unclear differences between the various AAC test streams

[jpiesing]

I'm working on the HbbTV .xml metadata files about our audio-only tests and I'm unclear about the differences between some of the AAC test streams. The master index of test content on https://cta-wave.github.io/Test-Content/ doesn't document any differences between some of them.

If we consider at1 to be the baseline then it looks to me like we have the following;

• at2; Chunks per Fragment = 4 iso 1
• at3; trun v1 instead of trun v0
• at4; without edit lists instead of with edit lists
• at5; No documented difference from at1
• at13; encoded from PN03 iso PN01
• at14; encoded from PN04 iso PN01
• at15; No documented difference from at13
• at16; No documented difference from at14
• at6; he-aac instead of aac-lc
• at7; he-aac-v2 instead of aac-lc

@yanj-github Is it possible to extend https://cta-wave.github.io/Test-Content/ to explain what these differences are?

[rcottingham]

Hi Jon,

Sparse Matrix is here: #58 (comment) - let me know if there is a better place to store it.
Yellow highlighted cells show the different property from the base stream at1.

[jpiesing]

Hi Jon,

Sparse Matrix is here: #58 (comment) - let me know if there is a better place to store it. Yellow highlighted cells show the different property from the base stream at1.

My comment was more that the public registry at https://cta-wave.github.io/Test-Content/ was missing some information.

Looking at the sparse matrix, the 3 streams mentioned above are the ones that are encrypted.
Do you know / remember who added those streams to the github.io page?

[rcottingham]

Hi @rbouqueau @nlsdvl - to address Jon's query we need to add the encryption status and track duration to database.json. Can you advise how best to do this so it appears in the public registry e.g. maybe add it to the representations like this?:

"caac_sets/aac_lc/at5/2023-04-27/":` {
            "representations": {
                "bitrate": "128000",
                "samplerate": "48000",
                "input": "PN01.wav",
       -->      "encryption": "Clearkey",
       -->      "duration":"30"
            },
            "channel": "Stereo",
            "segmentDuration": "1.984",
            "chunksPerFragment": "1",
            "trunVersion": "v0",
            "editList": "With",
            "mpdPath": "https://dash-large-files.akamaized.net/WAVE/vectors/releases/1/caac_sets/aac_lc/at5/2023-04-27/stream.mpd",
            "zipPath": "https://dash-large-files.akamaized.net/WAVE/vectors/releases/1/caac_sets/aac_lc/at5/2023-04-27/at5.zip"
        },

[nlsdvl]

Hi @rcottingham for some reason the database has been designed with a root key "CENC".

With more encryption options coming up on test vectors, it may indeed be a good time to update the way the information is stored : cta-wave/Test-Content-Generation#74

Note that changing the database implies updating the webpage's code as well.

With that said, I would store both duration and DRM information on the test vector object rather than on the 'representation' object.

[jpiesing]

If we want to be complete, we should also add the codec string from the MPD, i.e. including the profile / level as I think some of the audio streams only differ by this.

mp4a.40.2
mp4a.40.5
mp4a.40.29

[rbouqueau]

Let me summarize:

• Encryption: add info on the test vector object (in "representations"?) instead of using the "CENC" root key.
• @rcottingham also suggests to add the duration.
• @jpiesing suggest to add the rfc6381 string (in "representations"?).

Do that that correct to everyone?

As Nils wrote, this means:

• Changing the generation scripts.
• Updating the database (manually sometimes).
• Updating the front-end.

[jpiesing]

Let me summarize:

* Encryption: add info on the test vector object (in "representations"?) instead of using the "CENC" root key.

* [@rcottingham](https://github.com/rcottingham) also suggests to add the duration.

* [@jpiesing](https://github.com/jpiesing) suggest to add the rfc6381 string (in "representations"?).

Do that that correct to everyone?

As Nils wrote, this means:

* Changing the generation scripts.

* Updating the database (manually sometimes).

* Updating the front-end.

Is a set of changes possible that are less work? Is this hours? days? more?

[rbouqueau]

I'd say a couple of days. @nlsdvl Do you agree?

[haudiobe]

DPCTF June 18, 2025:

We appreciate @rbouqueau suggestions.

• Changing the generation scripts

• Updating the database (manually sometimes).

• Updating the front-end.

We need to identify the minimum we need to do and see who can execute. @yanj-github