALLY ask for. But at least, it's not because we aren't greatfull, of the work beeing done. If we did'nt find Subrip importent and usefull, we proberly would'nt bother asking. It's just because that most of us don't realize the huge amount of effort that goes into even changes that seems, barely noticable to us.


Best regards.

Last edited by Saligia; 13th September 2005 at 01:01.
Saligia is offline   Reply With Quote
Old 13th September 2005, 00:33   #206  |  Link
Esc
Registered User
 
Esc's Avatar
 
Join Date: Jan 2005
Posts: 74
Hey, ai4spam! Thanks for keeping working on it.
I opened the new version and the toolbar battons look kind of strange. Like the pictures do not fit the buttons and go over the edge. A screenshot of 3 of them is attached to this message.
Attached Images
 
Esc is offline   Reply With Quote
Old 13th September 2005, 00:55   #207  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
@Saligia: No rush with the documentation, take your time - I won't be able to do anything until the weekend anyway. Don't go into too much details either, I just need some examples where the "low-res" detection fails, but the upsampled version works. PM or email is fine (lookup the address in the SubRip manual). As for the .pgm files, they were designed for removal using a VirtualDub and special version of DeLogo that I was working on. Unfortunately, debugging it is a royal pain and I simply don't have the time. I've sent the source to one of the members here, maybe he can work on it. This option will probably be replaced with saving regular .idx/.sub pairs, just like PodPis does, if we can get the removal part grafted onto VSFilter to remove subs intead of putting them in, yet still use the .idx/.sub loagind routines.

@Esc: I've changed the button components because the old ones did not have hints in UniCode, so they didn't work with the new UniCode language format. I'll make them a bit larger if needed. Also, I've shortened a couple of captions in the "Enter new char" form to make all the buttons fit in. All the languages need updating.

Last edited by ai4spam; 13th September 2005 at 01:27.
ai4spam is offline   Reply With Quote
Old 13th September 2005, 16:27   #208  |  Link
chkp45
Registered User
 
Join Date: Sep 2005
Posts: 6
@ai4spam: This is really something different.
The 'prisma' set shows bitmaps as black on black initially, and processes nothing. (All bitmaps seem to be read and displayed but nothing is offered to be recognized.

When I go to 'change text color' I see texts in red with blue outline on black background. Then I uncheck black and blue and conversion starts. Trouble is, it shows only the character to be recognized, (blue on black in thin red outline box). Other characters are black on black ??? (This is on main window where tho whole bitmap is shown). You can actually do the conversion but you get errors as the whole text is not visible.

Funny it does not behave like this on the other sample ??

Another minor detail, as these recordings have ridiculously long file names they spoil the apperance of the status line where the file name is shown..

Best regards
chkp45 is offline   Reply With Quote
Old 13th September 2005, 16:58   #209  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
@chkp45: I've looked into it. It seems like I overdid it: the bitmaps are always 4 colors (0..3) even though the palette may have more, and they always refer to shades of gray. I guess I'll just ignore the "Palette=" lines and be done with it.

Last edited by ai4spam; 14th September 2005 at 16:36.
ai4spam is offline   Reply With Quote
Old 15th September 2005, 05:39   #210  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
SubRip 1.40 Beta 2 is now up on the home site.

ChangeLog:
- Fixed a few bugs.
- Added an option to automatically fill the best guess (look in the global options window for how to turn it on).
- Updated Portuguese language by Vítor Vieira

Future plans are to implement something to automatically look in all your fonts for a match, instead of asking you to choose the font. It may take a long time to do (got tons of other things to worry about in my own life).

@Saligia: now it's your turn to hold your end of the bargain .

Last edited by ai4spam; 15th September 2005 at 06:10.
ai4spam is offline   Reply With Quote
Old 15th September 2005, 11:03   #211  |  Link
masken
uhm... ?
 
Join Date: Oct 2001
Location: Gothenburg, Sweden
Posts: 280
@ai4spam, how did it go with the suggestion to allow only whole words to be formatted? (italics, bold etc, the issue where some characters are formatted with italics and some not, causing hard-to-read and large files)?
masken is offline   Reply With Quote
Old 15th September 2005, 17:11   #212  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
@masken: I totally forgot about it . I'll see when I get the time.

Last edited by ai4spam; 15th September 2005 at 18:34.
ai4spam is offline   Reply With Quote
Old 15th September 2005, 17:30   #213  |  Link
masken
uhm... ?
 
Join Date: Oct 2001
Location: Gothenburg, Sweden
Posts: 280
np m8, you're doing a great job just a suggestion
masken is offline   Reply With Quote
Old 15th September 2005, 18:56   #214  |  Link
movax
Member
 
Join Date: Nov 2004
Location: Michigan
Posts: 217
ai4spam, I'll checkout the source and email/PM you some of the changes if I make any :P

*EDIT* PM I guess, your email is secret.
movax is offline   Reply With Quote
Old 15th September 2005, 19:33   #215  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
Quote:
Originally Posted by movax
PM I guess, your email is secret.
No, it's not. Look in SubRip's Credits.txt. Don't use Yahoo Mail if you send attachments. Mark your changes for ease of integration.
And remember: I am to blame for only part of the mess in the code .

Last edited by ai4spam; 15th September 2005 at 23:29.
ai4spam is offline   Reply With Quote
Old 17th September 2005, 20:59   #216  |  Link
giannid
Registered User
 
Join Date: Nov 2004
Posts: 30
Hi,
I opened an .srt file in the 1.40 Beta 1, and I tried to use the "Post OCR Spelling Correction", but this option was not available. I think it's available from a fresh rip but not for an old, previously ripped .srt file.
giannid is offline   Reply With Quote
Old 17th September 2005, 23:22   #217  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
The first thing that happens when you open an old file is that SubRip asks you what CodePage/CharSet to use to convert to UniCode. Chose the values you want, then press "Done". Then all the other options will be enabled again.
The whole purpose of this is to deal with non-standard CharSets. For example, for Eastern European, if you don't choose that CharSet, some UniCode characters (the diacritics) will be converted to the wrong values. The reason the text shows is so that you can try different CharSets until you are satisfied with the conversion.
Anyway, I recommend other tools for better OCR and spelling correction, like SubtitleWorkshop.
ai4spam is offline   Reply With Quote
Old 19th September 2005, 06:57   #218  |  Link
ai4spam
Programmer
 
ai4spam's Avatar
 
Join Date: Sep 2003
Posts: 378
Quote:
Originally Posted by Esc
A screenshot of 3 of them is attached to this message.
Well, it's been pending approval for a while, so I guess you need to send it over via email or put it on the web someplace.
ai4spam is offline   Reply With Quote
Old 20th September 2005, 21:17   #219  |  Link
TiaoMacaleh
Registered User
 
Join Date: Aug 2004
Posts: 24
Anyone have problem with subrip not recognising whole lines?

I tryed changing colors etc, even it looking perfect it doesnt recognise the line. Some subs have more than others this problem... sometimes have to type the whole line dialogue 20-30 times...

Is there any way to solve that? maybe a config etc?


thanks
TiaoMacaleh is offline   Reply With Quote
Old 20th September 2005, 22:17   #220  |  Link
LeMoi
Registered User
 
Join Date: Sep 2004
Location: France
Posts: 322
On french windows, with default XP theme, when i save subs, i can see the end of the last charsets, so i don't know if they are utf-8 or utf-16 or other, i just see until UTF
and with french language selected, i can't the char proposed by the soft, it appears under the words "meilleure réponse" :|
LeMoi is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 16:58.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2017, vBulletin Solutions Inc.