Subtitle Edit 4.0.4

[Nikse555]

@JayJayH: Sorry, SE only supports a simple preview of the subtitle (no position + only one sub).

@jpsdr: You could try editing Settings.xml and change "ShowBetaStuff" to true - and then try the "New image compare" ocr...

@Thunderbolt8: thx, the "remove text for HI" issue should be fixed in latest beta: http://www.nikse.dk/SubtitleEdit.zip (portable version)

@minhjirachi: Cannot re-create .sup creating crash - what OS are you using?
I don't have Scenarist... any idea why it does not work?

Also, SE 3.4.6 is out (well, 18 days ago)

[jpsdr]

Is there a way to get the french dictionnary ? The PC where i'm using SE is totaly offline, and will never be connected.

[Betsy25]

Originally Posted by jpsdr

Is there a way to get the french dictionnary ? The PC where i'm using SE is totaly offline, and will never be connected.

You could try the local library, but I don't know if that one will be of much help here.

Otherwise, while you might be connected, Spell Check / Get Dictionaries... / French / Download.
You can then copy them on a USB and put them in the dictionary folder on the unconnected PC, "Open dictionaries folder..." copy/paste the "fra*.*" and "hyph_fr*.*" files.

[Nikse555]

Yes, do like Betsy25 said - or do it manually:

1) find .oxt file from libre office (or open office), eg. http://extensions.libreoffice.org/ex...es-fr-v5-3.oxt

2) Download the file and rename the .oxt file to .zip

3) Unzip the file

4) copy the contents of the "dictionaries" folder to SE's "Dictionaries" folder.

[jpsdr]

Finaly, i've installed SE on a PC with connection, downloaded langage i needed for both things (Teressac & dictionnary) and copied it on USB stick and put them on the other PC. Gladly, it worked. I was afraid there could be some things set in registers or others, but apparently not.

Thanks Nikse555 for this very usefull tools, handling a lot of formats, and i've tested Teressac with the correct langage on sub/idx, results are very good, even excellent.

[JayJayH]

Hi Nikse555, Thanks for your reply at post #341.
I guess what you're saying is that all versions of SE which are later than v3.3.15 no longer support previews of .SSA subtitles showing where they will finally display.
The screen-shots (taken from SE v3.3.15) attached to my post #337 clearly show that it used to happen with .SSA subtitles in that (and the earlier versions if I remember correctly) version - please see the (upper) RED and (lower) GREEN subtitles.
It seems that feature is now "lost" in the later versions - all that they now display is the WHITE subtitle preview (again, see the screen-shot)

[minhjirachi]

@Nikse555: I'm using Windows 8.1 64-bit. I have known why the Subtitle Edit got crash. Because it meet the too long line. With the old version (e.g 3.4.2), it forced export to sub and show up the lines, which are too long.

With the Scenarist, please view this picture:

[ndjamena]

I've been bashing away at a little program to convert 608 captions taken from m4vs to srt. I'd like to upgrade to Advanced SubStation Alpha, but given the lack of background colours and positioning in the srt format I've pretty much done everything I can for now (other than Flash On and alternate channels). The code is a complete shambles (I was learning as I went) and completely unoptimised but on comparing the SRTs played on VLC to the original captions played through iTunes they're pretty much exactly the same.

Then I discovered Subtitle Edit could load 608 from an m4v. Sadly, it seems to be rather inferior to my programs abilities at the moment, hopefully that can be fixed though.

This is the first few lines as Subtitle Edit extracts it:

1
00:00:07,007 --> 00:00:09,041
HI, THERE. JUST WANTED
TO WELCOME YOU TO MY SHOWSTARRING ME--KUZCO.

2
00:00:09,042 --> 00:00:11,844
SO, NO CHANGING
THE CHANNEL. UNDERSTAND?

3
00:00:11,845 --> 00:00:13,212
NO CHANGING.
OK. THEME MUSIC.

4
00:00:14,982 --> 00:00:16,883
♪ HE'S ON HIS WAY
TO THE THRONE ♪

This is what my program spits out:

1
00:00:04,703 --> 00:00:07,007
HI, THERE. JUST WANTED
TO WELCOME YOU TO MY SHOW

2
00:00:07,073 --> 00:00:09,042
STARRING ME--KUZCO.

3
00:00:09,109 --> 00:00:11,845
SO, NO CHANGING
THE CHANNEL. UNDERSTAND?

4
00:00:11,911 --> 00:00:14,981
NO CHANGING.
OK. THEME MUSIC.

5
00:00:16,950 --> 00:00:18,551
♪ HE'S ON HIS WAY
TO THE THRONE ♪

I'm pretty sure mine is correct, some of the conversions Subtitle Edit produces don't actually make sense. It misses line breaks for some reason I can't fathom and seems to treat each frame as a frame... which is wrong.

Anyway, if anyone has m4vs they'd like to try my program on, you first have to extract the captions with mp4box [-nhml TrackID:Full] then drop the .media file onto the EXE. The srt should be forthcoming.

http://www.mediafire.com/download/yc..._Final_Cut.exe

It was written in c# and it needs to be completely rewritten, I need to add Flash On (although I doubt iTunes would ever use it) and try to figure out how ASS subtitles work so I can try retain the positionings (and maybe the colours as well).

Getting Subtitle Edit to do it for me would work too...

[Nikse555]

@JayJayH: If you use "DirectShow" video player in SE (default) and install DirectVobSub/VSFilter (use 64-bit if you have 64-bit os) then you should have a "preview" (it's not an intentional feature

@minhjirachi: I'll do some testing on win 8.1 (my normal computer runs win 7). Also, what happens if you convert all images to 8-bit before importing into Scenarist?

@ndjamena: cool I do know SE has some issues regarding/extraction 608/mp4 . If you want to help improve SE perhaps you could share the source code? Also, you can check SE mp4 extraction + 608 decoding - e.g. start from here: https://github.com/SubtitleEdit/subt.../Boxes/Stbl.cs
(I've not seen any nice documentation about either mp4 subtitle extraction or c608)

[minhjirachi]

Originally Posted by Nikse555

@minhjirachi: I'll do some testing on win 8.1 (my normal computer runs win 7). Also, what happens if you convert all images to 8-bit before importing into Scenarist?

After using the BDSup2Sub to convert 8-bit images. The Scenarist can import them fluently.

[ndjamena]

For what it's worth (which isn't much) here's the source code:

http://www.mediafire.com/download/5g...8_Captions.zip

Bear in mind, originally I was just trying to split the track into files containing individual frames so I could have a look at them, then I tried translating each of the "commands" into strings I could output into a text file, then I though I'd try applying the commands to a buffer and finally I though I'd try converting the buffer into an srt. So it's just one thing built onto another, I hadn't really figured out how any of it worked until I'd got a few srts out of the way so I had to patch it up on the way and make it do things it wasn't designed for, and I'm pretty sure there's still some things I'm missing. Basically the code is junk, testament to my learning process but otherwise rather worthless and needs to be rewritten from scratch. (I've started redoing the buffers so I can swap them easier and use the second channel but haven't gotten very far yet, so there's some unused code in there too.)

It's not obvious from any of the documentation I've found, I assume they don't bother mentioning it because when used with it's original transmission medium (ie analogue NTSC) it should be fairly bloody obvious.

http://en.wikipedia.org/wiki/EIA-608

OK,

It uses a fixed bandwidth of 480 bit/s per line 21 field for a maximum of 32 characters per line per caption (maximum four captions) for a 30 frame broadcast.

480bps / 30 fps = 16 bits = 2 bytes

So basically, line 21 of each field contains 2 bytes worth of information, which means each 2 bytes in 608 captions, regardless of how it's stored or how it's being transmitted, increases the current timecode by 1/30 of a second (which is when the next even field begins) and every 608 "frame" (2 bytes) is a P-frame that inherits the state from the previous frame. (Captions update 30 times a second, to convert to srt you need to figure out where the actual "display" memory changes, capture it's current state and convert THAT to srt, rather than the commands themselves.)

The way MP4 stores the captions is misleading, there are no frames in 608 beyond the 2 bytes on each line 21, which leaves the mp4 frames themselves as having durations of [Number_Of_Bytes / 60] seconds. If the timecode of the current mp4 frame is later than the end of the last mp4 frame then that's the equivalent of the time in between the end of the last frame and the beginning of the current frame being filled with nulls. If the end of the last frame overlaps the beginning of the current frame then that's an authoring error.

[end of caption] SWAPS the displayed memory with the non-displayed memory. So basically, if I load the words "Rabbit Season!" into Non Displayed Memory [resume caption loading] then send [end of caption] the contents of the Non Displayed Memory and the Display Memory will be swapped and "Rabbit Season!" will be displayed on the screen. If I then load "Duck Season!" into Non Displayed memory and then send [end of caption] it will now display "Duck Season!" if I then wait a few seconds and send [end of caption] again, the buffers will be swapped and it will display "Rabbit Season!" again. I don't know if iTunes will ever use it like that, but they can if they want to.

The only way to move to a new line is to explicitly tell it to, if you've filled the last column in a row (column 32) and are told to display a new character, the standard is to replace the character in column 32 with the new character and keep doing that until you receive a command to move the cursor. If they send a [backspace] command and you're in column thirty two, the recommendation is to determine if you've just written to column 32, if you have delete the contents of column 32, otherwise delete the contents of column 31. (I keep forgetting that I've left a bug in my code in regards to that, my code sets the column number to -32 once column 32 has been written in to, I was supposed to add math.abs functions to all the array coordinates so they don't error out but haven't got round to it yet [FIXED].)

[resume direct captioning] writes the following commands directly to the display buffer. So if someone's screaming the same word over and over with increasing volume, you can pop up a caption saying "You!" the first time, then for the second, switch to Direct Captioning and send just a "!" to make it "You!!" and then send another "!" each time they yell, again I'm not sure iTunes would use it that way, but they can.

Most of my code was written with nothing but Wikipedia as a reference, I did find another document that filled in a few blanks, but it's a PDF and it wastes so much effort explaining the differences between how each generation of 608 decoders handle each command that it's hard to find anything useful in there. It doesn't help that they've denied copying permissions, so I can't copy/paste important parts or convert it to a Word Document to make finding things easier.

Anyway, that's most of it. The positioning data makes flawless conversion to srt impossible in every possible case, I can imagine situations where it would be almost impossible to figure out the order in which words are said without user intervention:

Hello |Hi,
How are you? |I'm fine thanks.
{Sorry, the forum is removing all the spaces separating the captions}


It would make perfect sense as a caption when you're watching the video, I could probably think of worse and more likely situations if I tried. If it was a comedy and they were deliberately playing with the captions maybe...

{If left to me my program will likely never get finished, my head's not the most stable construct on Earth and I'm surprised I got this far}

{I had more, but my head hurts so I'll stop now.}

-edit- [Replace 60fps with (60000/1001) fps if you like, or 30 fps with(30000/1001) fps]

[Thunderbolt8]

Originally Posted by Thunderbolt8

some more SHD removal fixes:

WOMAN: <i>Mr. Sportello?</i>|- Mm-hm.

gets changed to:

<i>- Mr. Sportello?</i>

but should get changed to: <i>Mr. Sportello?</i>

--> the WOMAN: speaker information part is not taken into consideration for SHD removal in combination with the hyphen from the speech part after the line break.

Originally Posted by Nikse555

@Thunderbolt8: thx, the "remove text for HI" issue should be fixed in latest beta: http://www.nikse.dk/SubtitleEdit.zip (portable version

hey, thanks. however, could you please also include the special "—" hyphen character in all those hyphen removal related rulesets as well? Current example is:

Uh— Well, I, uh—

gets changed to

— Well, I —

its correct for the last part in which the hyphen has to remain after the ", uh" removal (because the "I" is still real speech information), but is not taken into consideration as part of the 'no real speech' SHD stuff with which it should be removed together if thats the only thing remaining, as seen at the first part of that line.

[ndjamena]

Grrr, I discovered CCExtractor could extract captions from MP4, so I thought I'd give it a try. I figured PGS would be the best output format to keep the positionings, there's an output format called "spupng" which looked promising, it created and xml and a bunch of png images of the captions, but so far I haven't found anything that can even open it, much less convert it to pgs. So I thought I'd try simple srt output, I tested it on a file I'd already converted using my crappy little program... AND THEY GOT THE DAMN TIMECODES WRONG. I know it's wrong, because I can play the files synced with iTunes, mine lines up exactly, theirs doesn't. It's CCExtractor!!! WTF???

Unfortunately I've encountered my first 608 with Roll-Up captions, for one thing they're not what I thought they were, for another it makes mincemeat of my program. I looked at my code and got embarrassed enough to remove it from mediafire, I'll pretend it was never there. I've started rewriting it into a single neater 608 Captions class... I doubt I'll manage to finish it though.

RECOMMENDATION: Service providers shall calculate whether a Backspace is being issued
before or after display in Column 32. If a character or mid-row code has been placed in Column 32,
send a transparent space or Delete to End of Row command (DER) to erase the 32nd column. If
Column 31 is also to be erased, send the transparent space or DER first, then the Backspace.
In general purpose captioning, the Backspace command should not be used as long as TC1
decoders continue to be supported.

---------------

The FCC rules specify that "a Backspace received when the cursor is in Column 1 shall be ignored," but it
does not specify how Backspace should be applied when it is received following a character displayed in
Column 32. Since the rules say, however, that "Backspace shall move the cursor one column to the left,
erasing the character or Mid-Row Code occupying that location," and since there is no Column 33, many
manufacturers have concluded that a Backspace received either before or after displaying a character in
Column 32 shall move the cursor to Column 31 and erase the character there. This application is legal under
the rules, and, although a different method might have been preferable, all decoders shall implement
Backspace in this manner. When erasing Column 31, the decoder may also erase any displayable
character or other code in Column 32 (as is currently done in TC2).

OK, I read that wrong, it's the people who send the captions that are supposed to monitor what's in column 32, not the ones who decode it, we're supposed to delete both columns 31 and 32. I need to rewrite that. I still can't figure out if PACs take up a space, or where I'm supposed to store their text style info... If I press process a backspace after a MidRow, do I cancel the style, or just remove the space and not remove the style unless there's another backspace? I don't know.

[Thunderbolt8]

it seems like quotation marks screw up the SHD line detection a bit:


- Cover him!|EAMES: Down! Down now! ---> - Cover him!|- Down! Down now!

working as intended


- "My father doesn't want me to be him."|EAMES: Exactly. ---> - "My father doesn't want me to be him."|Exactly

the hyphen indicating a change of speaker after the line break is missing

[izanami]

i dont see my font in waveline . pls help me how can i do
http://postimg.org/image/557fqwv9b/

[Nikse555]

@Thunderbolt8: thx, this case should be fixed in next update.

@izanami: could you test latest beta: http://www.nikse.dk/SubtitleEdit.zip ?
(the text in the waveform will appear double... hopefully one of them will be correct)

@ndjamena: Sorry, I've not had too much time to look at it... will have more time in about 14 days I think

[minhjirachi]

Originally Posted by Nikse555

@Thunderbolt8: thx, this case should be fixed in next update.

@izanami: could you test latest beta: http://www.nikse.dk/SubtitleEdit.zip ?
(the text in the waveform will appear double... hopefully one of them will be correct)

@ndjamena: Sorry, I've not had too much time to look at it... will have more time in about 14 days I think

I don't know why the bit depth of the exported bdn/xml files always 32-bit. Or I think that you should add the BDSup2Sub to Subtitle Edit, which process BDN/XML files really good.

[izanami]

Thank You Nikse555 . My font is ok in SE 3.4.6 . I really appreciate for your help

[ndjamena]

In case you find something useful in it here is the other document I was using:

http://www.mediafire.com/view/0j4whl...8.unlocked.pdf

(I've unlocked it.)

I suppose it's possible iTunes is buggy and CCExtractor time-codes are correct, but the muxing mode is Final Cut Pro, which is a program owned and written by apple... someone with a Mac could check how FCP handles the time-codes, if it agrees with iTunes then that's pretty much the end of the story. I couldn't figure out where CCExtractor was getting it's time-codes from but 608 captions are line 21 captions for NTSC analogue broadcasts, theoretically if you play the M4Vs back on an analogue TV you should be able to write the captions back into line 21, which isn't possible with the CCExtractor time-codes because you'd have to start sending the information for the first caption before you even begin playback of the file to get it to display at the right time, and all the other time-codes are off too.

CCExtractor is open source... Apparently. Beyond that VLC has a 608 decoder if you can find it.
https://wiki.videolan.org/VLC_Source_code/
http://ccextractor.sourceforge.net/a...extractor.html

My test program has been downloaded 3 times now... I probably should have deleted it but I guess it kind of works if all you want from it is SRTs and you don't use it on files with roll up captions... And unless I missed a program it does seem to be the only way to get the correct time-codes short of owning a MAC... well, this looks like it will do the job but it costs $900:

http://www.drastic.tv/index.php?opti...211&Itemid=303

There's probably something better out there but blow if I can find it.

That's about all I have to contribute.

[Dean007]

Hi. I have a question. Can I sync subtitles with a press of a button like Subtitle Workshop has (alt+m).
For example; I load subs and a video, mark from which subtitle I want to sync, play the video to that subtitle and simply press alt+m and subtitles that were marked are all sync with the video.