HEVC switching sets referencing invalid segment URL

[FritzHeiden]

Running any switching set test with HEVC content fails with error:

Failed to load resource: https://dash.akamaized.net/WAVE/vectors/switching_sets/12.5_25_50/ss2_chh1/2025-01-15/1/0.m4s the server responeded with a status of 404 ()

same with local content:

Failed to load resource: /content/switching_sets/12.5_25_50/ss2_chh1/2025-01-15/1/0.m4s the server responeded with a status of 404 ()

It appears the URL for the referenced segments is wrong

[rbouqueau]

Thanks for reporting.

I confirm the pattern for the filenaming in the manifest seems incorrect, the $RepresentationId$ seems missing:

    <AdaptationSet segmentAlignment="true" maxWidth="1920" maxHeight="1080" maxFrameRate="50" par="16:9" mimeType="video/mp4" startWithSAP="1" contentType="video" containerProfiles="cmf2 chh1">
      <SegmentTemplate media="1/$Time$.m4s" initialization="1/init.mp4" timescale="12800">
        <SegmentTimeline>
          <S t="0" d="25600" r="29"/>
        </SegmentTimeline>
      </SegmentTemplate>
      <Representation id="../../../../chh1_sets/12.5_25_50/t12/2025-01-15/1" codecs="hvc1.2.4.L123.90" width="1280" height="720" frameRate="25" sar="1:1" bandwidth="3100000"/>
      <Representation id="../../../../chh1_sets/12.5_25_50/t11/2025-01-15/1" codecs="hvc1.2.4.L123.90" width="1920" height="1080" frameRate="25" sar="1:1" bandwidth="5000000"/>
      <Representation id="../../../../chh1_sets/12.5_25_50/t10/2025-01-15/1" codecs="hvc1.2.4.L123.90" width="1920" height="1080" frameRate="50" sar="1:1" bandwidth="8000000"/>
    </AdaptationSet>

CC @nlsdvl

[nlsdvl]

Indeed.

[FritzHeiden]

I can confirm the hosted stream MPDs are now fixed, however, when downloading the content it still has the old version.

[nlsdvl]

I can confirm the hosted stream MPDs are now fixed, however, when downloading the content it still has the old version.

@FritzHeiden the mpd was fixed and updated, but as far as I understand, the representations didn't need to be changes and so weren't re-uploaded.

[FritzHeiden]

We are running into this issue again. The MPD hosted on akamai works fine, however, the MPD contained in the zip file still is wrong. Running test with local content still fails with the mentioned error message.
For now, we run the tests with the online content, however, this may skew test results.

[rbouqueau]

@FritzHeiden : I see some issue, let me make sure I understand and propose a PR.

[rbouqueau]

Ok, can you retry please? (and re-close if ok)

[FritzHeiden]

The directory structure inside the zips appears to be off. For example in this zip:

https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/12.5_25_50/ss5_chd1/2025-09-23/ss5_chd1.zip

I expect the structure to be:

/switching_sets/12.5_25_50/ss5_chd1/2025-09-23

But it is actually:

/ss5_chd1/switching_sets/12.5_25_50/ss5_chd1/2025-09-23

This appears to be the case for the following zip files:

https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/12.5_25_50/ss2_chh1/2025-09-23/ss2_chh1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/14.985_29.97_59.94/ss2_chh1/2025-09-23/ss2_chh1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/15_30_60/ss2_chh1/2025-09-23/ss2_chh1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/12.5_25_50/ss3_cud1/2025-09-23/ss3_cud1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/14.985_29.97_59.94/ss3_cud1/2025-09-23/ss3_cud1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/15_30_60/ss3_cud1/2025-09-23/ss3_cud1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/12.5_25_50/ss4_clg1/2025-09-23/ss4_clg1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/14.985_29.97_59.94/ss4_clg1/2025-09-23/ss4_clg1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/15_30_60/ss4_clg1/2025-09-23/ss4_clg1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/12.5_25_50/ss5_chd1/2025-09-23/ss5_chd1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/14.985_29.97_59.94/ss5_chd1/2025-09-23/ss5_chd1.zip
https://dash-large-files.akamaized.net/WAVE/vectors/switching_sets/15_30_60/ss5_chd1/2025-09-23/ss5_chd1.zip

[nlsdvl]

@FritzHeiden I followed the same pattern for the hevc switching sets archives that I observed had been previously used for avc content's switching set (ss1).

Note, the archive contains not only the switching_sets, but also a copy of all test vectors it depends upon:

ss5_chd1/switching_sets
ss5_chd1/switching_sets/12.5_25_50/ss5_chd1/2025-09-23/
ss5_chd1/chd1_sets
ss5_chd1/chd1_sets/12.5_25_50/t70/2025-01-15
ss5_chd1/chd1_sets/12.5_25_50/t71/2025-01-15
ss5_chd1/chd1_sets/12.5_25_50/t72/2025-01-15
ss5_chd1/chd1_sets/12.5_25_50/t73/2025-01-15