Podcast support

[Shanachie]

Earlier today I was thinking of adding some support for podcasts to Rockbox and here's what I came up with:

I would like to extend the bookmarks-concept of Rockbox to suit podcasts. Most podcasts, at least the once I listen to, contain a few topics. It would be wonderfull if you could jump past a topic or song that doesn't interest you to the next part of the show.

So what we need is timecodes, which could be added to the rss-feed by the podcasters,  parsed by the feed-aggregator and converted to some format rockbox would understand. Podcasters could do this easy during recording or editing theire shows. I could even be done by podcast-production software.

On the Rockbox side it would work like this:
There's a top-level podcast folder which has special meaning to Rockbox.
Every podcast-file is added to a directory with the name of the podcast, this could be a dotted-name (thus hidden) because we wont need to see the files themselves.
In the podcast-directory files are placed that contain info about a podcast:

Code: [Select]

podcast title | linenr. of next item
|file (relative path)
|progress in the file (same as a bookmark entry)
|timecodes (parsed from timecodese entry in rss)
|| discription | timecode (again, same as bookmark entry)
Podcast title | next item
|
|

When playing a podcast the next, and previous actions should be changed to next- and previous-timecode. So skiping that boring song in a podcast would become easy.

Please tell me what you think while I figure out how to implement this.

Greets Shanachie

[bk]

As far as I know, the podcasts that support "enhanced content" (as opposed to straight audio) like bookmarks, topic separation, website links, etc, are in "enhanced" AAC format. This requires the Rockbox AAC codec to be optimized beyond realtime (which it isn't yet) and also support this (proprietary?) metadata format.

So it looks like Apple has already thought of this.

[Shanachie]

Hey bk, thanks for replying.
I know of enhanced AAC's but I'm thinking of support for any audio (and video) format used in podcasts. The idea is to have the timecode info in the podcast feed, that way a podcast-fetcher can do the conversion to the format that any mp3 player supports, even cut up the mp3 if the player supports nothing else.
Of course in rockbox we can have go all the way like I mentioned.

[bk]

I think it would be smarter to add the timecode information as metadata tags within the file. For instance, APEv2 tags on MP3/WavPack/MPC files are flexible enough to have an arbitrary number of timecodes + titles, and can co-exist with ID3. Vorbis comments are already flexible enough for this.

However this is mostly beyond the scope of Rockbox, although support for proper timecode tag handling would have to be added. The majority of the work would be coming up with a comprehensive standard and then convincing podcasters to adopt this new format. Go for it, I think an open alternative to "enhanced" AAC is definitely worthwhile.