[Hpr] "HPR Dodgers, in the 21th-and-a-half centuryyyyy....." (Future HPR features)

Ivan Privaci epicanis+hpr at dogphilosophy.net
Mon Feb 9 16:14:56 PST 2015

> While we both may love metadata, we do need to keep in mind our mission
> of ensuring the maximum audience for HPR. To that end it explains why we
> still have a mp3 feed at all. 

No argument there - actually even LESS of an argument from me than as recently 
as last year, now that I believe there are ways of working around the 
remaining Intellectual Precious issues (there'll be an HPR episode forthcoming 
about this, I promise!)

I should point out that although I wouldn't shed any tears if HPR actually did 
drop mp3, I'm not one of the participants who has actually advocated that.

> But keeping a simple mp3 file does not
> mean we cannot experiment with new features. There are loads of ways
> where we can move with the times. eg:
> - sending out a lower bitrate mp3, swapping the spx to opus

This may seem strange, but for mp3 what I'd personally advocate involves
sending a HIGHER bitrate mp3...

In order to be most compatible with the original standard and bypass the 
remaining patents completely at the same time, some of the later features of
mp3 encoders ought to be switched off (constant bitrate rather than variable 
or average, for example). In order to maintain similar quality, the bitrate 
may need to be bumped up to 80kbps or 96kbps (if the quality at 64kbps with
the maximum-compatibility/legally-clear settings drops too much).

I'm trying to find the reference again, but years ago I would swear I saw a 
guide to making the "most compatible" mp3 files that suggested only producing 
STEREO (counterintuitively) files, as apparently some of the original mp3 
players assumed that all files would be "music" and therefore hard-coded to 
assume 2 channels. I think this would need to be "real" stereo (two identical 
tracks in the case of HPR's mono source material) rather than "joint stereo" 
if that was the case, though. I'm not sure
if this is any more of a compatibility problem than, say, id3 tags. I've never 
actually heard of an mp3 player that had trouble with mono files, but on the 
other hand I *did* once run into an ogg-vorbis-supporting mp3 player that 
would lock up on mono .ogg files but play stereo ones.

I might also suggest dropping ID3v2.3 entirely and only including ID3v1, if
anything. HPR's currently-used minimal tag set is the same handful used in 
ID3v1, except for the limited space for ID3v1 "comments" which could just be 
summarily truncated or left out.) ID3v1 being a single frame stuck at then end 
of the file, even if it's played on a REALLY bad old player, it should at 
least play the file before choking.  Alternatively, dump all the metadata 
entirely, which would be a minor inconvenience for some listeners but would 
increase the range of older players that could handle the files. I personally 
wouldn't go to that extreme though.

> - adding links to the enhanced media
> - providing a experimental feature rss feed.
> Running an experimental feed would allow you to determine which players
> have issues and which don't. Additionally as the feed would be opt in,
> the participants would have motivation to report issues to you. This way
> we could build a database of known formats with known players. Something
> which would be of benefit to the podcasting community and would
> definitely meet our goal of sharing knowledge.

An experimental feed would be nice, especially if we can get more than just me 
interested in doing experiments with it (and participating in the testing).  
It could double as "the opus feed" until the time when we can replace .spx 
with it. I like that idea.

More information about the Hpr mailing list