Test Content for Splicing Tests

[louaybassbouss]

Test Content for following tests is missing:

• 8.8 Playback over WAVE Baseline Splice Constraints
• 8.13 Restricted Splicing of Encrypted Content
• 8.14 Sequential Playback of Encrypted and Non-Encrypted Baseline Content
• 9.4 Splicing of WAVE Program with Baseline Constraints

[jpiesing]

Test Content for following tests is missing:

* 8.8 Playback over WAVE Baseline Splice Constraints

* 8.13 Restricted Splicing of Encrypted Content

* 8.14 Sequential Playback of Encrypted and Non-Encrypted Baseline Content

* 9.4 Splicing of WAVE Program with Baseline Constraints

Which mezzanine content does this correspond to?

[FritzHeiden]

I moved the already generated tests for "8.8 Playback over WAVE Baseline Splice Constraints" and "8.14 Sequential Playback of Encrypted and Non-Encrypted Baseline Content" into the splicing-demo directory to avoid confusion. These tests exist for the purpose of demonstrating the functionality of the playout parameter in a splicing test scenario. The used content is not intended for these tests

[rbouqueau]

I'd like to make a status here, so that we have it written somewhere.

The current Test-Content-Generation implementation takes an A/V content and processes it in one step (single call to encode_dash.py from run-all.py - these are the two only scripts involved here). This steps includes the encoding, the packaging and the annotations. The archiving and the encryption are done "over-the-top" with direct calls from run-all.py.

Processing in one step causes two kinds of issues: 1) the content is bundled in folders (see cta-wave/dpctf-tests#59) and 2) the composition of content (#19, #4, audio profiles only, ...) is also more complex than expected. The good news is that gpac allows to split them ; my task will be to split the code appropriately to implement this.

Once this is done, generating the splicing content is just a matter of assembling the annotations. We don't need to generate some new CMAF content again. This part of the process needs to be done manually at the moment ; let's see how and how much it can be automatized.

[rbouqueau]

Some first splice content:

• http://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_clear_9.4/stream.mpd
• http://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_clear_cenc_8.14/stream.mpd
• http://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_cenc_8.13/stream.mpd

Any feedback welcome. Only the first stream (in clear) is playable. The manifests are very similar (some periods are in clear while others are encrypted). If it looks ok then I will update it to the Akamai server.

[rbouqueau]

Ping @louaybassbouss :)

[jpiesing]

@louaybassbouss Ping ...

[louaybassbouss]

@FritzHeiden FYI new splicing test content is now available.

[louaybassbouss]

@rbouqueau the content is avi

Some first splice content:

• http://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_clear_9.4/stream.mpd
• http://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_clear_cenc_8.14/stream.mpd
• http://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_cenc_8.13/stream.mpd

Any feedback welcome. Only the first stream (in clear) is playable. The manifests are very similar (some periods are in clear while others are encrypted). If it looks ok then I will update it to the Akamai server.

@rbouqueau can you provide each content in multiple MPDs (2 MPDs one for main and one for spliced) instead of multi-periods ? see issue cta-wave/dpctf-tests#22 .

We check the content manually the clear content is playing properly. Regarding cenc content, we were able to play it even if we don't pass the cenc Key ID and key value. Can you check it please.

[rbouqueau]

@louaybassbouss Thanks for the feedback.

Can you provide each content in multiple MPDs (2 MPDs one for main and one for spliced) instead of multi-periods ? see issue cta-wave/dpctf-tests#22 .

Here are all the links. I'd be happy if we could clarify this (separate manifests vs multi-period manifest):

• Main content, clear: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_main/2021-01-18/stream.mpd
• Main content, CENC: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_main-cenc/2021-01-18/stream.mpd
• Ad content, clear: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_ad/2021-01-18/stream.mpd
• Ad content, CENC: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_ad-cenc/2021-01-18/stream.mpd

Regarding cenc content, we were able to play it even if we don't pass the cenc Key ID and key value. Can you check it please.

See links above. I checked and the content is encrypted. The content still use the same keys, could that be why you can play it back?

[louaybassbouss]

@louaybassbouss Thanks for the feedback.

Can you provide each content in multiple MPDs (2 MPDs one for main and one for spliced) instead of multi-periods ? see issue cta-wave/dpctf-tests#22 .

Here are all the links. I'd be happy if we could clarify this (separate manifests vs multi-period manifest):

The final decision was in cta-wave/dpctf-tests#22 to use multiple MPDs. There is no multi-period at all.

• Main content, clear: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_main/2021-01-18/stream.mpd
• Main content, CENC: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_main-cenc/2021-01-18/stream.mpd
• Ad content, clear: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_ad/2021-01-18/stream.mpd
• Ad content, CENC: https://reference.dashif.org/dash.js/latest/samples/dash-if-reference-player/index.html?mpd=https://www.gpac-licensing.com/downloads/CTA-Wave/splice_ad-cenc/2021-01-18/stream.mpd

Thanks we will check and let you know

Regarding cenc content, we were able to play it even if we don't pass the cenc Key ID and key value. Can you check it please.

See links above. I checked and the content is encrypted. The content still use the same keys, could that be why you can play it back?

Thanks for the hint this can be the reason

[rbouqueau]

@louaybassbouss Note that the URLs (server) will evolve once this is validated on your side.

[louaybassbouss]

@rbouqueau the streams are working fine. Can you make them available via database.json? @jpiesing regarding content durations 10s for main content and 5s for Ad content are fine from your side? From Test implementation perspective the duration doesn't matter.

[rbouqueau]

Agreement during the call to create a new "CFHD-splicing" database + front-end tab.

[louaybassbouss]

@rbouqueau when do you think can you make the new splicing content available in database.json?

[rbouqueau]

Added by 5ec1ca4. Let me know if this is ok for you.

[louaybassbouss]

Thanks @rbouqueau . There is still a mismatch of the ID structure for few elements most are related ss1 streams (Example https://github.com/cta-wave/Test-Content/blob/master/database.json#L472) can you make all IDs Zip/MPD URLs following the same structure?

[rbouqueau]

Thanks for reporting. That's a bug in the script. The ss1 is handled separately. I should have a look this evening.

[rbouqueau]

That's fixed, thanks again. The script is safe, it was an error when merging the database.

[louaybassbouss]

@FritzHeiden can we close this issue?

[FritzHeiden]

@FritzHeiden can we close this issue?

Yes. For now we are using the provided splice_main and splice_ad content

[jpiesing]

@FritzHeiden @rbouqueau I would like to update the Google spreadsheet to include splice_main and splice_ad and ensure all our test streams are included. @Romain Do you have the information on their properties? Do we need to add any more rows to the table to capture important details?

[rbouqueau]

I arbitrarily copied the properties from T1.

[jpiesing]

I arbitrarily copied the properties from T1.

Just now or when you made those streams?
Do splice_main and splice_ad_content need to be regenerated with the most recent changes?

[rbouqueau]

Just now or when you made those streams?

When I made the streams. I had asked during a Tuesday call what to do and the answer was "whatever is conformant".

Do splice_main and splice_ad_content need to be regenerated with the most recent changes?

I don't think so.

[gitwjr]

@FritzHeiden @rbouqueau need to discuss next steps/status to resolve this issue.

[rbouqueau]

NB: while digging I found randomly that the origin of this request is here: cta-wave/dpctf-tests#101.

[rbouqueau]

Today's notes: @rbouqueau evaluate the situation tomorrow after JCCP call + @nicholas-fr to check with fractional frame rate. If ok, then generate asap.

[rbouqueau]

See my comment in cta-wave/Test-Content-Generation#55

[rbouqueau]

Content updated.

[gitwjr]

@FritzHeiden @louaybassbouss Fritz and Louay to check and then close if OK.

[rbouqueau]

@FritzHeiden @louaybassbouss Fritz and Louay to check and then close if OK.

Thx!

[gitwjr]

Closed. This was checked during the HbbTV plugfest.