Open WernerBleisteiner opened 3 years ago
I'm not sure whether there are really no features at all from BS.2076-2, but the ADM generated should be in general compliant to BS.2076-1. For ADM writing, the EBU ADM Renderer Python implementation is used and that one is currently only supporting Revision 1, see also https://github.com/ebu/ebu_adm_renderer/issues/18 . I don't why audioProgrammeNames from the EPS are not identified by FhG software but I'm also not too familiar with the changes in Revision 2. The argument, however, sounds not really convincing to me... I would be rather surprised if ADM generated by wav2bw64 would work then with the FhG tools. :-/
I'm also a bit sceptical. I'd consider that an ADM 'audioProgramme' maps to an MPEG-H 'preset'. Wrong idea? I'll check later and let you know. If you're interested yourself: These are the assets discussed: https://share.ard-zdf-box.de/index.php/s/W2pQBcZ7zegs8tF
Of course they recommend the use of MHAS 3.6 (which wasn't available when I embarked on the idea last autumn and - as we all know FH's ADM implementation was rather poor if not 'non-existabt' in earlier versions ;) ). Will check that, too. Yet, if that works, it might be a way to go at least when a DAW like Sequoia is involved (as long as there's no native support within)
More cross-checking made.
In brief: my 5 programme ADM test-file created from scratch (14 ch wav) with wav2bw64 is not accepted by FH ADMInfoTool and therefore also cannot be converted to MPF format (error log c.f. "FH-ADMInfo-Tool-ErrorLog-Compare-wav2bw64-MHAS-ADM-EPS-Prepare-Export-14ch_wav2bw64-ADM.txt")
Main issues being 14x "bit depth", "Invalid configuration of positionInteractionRange sub-elements...." plus warning. Didn't manage to track the issues yet.
However: Very interesting results when cross-checking various ADM creations from the same assets. I'll leave here the extracted xml from the various file variations. In order to identify the genealogy and relations of the files, have a look into the comment-data of the files. If that's of any further interest or you have questions, please do come back to me.
I suppose the main issue is that EPS and the respective ADM libraries need to be switched to BS.2072-2, wouldn't it? Referring to your issue and my recent question
Just to make sure: which ADM version does wav2BW64 support: BS.2076-1 or BS.2076-2? (I guess the latter). Found out that a multi-program (though channel-based only) ADM created with the present EPS does not forward the names given for audioProgrammes to MPEG-H tools (ADMInfo and EncMux - although the audio itself can get encoded). FH IIS states, that's due to the fact EPS only works on "ADM Revision 1" (BS.2076-1). Still have to check if wav2bw64 delivers accurate results.