Rockbox General > Rockbox General Discussion
WMA support status?
soap:
--- Quote from: danmint on March 10, 2007, 09:45:16 PM ---Another thought is people with a clue wouldn't use mp3 either...
--- End quote ---
MP3 is universally supported by audio players, and while it might not be the most advanced codec when judged on technical specifications alone, the existence of a mature encoder (LAME) makes up most the difference.
Vorbis, AAC, MPC, and MP3 all reach transparency at about the same bitrate.
CSMR:
All the tests I have seen have put Vorbis (with the later aotuv encodings), AAC and MPC ahead.
I haven't seen anything about reaching transparency. Most tests are done at under 200kbs at which none of the codecs are transparent (last I heard).
Trouble with wma is it is a lot of different codecs. Wma standard is the most popular and well supported but not the best choice for encoding.
saratoga:
--- Quote from: CSMR on March 23, 2007, 11:30:54 AM ---All the tests I have seen have put Vorbis (with the later aotuv encodings), AAC and MPC ahead.
--- End quote ---
This is the most recent widely accepted cross format listening test:
http://www.listening-tests.info/mf-128-1/results.htm
As you can see, all major formats tied for first place.
--- Quote from: CSMR on March 23, 2007, 11:30:54 AM ---I haven't seen anything about reaching transparency. Most tests are done at under 200kbs at which none of the codecs are transparent (last I heard).
--- End quote ---
In the above test, all codecs were essentially transparent at 128k. People encode at higher bitrates to guard against the occasional problem samples, because of habit, or for peace of mind.
soap:
--- Quote from: CSMR on March 23, 2007, 11:30:54 AM ---All the tests I have seen have put Vorbis (with the later aotuv encodings)
--- End quote ---
I love the idea of Vorbis, but unfortunately developement is essentially dead. (At least for the time being.)
The aotuv tunings are just that, tunings. LAME, on the other hand, routinely bores and strokes the engine of MP3, proving that active, consistent, development of a more mature encoder can make lemonaid out of lemons.
AAC might not be nearly as warm and cuddly as Vorbis, but Nero is showing, much as the LAME project does, that working and reworking an encoder can produce significant improvements in output quality, even while the specification remains unchanged. AAC is, most likely, the successor to MP3. Cell phones support it almost universally now, iPods support (a subset) of it out of the box, and the quality at low-bitrates is improving quickly. The transparency battle has already been fought, and it was a stalemate, with favor going to the one with most penetration (MP3). The low bitrate war is already in full-swing, and unless Vorbis awakes from its multi-year coma, it doesn't have a chance. (MP3 isn't even showing up to this fight.)
Superior technology is rarely enough of a mega-weapon to win a war*, and superior specifications even less so. Marketshare and mindshare almost always rule the day, and Vorbis offers no compelling reason to those other than ideological converts.
But that's just my opinion, I might be wrong.
*Not to mention that Vorbis, as it stands today, is not superior to AAC in terms of features.
EDIT:
Deleted my later message, would delete this one too if it hadn't already been replied to. I've already gone far enough off-topic.
CSMR:
saratoga
Yes lame does look pretty good in that test and is not behind the others (significantly).
The best the test could do is not disprove transparency. I don't know the rules of that test. Would the original wav have got 5 exactly?
Here is a test including the same LAME and vorbis encoder which doesn't get transparency for any codec at 180kbs:
http://www.hydrogenaudio.org/forums/index.php?showtopic=36465
Average equipment but probably good ears and experience. LAME is significantly (statistically) inferior to aotuv vorbis. And everything abxable from wav.
I think this shows that those codecs are not transparent at 180kbs, so also presumably not at 128kbs.
"Transparent to some" no doubt, but not to all, so... not transparent.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version