Do Lossless Files From Tape Capture Carry Caption

[dellsam34]

I'm trying to figure out if caption is carried over to the digital captured files, I use the Snell & Wilcox TBS800 capture device that gives the option to output the caption, however I'm not sure what type of output is mentioned in the manual, Is it SDI or analog output only? So how do I check if the captured file contains digital caption? Do lossless files carry any metadata at all? if so how do I pass that over to encoded files or uploading to a platform like YouTube?

The TBS800 does detect the caption and display the scan line containing it, If I stop the tape or play across commercials the caption turns into "None", As you can see from the screenshots I enabled all type of captions and teletext but as I said I'm not entirely sure if SDI is considered as an output as mentioned in the manual, I'm attaching a lossless sample captured from a caption section on the tape as well as the user manual here if anyone can interpret it better than me.

[Sharc]

You can see the caption on the top few lines of the .avi. The frame is 486 hight instead of the usual 480 for NTSC.

[pandy]

As pointed by Sharc - captions are located on area normally not active video (such as VBI lines) - to keep such signals your capture size must correspond to not only active video lines but also those normally not used to carry video signal. If you can capture video from line for example 6 and store such samples as video signal then you have RAW coded data lines alternatively your capture device may extract such data from invisible lines and save them as associated with capture data metadata or as separate file.

Being on your side (seem you are seriously dedicated to capture) and if you have space then i would search for TEKTRONIX VM700 - this is state of the art video signal analyzer so you can not only monitor your setup but also perform real video measurements (and quickly address your questions).
This is probably overkill but still something you may consider seriously - i worked a lot with VM700 and from my perspective this machine cover all you need and much more on regular SD signal. Nowadays it is not so expensive as i see on ebay.

[dellsam34]

You can see the caption on the top few lines of the .avi. The frame is 486 hight instead of the usual 480 for NTSC.

That's not what I'm asking, My question was, does the caption data get transferred from the analog scan lines to digital metadata like the ones in Blu-ray or mkv files?

[dellsam34]

I think I found a similar topic on this forum: https://forum.videohelp.com/threads/362776-Question-how-to-extract-closed-captions-fro...tured-TV-video

[pandy]

That's not what I'm asking, My question was, does the caption data get transferred from the analog scan lines to digital metadata like the ones in Blu-ray or mkv files?

Then answer is no - your SDI can pass VBI from input to output - in some cases you can select line where such data are reinserted - to capture those data you need capture VBI lines and later use tool that can extract data from signal and offer it as for example txt file.
Probably ffmpeg offer such parser for close captions also there is dedicated group of utilities https://github.com/zapping-vbi/zvbi

[dellsam34]

Oh okay, So the TBS800 does not have the capability to extract and embed the data into the file's header, This means that it can only create caption or teletext lines from SDI into analog out with special caption or teletext consoles like this one, Although in that video he is using the analog input/output version, not the digital input/analog output version.

But the question still remains, what if the TBS800 is indeed outputting caption and teletext data over SDI and the ingest software is ignoring it, What if there is a software out there like some pro NLE that can detect that data in the SDI stream, I know for a fact that SDI is not just video and audio, it include other data like timecode, signaling, HDCP ... etc. so caption and teletext and other proprietary messages and barcodes should be part of that data.

[pandy]

Oh okay, So the TBS800 does not have the capability to extract and embed the data into the file's header, This means that it can only create caption or teletext lines from SDI into analog out with special caption or teletext consoles like this one, Although in that video he is using the analog input/output version, not the digital input/analog output version.

This is how i understand provided by you manual for TBS800 - there is no detailed description there except rather brief information.
This is also inline with SDI functionality - just digital video interface.

But the question still remains, what if the TBS800 is indeed outputting caption and teletext data over SDI and the ingest software is ignoring it, What if there is a software out there like some pro NLE that can detect that data in the SDI stream, I know for a fact that SDI is not just video and audio, it include other data like timecode, signaling, HDCP ... etc. so caption and teletext and other proprietary messages and barcodes should be part of that data.

Normally such data are located in area not used to transport video so obviously device with SDI input need to have some form of data extractor - if your capture device is able to capture not only active video lines but also those VBI's then extraction should be not difficult - they will be accessible as RAW video and can be sliced and extracted.

[dellsam34]

The digital video capture active portion is only 486 (576 PAL/SECAM) unfortunately, there is no other way of outputting more scan lines, The device does recognize and read any kind of data as shown in the menu from the incoming signal from the VCR and can probably output it to a TV connected to its analog output (I say probably because I haven't tried it yet) but I know it outputs 525/625 lines out of an analog output because that's how analog video works, so the caption should be there.
There is no question that I can decode the caption from the device analog output or directly from the VCR itself unused composite output, But this is a different process, my inquiry was to see if the TBS800 has an automatic way of doing this for me and put it in the header of a digital file, which doesn't look like it's happening.

Mediainfo doesn't give too much information about the video, here they are:

General
Complete name                            : D:\Capture0000.mov
Format                                   : QuickTime
Format/Info                              : Original Apple specifications
File size                                : 784 MiB
Duration                                 : 27 s 160 ms
Overall bit rate                         : 242 Mb/s
Frame rate                               : 29.970 FPS
Encoded date                             : 2024-01-07 04:37:37 UTC
Tagged date                              : 2024-01-07 04:37:37 UTC
Writing library                          : Apple QuickTime

Video
ID                                       : 1
Format                                   : YUV
Codec ID                                 : v210
Codec ID/Hint                            : AJA Video Systems Xena
Duration                                 : 27 s 160 ms
Bit rate mode                            : Constant
Bit rate                                 : 224 Mb/s
Width                                    : 720 pixels
Clean aperture width                     : 704 pixels
Height                                   : 486 pixels
Clean aperture height                    : 480 pixels
Display aspect ratio                     : 4:3
Clean aperture display aspect ratio      : 4:3
Frame rate mode                          : Constant
Frame rate                               : 29.970 (30000/1001) FPS
Standard                                 : NTSC
Color space                              : YUV
Chroma subsampling                       : 4:2:2
Bit depth                                : 10 bits
Scan type                                : Interlaced
Scan type, store method                  : Interleaved fields
Scan order                               : Bottom Field First
Compression mode                         : Lossless
Bits/(Pixel*Frame)                       : 21.333
Stream size                              : 724 MiB (92%)
Language                                 : English
Encoded date                             : 2024-01-07 04:37:37 UTC
Tagged date                              : 2024-01-07 04:37:37 UTC
Color primaries                          : BT.601 NTSC
Transfer characteristics                 : BT.709
Matrix coefficients                      : BT.601

Audio
ID                                       : 2
Format                                   : PCM
Format settings                          : Little / Signed
Codec ID                                 : lpcm
Duration                                 : 27 s 160 ms
Bit rate mode                            : Constant
Bit rate                                 : 18.4 Mb/s
Channel(s)                               : 16 channels
Sampling rate                            : 48.0 kHz
Bit depth                                : 24 bits
Stream size                              : 59.7 MiB (8%)
Language                                 : English
Encoded date                             : 2024-01-07 04:37:37 UTC
Tagged date                              : 2024-01-07 04:37:37 UTC

Other
ID                                       : 3
Type                                     : Time code
Format                                   : QuickTime TC
Duration                                 : 27 s 160 ms
Frame rate                               : 29.970 (30000/1001) FPS
Time code of first frame                 : 00:00:00;00
Time code of last frame                  : 00:00:27;03
Time code, stripped                      : Yes
Language                                 : English
Encoded date                             : 2024-01-07 04:37:37 UTC
Tagged date                              : 2024-01-07 04:37:37 UTC

[pcspeak]

Using Mediainfo - Your sample, which I downloaded, does not show ID : 3 (QuickTime TC)
Is this correct?


You uploaded an avi. Your mediainfo output shows a mov (apple) file

[pandy]

The digital video capture active portion is only 486 (576 PAL/SECAM) unfortunately, there is no other way of outputting more scan lines, The device does recognize and read any kind of data as shown in the menu from the incoming signal from the VCR and can probably output it to a TV connected to its analog output (I say probably because I haven't tried it yet) but I know it outputs 525/625 lines out of an analog output because that's how analog video works, so the caption should be there.
There is no question that I can decode the caption from the device analog output or directly from the VCR itself unused composite output, But this is a different process, my inquiry was to see if the TBS800 has an automatic way of doing this for me and put it in the header of a digital file, which doesn't look like it's happening.

Then seem you are out of luck i must wrote sadly - from your perspective probably the best (archive purpose) will be to capture also VBI lines (from 10 to 20 accordingly to SMPTE 170 and perhaps to 21 also from 273 to 283) as they can carry some video associated data...

[dellsam34]

Using Mediainfo - Your sample, which I downloaded, does not show ID : 3 (QuickTime TC)
Is this correct?


You uploaded an avi. Your mediainfo output shows a mov (apple) file

I did a quick lossless compression to reduce the size of the sample to fit into the 500MB limit, The media info is for the raw capture from SDI put into lossless Quicktime by the HyperDeck, it's 10bit, so the files are huge.

[dellsam34]

Then seem you are out of luck i must wrote sadly - from your perspective probably the best (archive purpose) will be to capture also VBI lines (from 10 to 20 accordingly to SMPTE 170 and perhaps to 21 also from 273 to 283) as they can carry some video associated data...

It was just curiosity, not something necessary, Although it would have been nice if it was possible to pass that data into the file header and can be picked up by YouTube.

[pcspeak]

You uploaded an avi. Your mediainfo output shows a mov (apple) file

I did a quick lossless compression to reduce the size of the sample to fit into the 500MB limit, The media info is for the raw capture from SDI put into lossless Quicktime by the HyperDeck, it's 10bit, so the files are huge.

Virtualdub dropped all but the video and audio.
Maybe do it with ffmpeg. (5 minutes) See if Mediainfo shows the time codes as being copied into the output.
get-test-video.cmd

for %%a in ("*.mp4", "*.avi", "*.mkv", "*.mpg", "*.mov") do ffmpeg.exe -i "%%a" -ss 00:0:00 -to 00:5:00 -c copy -y "%%~na-test%%~xa"
pause

[dellsam34]

I don't think vdub dropped anything, I will make a short capture later of few seconds and upload it here in its raw form to prevent any manipulation, But I'm sure it will not have any data other than audio and video.

[pcspeak]

I don't think vdub dropped anything, I will make a short capture later of few seconds and upload it here in its raw form to prevent any manipulation, But I'm sure it will not have any data other than audio and video.

Thanks for that. Your upload of the Mediainfo results from D:\Capture0000.mov shows QuickTime time codes. I'm getting a little confused.

[dellsam34]

Blank timecode, blank 14 audio tracks, that's how the the HyperDeck works, There is no timecode in VHS captures.

[dellsam34]

Here is a raw sample.

[pcspeak]

Deleted

[Selur]

" Capture0006.mov " => That file does not include any caption.

[dellsam34]

Yes, that's what I was trying to demonstrate to the member, that vdub does not strip any metadata.

[Selur]

That only shows that the source did not have cations. It does not how whether VirtulDub strips metadata or not.

[aramkolt]

Pretty interesting topic. In that sample you posted, the flickering white lines at the very top of the frame do look like closed captioning data to me. Since this was recorded off air (noting the TBS logo) seems like it would be on those lines. It's interesting that the metadata somehow knows that both the video and audio language is English if there's really no data there?

While not the topic of this thread, am I seeing motion blur with the slight head movements? Not really what I'd have expected with a lossless encode. Granted, this is a really low movement clip so probably hard to say...

[dellsam34]

That only shows that the source did not have cations. It does not how whether VirtulDub strips metadata or not.

Yes, that's the point I was trying to make.

[dellsam34]

Pretty interesting topic. In that sample you posted, the flickering white lines at the very top of the frame do look like closed captioning data to me. Since this was recorded off air (noting the TBS logo)

It is indeed the caption lines, But in this experiment I was not interested in getting them out of the image myself, plenty of other scripts that can capture the full 525 lines and do that efficiently.