STV Import: Closed Caption Playback (v1.2.2 28/February/2008)

[JREkiwi]

Originally Posted by coryking

Show Analyzer seems to output SMI files that cannot be read by either this STV or the SageMC one.
4) You cannot see it above, but ShowAnalyzer saves it's files as either UTF-16 or UTF-8, which my copy of UltraEdit will pick up as unicode files. The mpg2srt ones are picked up as plain old ASCII files.

Apparently there is an option in version 0.8.8 of ShowAnalyzer and later to output .smi as ANSI. When I do a conversion of your file to ANSI it works. Also on the ShowAnalyzer forum, there is talk of captions not working properly with 0.8.11, but working well with 0.8.8

John

[Doubletop]

Yes that did it.

I used a clean copy of Sagetv3.xml and it loaded fine. Could it be something to do with the order of loading STVI's?

Anyway thats fixed it thanks; now I just need to find a confirmed source of CC data that's not just DVDs

[coryking]

This is the first version of ShowAnalyzer that I tried to get CC working. They must have taken out the ASCII format because all I'm getting is UTF-8 & UTF-16.

The ShowAnalyzer guys must have used MSDN for a reference, because they follow it perfectly. Given that Microsoft is the one who developed the file format, I'd wager it is actually mpg2srt that is broken.

The fileformat spec:
http://msdn2.microsoft.com/en-us/library/ms971327.aspx

Some samples & more info:
http://msdn2.microsoft.com/en-us/library/ms867679.aspx

That spec page has a download, samioverview.exe, that has some sample smi files.

Edit: Er, I see that you've gotten it to work if it is done as ASCII.... either way, besides the encoding, ShowAnalyzer nailed the spec :-)

[OferL]

Request:
Can you add an option to center justify the text or right justify, instead of left justify? or is there a way to change this in some .ini or .properties?

[JREkiwi]

Originally Posted by OferL

Can you add an option to center justify the text or right justify, instead of left justify?

I'll have that for you in the next update

John

[dblaine2]

So for those using this plugin with SA, did you switch SA outputting to srt to get it to work?

[JREkiwi]

Originally Posted by dblaine2

So for those using this plugin with SA, did you switch SA outputting to srt to get it to work?

srt won't work. I'll have a look at what's involved in getting the SA Unicode files working. No promises on that one though.

John

[JREkiwi]

Originally Posted by coryking

This is the first version of ShowAnalyzer that I tried to get CC working. They must have taken out the ASCII format because all I'm getting is UTF-8 & UTF-16.

OK, I've had a bit more of a look at this. The import will read UTF-8 files, but the ones from ShowAnalyzer you supplied use frame based timing (not the MS spec and means they run nearly 30x too fast) while the ones from mpg2srt and subtitle workshop use time based timing (work OK as ANSI or UTF-8)

John

[JREkiwi]

Just to add some more info from the SAMI(smi) spec.

"SAMI introduces the <Sync Start=> tag to support this functionality. The Start parameter is assigned the media's Elapsed Time value in milliseconds. "

John

[JREkiwi]

Originally Posted by OferL

Can you add an option to center justify the text or right justify, instead of left justify?

New version up for download

(1.0.8)
- Option to select Text Alignment of Captions (thanks Opus4)

John

[OferL]

Originally Posted by JREkiwi

New version up for download

(1.0.8)
- Option to select Text Alignment of Captions (thanks Opus4)

John

Thank you very much!

[Eckwell]

Will this import work if I am rendering using overlay? I thought I read somewhere that it must be VMR9, but I can't get smooth playback with VMR9.

thanks in advance

[JREkiwi]

Originally Posted by Eckwell

Will this import work if I am rendering using overlay?

Yes, the .smi file based captions will work in overlay.

John

[Eckwell]

I am using the OnAir GT USB Tuner to record HD via QAM. I know that the file that SageTV records does include the CC as I can take the file and watch it with the original OnAir Program and the CC appears. The CC is not analog, but digital.

The STV Import here is not rendering the embedded captions though. I think I might be able to extract the CC to a .smi file but none of the programs I have tried finds the captions.

Could someone here take a look at a sample mpg file and see if they can find the captions? So I can get them either rendered, or output to smi?

My wife has great difficulty watching TV without Captions and that is putting my whole SageTV based HTPC for HD recordings project on the line! Ah!!!

Thanks in advance.

Jon

[ke6guj]

In another thread, Eckwell let me know about a CC extractor that works on digital captures, http://ccextractor.sourceforge.net

The CC data is in my HDHR recordings and in my Firewired STB recordings.

I was able to extract .srt and .smi files with valid CC data in them. The files end in xxxxxxx_1.smi or xxxxxxxx_1.srt, so the files would need to be renamed, or perhaps the .stvi could be altered to look for the _1 file. Also, the formatting of the file is different than mpg2srt and subtitle workshop, so it would take some tweaking of the .stvi. Right now, I get the invalid captioning error.

If I take the .srt file and run it through srt2smi, then the resulting .smi will work. Also, there are some timing issues, but I dunno where the error is from.

edit: Timing issues seem to be with CC extraction from digital captures and not analog capture extractions. CCextractor files created from analog recordings in .srt and converted to .smi are in time, whereas digital recordings are off by seconds.

[JREkiwi]

On first glance I can see a problem with the case of SYNC start and the quotes on UNKNOWNCC, which should be easy to fix.

Is the timing out by a consistent amount? or is it variable?

John

[Eckwell]

I was watching "meet the Fockers" that I had recorded, ripped the CC using ccextractor, and converted the .srt to .smi (renaming the file) and the CC seems to be time aligned very well... until. About 30 minutes in there is a point were the CC is off by 10 seconds or so, immediately, as if there is a miscue in the file. I need to look further, but that seems to be the reason. I could back up 15 minutes into the show and it would realign, and them watch as it "fell" of out sync. It was a very specific point, not variable, or a consistent falling out of sync rate.

[ke6guj]

Did some quick tests for you...

View a show and watch the display, jump forward minutes and see if still in sync ( I can look at the .srt timestamps of the source file to double check to see if display is off, or the .smi file is off)

Andy Barker, QAM'ed HDHR, pretty much on time, maybe .25-.5 second
fast, perhaps some offset would fix that one. Sometimes gets out of sync (fast or slow) when FF'ing a lot, but a 10sec REW would get it back on sync..

Friday Night Lights, QAM HDHR, runs slightly fast, gaining seconds as it goes on. About 8 seconds fast in 30 minutes, 15 seconds in 1 hour.
edit: Tested another FNL, QAM, HDHR, perfect timing, only slightly out of sync on FF occasionally, a REW got it back on sync.

Brothers & Sisters, KABCDT, Firewired STB, some parts on time, some parts 1-5 seconds fast, and others 1-5 seconds slow.

Weeds, Showtime, FW STB, seems to be running about 10% fast, 1 minute fast over 10 minutes, 2 minutes over 20.

Tuders, ShowtimeHD, FW STB, similar 10% fast.

Entorage, HBO, FW STB, 55 seconds at 5, 90 at 10, 120 at 20.

[Eckwell]

I just checked another forum where the author of ccextractor had posted and he said he placed a .20 version on line that supports direct .smi production. You just need to use a command line parameter of -sami. I am rerunning an smi for Meet the Fockers and I am going to check the timing... that is if it is clean enough for SageTV to render it. It might be... we shall see.

[ke6guj]

I'm already using .20 and have already created some .smi files with it. The file syntax is different than the currently used .smi creators, and JREkiwi is looking into adding support into the .stvi.