Subtitle Edit

[von Suppé]

I'm gonna try this out too, sounds promising.

Yes, you rule !:)
But there is a minor problem : curiously, when I open the sup with modified timecodes (thus after having imported new time codes and exported in Blu-ray sup), the timecodes are all displayed 44 ms sooner than those in the srt used as new time codes.
For example, if the time code of the first line of the good srt is ;
1
00:00:07,921 --> 00:00:12,241
I get this in the corrected sup ;
1
00:00:07,877 --> 00:00:12,197

But as I can change srt's timecode easily with SE to compensate this little change, that's not a big problem.;)

Or, after importing the new timecodes, you could also change all their in/outs manually :D

[Betsy25]

The "Fix common OCR errors" is a bit too greedy, for example in the example below, it will propose to replace "for" with "For", however this should not be the case.

Any idea how to make the method stay away from automatically capitalizing the new subtitle line when the previous one ended with some "end of sentence" string ? There are quite a lot of "false positives" by default.

43
00:08:07,379 --> 00:08:11,033
I warned their president...
- Say it John, Say it !

44
00:08:11,133 --> 00:08:13,173
for doing business with them.

[von Suppé]

Nikse555, is it possible to add this to the "Fix common error tools":

If there are 2 lines, and both begin with a dash, being able to remove the dash (and the following space, if it's there) in the first line?
This would be as opposite of the tool "Fix (add dash) line pairs with only one dash (-)"
Maybe call it "Fix (remove first dash) line pairs with two dashes" or something?

Example:

- Hi, how are you?
- I'm fine, thanks.

-->

Hi, how are you?
- I'm fine, thanks.


Thanks for your hard work :)

[Betsy25]

Proposition for the "Fix Unneeded Spaces" :

In case of <somestring><DASH><space><nextstring>, when <nextstring> is "and" or "or", do not remove the <space>.
("en" or "of" in Dutch)

Examples :

(English)
What are your long- and altitude stats ?
There is an X- and a Y-chromosome.
Did you buy that first- or second-handed ?

(Dutch)
Wat zijn je voor- en familienaam ?
Was het in het voor- of najaar ?

I've commited some brewing on my github (https://github.com/Betsy25/subtitleedit/commits/My-fixes2) that probably is far below your coding standards, but perhaps you can refine it.

[Ghitulescu]

If there are 2 lines, and both begin with a dash, being able to remove the dash (and the following space, if it's there) in the first line?
This would be as opposite of the tool "Fix (add dash) line pairs with only one dash (-)"
Maybe call it "Fix (remove first dash) line pairs with two dashes" or something?

Example:

- Hi, how are you?
- I'm fine, thanks.

-->

Hi, how are you?
- I'm fine, thanks.
)
Why would be this a correct formatting?
Yes, I've seen this also on commercial movies, but I still have no explanation why one line is treated differently than the other one, although both are dialogues ...:confused:

[von Suppé]

I'm not saying it's correct formatting, if there is any. I and many people just like it this way and find the first line dash not necessary.
I myself do a lot of manual subtitling and there's a lot of feel to it, which I find more important than correct formatting or not. A first dash disturbs my feel to the text in most cases.

[Betsy25]

I'm not saying it's correct formatting, if there is any. I and many people just like it this way and find the first line dash not necessary.
I myself do a lot of manual subtitling and there's a lot of feel to it, which I find more important than correct formatting or not. A first dash disturbs my feel to the text in most cases.

the - normally "tells" the viewer that a different person than the one who told the last line is now speaking, so it's perfectly normal to have both lines beginning with - in that case. (at least, that's the way I understand it):)

[von Suppé]

Yes, I know what you mean.
If one would consist in that way, in every new subtitle display where the speaker of the first line is another than the last one in the previous subtitle display, how many cases would you have to start with a dash to tell the viewer this? It feels unnatural to me to do so.

Example:

Display 1:

Son, as I don't have much time today,
will you do me a favor and wash the car?

Display 2 would have to be like this:

- I will in the afternoon dad, because
I promised mom to fix her computer first.

Or similar things. Doesn't feel right. But everybody should of course do as they see or feel fit.

[TheSkiller]

I agree with von Suppé.
Whenever I'm creating subtitles myself I put the dash (which in my humble opinion should always be a so called "em dash" —, not a "hyphen" - ) in front of the second line, in other words exactly where the change of the speaker happens within the same subtitle.
I also think it's easier to follow such a subtitle if both lines are left-aligend but as a whole the subtitle should of course be centered within the video frame, with the top line not starting before the em dash of the bottom line.

http://picload.org/thumbnail/lwgpdww/subtitle_speakerchange_with_em.png (http://picload.org/image/lwgpdww/subtitle_speakerchange_with_em.png)
(click to enlarge)

That's what I personally consider "most aesthetic" and easiest to read.



I'm quite sure the main reason for putting a dash in front of both lines is rather just a technical shortcoming because it is easier to format this way. But of course it also depends on what the viewer is expecting and used to.

[mood]

@Nikse

It is possible add boxing options to context menu???

[von Suppé]

@ mood

Can you tell me what "boxing" is?

[mood]

@von Suppé

is in version 3.4 change log.

boxing is this:

http://i.imgur.com/xlNf85L.png


you can add tags manually (<BoxSingleLine>) and (<BoxMultiLine> to the lines to get the effect shown in the image above.

is a black box on background subtitle, you can have lines with black box on background and without them in same subtitle.

this only on sub image based.

this is great, but only can do this manually or in export window.

And in export window not work very well because the effect only work in one line whatever more lines you add the effect

I use this in idx/sub and I need this on context menu like italic tag and others

this effect is useful when you have hardsubs on video.

[von Suppé]

I understand. Thanks for clear explanation. :)

[Nikse555]

About the dialogues and hyphen, try File - Plugins - "Dialogue one hyphen only".
The rules regarding one or two hyphens might be different depending on language I think...

@mood: In export you can choose default "Border style" incl. "box single line" or "one box".
The list view (and hence the list view context menu) in the export will be improved with Ctrl+a for select all, ctrl+d for de-select + ctrl+shif+i for inverse selection. Changing selection will also be much faster (was very slow previously).

[Betsy25]

Nikse, you're willing to take a look at my 2 latest brewings here : https://github.com/Betsy25/subtitleedit/commits/My-Fixes-2

I'm a complete noob in C#, but I'm pretty sure you'll understand what I'm trying to do. (Regarding the dash cases followed by and or or)

[mood]

@Nikse555

thanks, but the "Border style" applies to all lines and I do not want to apply the style to all lines, only the lines that I choose.

In export window the boxing effect only work for one line, if you applie more than one line, the effect you applie on line before disappears, the tags <BoxSingleLine> or <BoxMultiLine> still there but without effect.

[Nikse555]

@Betsy25: I'll take a look... sorry I've looked earlier (was on vacation).

@mood: thx, a bug - hopefully fixed now (on Github and in this portable beta - http://www.nikse.dk/SubtitleEdit.zip )

[mood]

@Nikse555

thanks, now work but I hope you add this style to context menu ;)

[Betsy25]

@Nikse,

when you have the time, will you take a look at your PM's here on Doom9 forum please ?

[mariner]

Greetings Nik.

Having trouble getting the BD sup to follow the font settings defined in ass subtitles. The preview panel seems to work fine, but the sup created uses only one font. Is this a known issue?

Many thanks and best regards.

[Nikse555]

Having trouble getting the BD sup to follow the font settings defined in ass subtitles. The preview panel seems to work fine, but the sup created uses only one font. Is this a known issue?

Thx for reporting this :)
Should be fixed on github + in this portable beta: http://www.nikse.dk/SubtitleEdit.zip

[mariner]

1. Thanks for the quick fix, Nik.

2. There appears to be a problem in the handling of overlapping subs, some ending up with 0 duration.

Many thanks and best regards.

[Nikse555]

1. Thanks for the quick fix, Nik.

2. There appears to be a problem in the handling of overlapping subs, some ending up with 0 duration.

1) np
2) where and doing what?

[mariner]

It seems the overlapped graphic elements need to be combined into one.

[Music Fan]

@ Nikse555 : did you find where does this little problem come from ?

there is a minor problem : curiously, when I open the sup with modified timecodes (thus after having imported new time codes and exported in Blu-ray sup), the timecodes are all displayed 44 ms sooner than those in the srt used as new time codes.
For example, if the time code of the first line of the good srt is ;
1
00:00:07,921 --> 00:00:12,241
I get this in the corrected sup ;
1
00:00:07,877 --> 00:00:12,197

But as I can change srt's timecode easily with SE to compensate this little change, that's not a big problem.;)

I talked about this in post #200 ;
http://forum.doom9.org/showthread.php?p=1686798#post1686798

[Nikse555]

@mariner: Yeah, I guess... perhaps a warning would be nice...

@Music Fan: Hm, perhaps to give the decoder a tiny bit extra time to decode the subtitles? I don't know ;)
The code is here: https://github.com/SubtitleEdit/subtitleedit/blob/master/src/Logic/BluRaySup/BluRaySupPicture.cs#L45

[Betsy25]

Perhaps a bug.

In case the user has unchecked "Start with uppercase letter after paragraph" in the Fix common Errors dialog (and the other "start with uppercase" choices), some items are still detected and presented to be Uppercased using the "Fix Common Errors (using the OCR replace list)" method.

For example, subtitles 444 and 1047 in the subtitle file below.

444 --> als ie tenminste nog capabel was...
1047 -> b kwadraat?

Subtitle : Dutch .SRT subtitle (https://www.dropbox.com/s/kgocs0sixny85e7/Into%20the%20Wild%20%282007%29%20%281080p%29.ned.srt)

[Thunderbolt8]

Maybe a bug, for some reason I couldnt delete an entry from the OCR fix list which I had added myself before. had to edit the .xml file to get rid of it.

[Nikse555]

@Betsy25: Hm, yes. "Fix common OCR errors" also changes first letter to uppercase...

@Thunderbolt8: I could not re-create this... let me know if it occurs again or if you can re-create it.
Also, changes to interjections require the 'Remove text for HI' window to be closed and re-opened for the changes to take effect... will be fixed next update.

[Thunderbolt8]

in the remove text for hearing impaired menu, would it be possible to add exceptions for removing text between ( ) and { } in case of .ass files? the screen position in some .ass subs is indicated between these markers. perhaps when such marker both occur in the same line or when the ( ) brackets are inside the { } ones?

currently, its like this:


{} ticked: {\an5\pos(958,176)}JACKIE: <i>It seems as though</i> -----> <i>It seems as though</i>
() ticked: {\an5\pos(958,176)}JACKIE: <i>It seems as though</i> -----> <i>It seems as though</i>

--> should be: {\an5\pos(958,176)}<i>It seems as though</i>


{} ticked: {\an4\pos(705,833)}<i>Today really marks</i> -----> <i>Today really marks</i>
() ticked: {\an4\pos(705,833)}<i>Today really marks</i> -----> {\an4\pos}<i>Today really marks</i>

--> should be: {\an4\pos(705,833)}<i>Today really marks</i>


{\an4\pos(556,902)}Huh? -----> {\an4\pos(556,902)}

--> should be: whole line should get deleted because its empt after the position information after SHD removal.

{\an4\pos(814,825)}- [ Panting ] -----> {\an4\pos(814,825)}-

--> should be: whole line should get deleted because theres only a hyphen left after SHD removal, but line doesnt get removed because the subtitle screen position information is considered as text information.

[Nikse555]

@Thunderbolt: thx for reporting this... I probably use srt files most of the time.
Should be fixed on github + in this portable beta: http://www.nikse.dk/SubtitleEdit.zip (also contains the interjections-fix)
Let me know if you find more stuff :)

[Betsy25]

@Betsy25: Hm, yes. "Fix common OCR errors" also changes first letter to uppercase...


So, there's no way to stop it from (false positive) uppercasing, without giving up the whole "Fix common Errors" package ?

[Thunderbolt8]

there are a few things I still noticed. some of them to come at a later point. but for now:


- if there are double hyphens "--" in line initial position already present before SHD removal has been run (e.g. --the wheather will be...) then those double hypens shouldnt be changed and taken into consideration for SHD removal.

e.g. "SPEAKER: --the wheather will be...." in this case the double hyphens should be disregarded from removal. this should only apply with cases in which the double hyphens were present from the start, not as a leftover result from a previous SHD removal (i dont know how the SHD removal stage works, if there are for examples 2 runs you do after another when pressing button only once)



- hyphens and other SHD information in a subtitle line (for .srt) should be taken into consideration if that subtitle line for example consists of two or more lines on screen.

e.g
- It's already done. (- It's already done.|BENJAMIN: You--)
BENJAMIN: You--

gets changed to
- It's already done. (- It's already done.|You--)
You--

while it should get changed to
- It's already done. (- It's already done.|- You--)
- You--

(this example can get fixed though with "fix line beginning with dash (-)" in the fix common errors tab, but that option might not be activated, possibly because it could interfere negatively in other situations within the same file. so it would be nice if this situation could be resolved already during SHD removal)



- same as above in case of .ass subtitles, if you have 2 or more subtitle lines with identical timestamps, meaning they belong together on screen.

e.g.
00:04:40:22 00:04:43:23 {\an4\pos(775,828)}- Hey, hey!
00:04:40:22 00:04:43:23 {\an4\pos(775,900)}- [ Laughing ]

gets changed to
00:04:40:22 00:04:43:23 {\an4\pos(775,828)}- Hey, hey!

while it should get changed to
00:04:40:22 00:04:43:23 {\an4\pos(775,828)}Hey, hey!



- in case of .ass subs apparently brackets like [] or () sometimes arent applied in each line, but each block with same timestamps which belog together

00:29:38:23 00:29:41:11 {\an4\pos(1177,684)}[ Ship's Horn Blows
00:29:38:23 00:29:41:11 {\an4\pos(1178,756)}In Distance ]

this get ignored by SHD removal (and "Fixing missing [ in line" only fixes the 2nd line, but doesnt list the first one for some reason), maybe these situations can be added to SHD removal as well in case of .ass subs.


it might already get tricky now that implementing new rules could interfere with other rules ;)

[mariner]

Greetings Nik. A few issues with BD sup rendering:

1. In a srt sub, if the italic tag is preceded by space after a line break, the placement of that line is off.
2. Can't seem to get the font formatting defined by AEGISUB style editor to work. Are these supported by SE?
3. Fade doesn't seem to work as well.

Thanks for looking into these issues.

[Thunderbolt8]

in which file are the user edited/added information of the "Remove interjections" list stored?

[Nikse555]

in which file are the user edited/added information of the "Remove interjections" list stored?
In Settings.xml... in a tag called "Interjections".


@mariner: thx for the info about (1) italic tag is preceded by space after a line break... will be fixed in next update.
(2) basic stuff like font style, color should work...
(3) I don't know anything about fade.

[von Suppé]

Hi Nikse555

Is it possible to make SE show the subtitles as SUP preview in the videopreview window? Preferably with being able to adjust SUP settings (font, size, position etc.) for all - or selected only.

I think I'm asking a lot here... :rolleyes:

cheers
von Suppé

[Thunderbolt8]

the "Fix common errors" window always opens on my primary screen, even though I have the program on my 2nd screen and dragged the window over already before. could you please fix that? other windows also open at that screen at which they have been closed down last.

[Thunderbolt8]

I added "Ha!" to the remove interjections list, but subtitle lines just consisting of "Ha!" dont get removed by that. bug?

edit: well it works if I dont include the exclamation mark.

edit²: but then again a line consisting of "- Ha!", resulting from a previous remove interjections run, does not get removed.

[Music Fan]

rev 3.4.2 ;
https://github.com/SubtitleEdit/subtitleedit/releases

[kalehrl]

Hi Nikse
I keep getting this error:
Message: Exception from HRESULT: 0x80040265
Source: SubtitleEdit

StackTrace:
at QuartzTypeLib.IMediaControl.RenderFile(String strFilename)
at Nikse.SubtitleEdit.Logic.VideoPlayers.QuartsPlayer.Initialize(Control ownerControl, String videoFileName, EventHandler onVideoLoaded, EventHandler onVideoEnded)
at Nikse.SubtitleEdit.Logic.Utilities.InitializeVideoPlayerAndContainer(String fileName, VideoInfo videoInfo, VideoPlayerContainer videoPlayerContainer, EventHandler onVideoLoaded, EventHandler onVideoEnded)
I already have the latest LAV Filters installed.

[Nikse555]

Hi Nikse
I keep getting this error:

I already have the latest LAV Filters installed.

Is it only one video file that gives this error or all videos files?
LAV Filters cannot decode all video files so you might switch to VLC media player for special files like 10-bit color or similar.

Also, from version 3.4.0 SE runs 64-bit on 64-bit operating systems which will require 64-bit lav filters too.

[kalehrl]

I have Combined Community Codec Pack installed and I guess it ships 32bit LAV filters.
I installed 64bit version of the filters, and now SE works.
Thanks

[Thunderbolt8]

theres still problem with hearing impaired stuff removal in .ass subs and the "remove text before a colon" option. when the "only if text is UPPERCASE" box is ticked as well, then the information about the speaker is not removed:

{\an4\pos(1335,891)}NIC: Shh! --> {\an4\pos(1335,891)}NIC:

or {\an4\pos(691,748)}WOMAN: (OVER PA) --> {\an4\pos(691,748)}WOMAN:

(both lines should be removed entirely as they dont include any speech)


and when I leave the "only if text is UPPERCASE" box unticked then everything before the colon is removed including the screen position information part of the line:

{\an4\pos(652,817)}NIC: "Your hack into MIT --> "Your hack into MIT

(screen position information needs to be preserved and not taken into consideration for removal if there is still speech present in the line)

[minhjirachi]

Does anyone know which software can make the effect for the subtitle? Like Fade In Effect or Fade Out Effect.

[Superb]

Does anyone know which software can make the effect for the subtitle? Like Fade In Effect or Fade Out Effect.Aegisub... for example... http://docs.aegisub.org/3.2/ASS_Tags/

[tuco76]

Does anyone know which software can make the effect for the subtitle? Like Fade In Effect or Fade Out Effect.

PotPlayer is sort of disneyland for the habitual subtitle tinkerer ;)
Standard Fade effect, Fade in/ out adjustable strengt 1-100%,
and an almost ridiculous lot of else subtitle settings, selections/ adjustments/features/gimmicks on top of that. Be aware! :)
You prefer it rather basic and readily comprehensible-
better dont bother with :)

[Mole]

Do you think it'll be possible to open several subs at the same time?

For example, sometimes I work with subs from TV series, and I'd like to be able to spell check them all with exactly the same settings, so that certain weird name not in the dictionary can be skipped automatically on all the subs or certain words be renamed on all subs. But I'd rather not add them all permanently in the program because occasionally there may be spelling variations on a particular name, which may be used in one movie, but spelled differently in another. So I want to manually specify which spelling I want to use.

This is often the case for us who work with English subs for non English movies.

Maybe each sub can be in it's own tab or something.

Also, you should separate word list which we have added manually into a dedicated XML file, so that each time we upgrade your program we can simply just copy this XML file over.

Currently all are stored in the eng_OCRFixReplaceList and each time I upgrade your program, I need to manually add in my own stuff in there, because each new version will write over this file.

Problem is that all the words will be sorted alphabetically in eng_OCRFixReplaceList, and often I don't remember which one was my own.

Hope you understand what I'm trying to say.

[minhjirachi]

I think the subtitle edit have a little bug with sub color. Check this script:

VAV STUDIO Hân Hạnh Giới thiệu
Bộ phim: <font color=yellow>"CUỘC CHIẾN LUÂN HỒI"</font>

Export as XML/PNG and *.SUP will have different results. XML/PNG get the correct color but can't import to Scenarist without BDSup2Sub and have a little bug about time stamp. Export as *.SUP will have bug about the sub color.

Hope you can fix it as soon as possible.

[Music Fan]

Hi Nikse,
I have a lot of problems with DVB-SUB (included in TS) ;

1)a lot of ts can't be opened ("unhandled exception")

2)with those which are opened, the size after conversion in Blu-ray sup without OCR is completely different (much more little) than what I see when I play the original TS in MPC-HC and VLC (which both handle DVB-SUB). But for SE the size of the lines are the same before and after conversion : I mean that just after parsing the ts, if the detected height of a line is 33 (in the OCR window), the size will still be 33 for this line when opening the Blu-ray sup created from this ts, thus I don't know what is the real size and why they are different for MPC-HC and VLC while they are the same for SE.:confused:
And when put on AVCHD and played with a standalone player, the size of the Blu-ray sup is also very little, same size than what I see in MPC-HC and VLC (when playing remuxed TS with the Blu-ray sup).
But I'm sure that the original TS is correctly displayed by MPC-HC and VLC because they display the same subtitle size than my DVB recorder.
That's why I'm sure SE reduce the size of these DVB-SUB, but why ... :confused:

3)when I make OCR with Tesseract, a lot of words are not correctly identified while they are well readable (but maybe a little bit too tiny for Tesseract)

4)even when they are correctly identified, I have to skip them all one by one because SE don't know these words while they are very common :o