Implement AUDIO header

[TheNailDev]

Format 1.1.0 introduced the mandatory AUDIO header as a replacement for the MP3 header.

USDX now requires songs with VERSION >= 1.1.0 to include the AUDIO header and reads the audio file from this header.
The MP3 header is ignored for songs with VERSION >= 1.1.0.

When saving songs, the AUDIO and MP3 headers are written depending on the VERSION of the song, as shown in the table below:

VERSION	MP3	AUDIO
< 1.1.0	✔️	❌
>= 1.1.0 / < 2.0.0	✔️	✔️
>= 2.0.01	❌	✔️

Footnotes

1. VERSION >= 2.0.0 is currently not supported by USDX. This is just future proofing ↩

[barbeque-squared]

Done a quick initial pass. Is this the only change in 1.1.0 or is there more? The last thing I want to do is half-support something now that people are actually putting effort into standardizing it. Because people will hack around it and/or report bugs because we're only implementing it partially (and then comes the hotfix but it misses some edge case and now we're stuck with n+1 things for the next 20 years)

Not sure if I'll get to really dig into the spec until the weekend, hence this quick review.

EDIT: personally I'm not really against partial support, especially if it's an added tag like this, but if it also for example demands that #START and #END use sane values, then this is not getting in in its current state. It's a bit of a chicken-egg problem, especially with bigger spec changes it'll be good to split them up across multiple PR's, but ideally it's still all in a single release. Like I said, I need to actually research the spec soon.

[TheNailDev]

Is this the only change in 1.1.0 or is there more?

There are more changes for 1.1.0 (see #921):
New headers in 1.1.0 are AUDIO, VOCALS, INSTRUMENTAL, TAGS and PROVIDEDBY.
Additionally 1.1.0 also changes GENRE, CREATOR, EDITION, LANGUAGE and TAGS to multi-valued headers.

IMO support for format 1.1.0 can be classified in distinct sub-features which can be implemented independently:

• Implementation of AUDIO header (this PR)
• Implementation of VOCALS and INSTRUMENTAL (use add karaoke mode via #INSTRUMENTALS or #KARAOKE tag, key to toggle karaoke mode is K while playing. #850 as base)
• Implementation of the multi valued headers GENRE, CREATOR, EDITION, LANGUAGE and (new header) TAGS
• I don't see a reason to specifically implement PROVIDEDBY, as it is mostly just an information, from which song hosting service (USDB, spanish site, etc.) the song was downloaded. I don't think this information is relevant in USDX. (Naturally we could implement filtering by the song providers, but is that really something people would use?)

[TheNailDev]

I just changed TSong.Mp3 to TSong.Audio.
I did not notice any unexpected side effects.

I also changed the English and German localization to use Audio instead of MP3.

[barbeque-squared]

re: #951 (comment)

Yeah independent (or at least multiple) PRs is the way to go for that. But can we release this without all the other things? Personally I'm fine with partial support as long as it doesn't really break the not-yet implemented bits, but I'm not an end user.

Also generally agree with your comments on the subfeatures.

[TheNailDev]

But can we release this without all the other things?

In my opinion yes. As long as we don't say that we fully support 1.1.0 there should be no Issues.

For songs with Format < 1.1.0 nothing changes.

For songs with Format >= 1.1.0 the AUDIO and MP3 tag should (hopefully) both be valid audio files (usually the same file).
So for those songs also nothing would change. (Unless AUDIO is an invalid file, while MP3 is a valid file. But then I would say that the user should have the responsibility of having valid files and this should not be our issue)

[barbeque-squared]

Code looks good. Verified that existing songs still work. Did not actually test the new header, but the code looks good enough. Will merge tomorrow.