[SAMPLE] Official CEA DTVCC Reference Test Suite for EIA-608 & CTA-708 captions

[bbgdzxng1]

[since Carlos is always looking for samples...]

Back in the day, the CEA DTVCC Working Group & WHD-TV put together a CTA-608/708 reference suite that could be used for testing captioning decoders, STBs & players, cycling through each main feature of 608 & 708. OEMs and vendors would use this reference when developing products.

The DVD used to be purchasable from the usual standards bodies for $500+ and eyewatering number for the amateur. Surprisingly, the media files never made their way out to archive.org, FFmpeg FATE test suite, MPV test suite or VLC test suite.

Well, it looks like the Official DTVCC Reference Suite is publicly available online from National Center for Accessible Media at WGBH (NCAM)'s FTP server, but was never indexed in any of the search engines.

• Kudos to NCAM for making the reference files available. Accessibility standards should be accessible.

Here is a link to the Official 608 Reference File (it is interlaced MPEG-2 Video in a Transport Stream, with the DTVCCs in the user data headers)...
http://ncamftp.wgbh.org/DTV/CEA%20test%20material/Iteration_1/CEAv1.2zero.trp
And here is the PDF support sheet indicating the timings of each test case...
http://ncamftp.wgbh.org/DTV/CEA%20test%20material/Iteration_1/It1.2ReadMe.pdf

Iteration_1 is for 608 captions in DTVCC compatibility bytes. Iteration_1 is particularly good for testing 608 extended characters, rollup behavior, overwriting, deleting, tab-offsets, italics and underlines.

Iterations 2-4 are for true 708s.

├── Iteration_1
│   ├── CEAv1.2zero.trp
│   └── It1.2ReadMe.pdf
├── Iteration_2
│   ├── It2A ReadMe V1.pdf
│   ├── IT2Afinal_110504.trp
│   ├── It2B ReadMe V1.pdf
│   └── IT2BFinal_v2_110504.trp
├── Iteration_3
│   ├── It3 ReadMe V1.pdf
│   └── It3final110504.trp
├── Iteration_4
│   ├── Hex data
│   │   ├── IT4_A.hex
│   │   ├── IT4_B.hex
│   │   ├── IT4_C.hex
│   │   └── IT4_D.hex
│   ├── IT4ReadMe V1.pdf
│   ├── Transport streams
│   │   ├── IT4_A.trp
│   │   ├── IT4_B.trp
│   │   ├── IT4_C.trp
│   │   └── IT4_D.trp
│   └── XML source
│       ├── IT4_A.xml
│       ├── IT4_B.xml
│       ├── IT4_C.xml
│       └── IT4_D.xml

For anyone involved in caption decoding of 608 & 708, this is an excellent reference for testing captions. Linking here so that Carlos & crew can download and secure themselves a backup of the Official DTVCC Test Suite in case NCAM ever goes offline.

With a bit of FFmpeg, it is possible to transcode these MPEG-2 video files into H.264 w/SEI side data and/or Quicktime MOV separate 608-track, and with ccextractor, it should be possible to extract a perfect SCC file representing the 608 byte-pairs.

[I am not sharing any of the files, I am sharing a link to the National Center for Accessible Media's public FTP server.]

Carlos & Co - feel free to close the issue once you have the suite - in case you do not already have a copy.

[bbgdzxng1]

Closing ticket since public link is now memorialized in a ticket. If anyone needs this official reference file, link to NCAM's hosted version is above.

[cfsmp3]

@bbgdzxng1 Wait wait :-) I saw it (thanks!) but didn't download it - @canihavesomecoffee can you add this to our official corpus please?

[bbgdzxng1]

@cfsmp3 - don't panic. Link to the main Iteration_1 608 file is active.

These test references were produced with professional tools in about 2003 during the development of 708. There's a couple of very minor errors where one or two of the tuplets in the 608 file is a little wonkey, but other than that, they are really good in that they cycle through all the main 608 features (including complex rollup cases, deletes and overwrites). The 608 track has second data channels (cc2) and the transport stream has second data fields (cc3/4).

MPV / FFmpeg starts to get confused with the more complex 608 tests (because FFmpeg's ccaption_dec.c needs a lot of love), so don't use that as a reference player. If you transcode to HLS, Quicktime Player does a really good job of showing what each test should like on screen. Quicktime is the better of the player (but still mis-converts apostrophes).

I don't believe these media contain any (c)Copyright notice, nor is there any statement on sharing the media, but you may prefer to keep your own copies in a private corpus.

Hope you are well - and thanks for all your work on recent versions of ccextractor.

Enjoy.

[cfsmp3]

@cfsmp3 - don't panic. Link to the main Iteration_1 608 file is active.

These test references were produced with professional tools in about 2003 during the development of 708. There's a couple of very minor errors where one or two of the tuplets in the 608 file is a little wonkey, but other than that, they are really good in that they cycle through all the main 608 features (including complex rollup cases, deletes and overwrites). The 608 track has second data channels (cc2) and the transport stream has second data fields (cc3/4).

MPV / FFmpeg starts to get confused with the more complex 608 tests (because FFmpeg's ccaption_dec.c needs a lot of love), so don't use that as a reference player. If you transcode to HLS, Quicktime Player does a really good job of showing what each test should like on screen. Quicktime is the better of the player (but still mis-converts apostrophes).

I don't believe these media contain any (c)Copyright notice, nor is there any statement on sharing the media, but you may prefer to keep your own copies in a private corpus.

Hope you are well - and thanks for all your work on recent versions of ccextractor.

Enjoy.

I could only download a few files:

• Iteration_1/CEAv1.2zero.trp (650MB)
• Iteration_1/It1.2ReadMe.pdf (125KB)
• Iteration_2/IT2Afinal_110504.trp (428MB)
• Iteration_2/It2A ReadMe V1.pdf (341KB)

The rest 404.
Still, super useful. Thanks!