How about a decoder for MP3Pro-files !?

I saw the first encoder in the Nero CD-Burning-Software these days ( www.ahead.de ), so this is starting to get REALLY interesting...
I mean MP3Pro's use half the diskspace of comparable quality MP3's - if it's true what they say......


Dr. Pupvogel
mk2 #080000169

There's no ARM implementation - yet.

Rob

Don't expect better quality than lame-vbr at a decent bitrate. Mp3pro seems to target the 64-96k range and doesn't show any better results at higher bitrates under initial testing. Here are some other links regarding mp3pro vs wma vs mp3 sound tests :

http://fastforward.iwarp.com/mp3pro_opinions.html
http://fastforward.iwarp.com/tamsjam.html

-Kit

Any more thoughts on adding a mp3pro decoder to the unit yet / if ever???

Martyn

I personally don't think theres any benefit to using the mp3pro format on the empeg. It seems like it would be better used for small handheld players with small amounts of ram. I would rather see ogg and mp+ decoding.

Sean

Mp3Pro achieves its efficiency by chopping off everything above 12 or 13 kz and interpolating the rest of the sound using the lower frequencies. This obviously doesnt always work very well.

The only benefit I see is that I can make another entry into my comparison matrix to give the empeg a few more points. :)

IMO, there are at least a hundred more useful things to do before thinking about supporting Fraunhoffer's latest licen$ed codec. Pro is in line to compete with WMA, it isn't competition for a well-encoded LAME MP3.

And if anyone is wondering, MP+ is using the MPC extension I believe. Pro still uses MP3 just to make things confusing.

Bruno

Yes. Mp3pro will play on a normal decoder, but the stuff above 12 or 13 khz will be stirpped out. I think mp3 pro is a waste of time, lame has already stretched mp3 quality about as far as it will go. Mp3pro is simply an attempt to save space, which is getting cheaper every day. (for the empeg anyway) The only limit on the empeg is the database size. MPC on the other hand, is a completely different animal. At high bitrates, its quality is unmatched by any other codec at the moment. I believe its based on mp2.

Sean

Does it? MP3 pro files sound awful though normal decoders because you get to hear the 22khz mp3 part only; everyone I know who has heard MP3 pro @64k has been very impressed.

Kind of beside the point though, as unless you're streaming... disk space is cheap.

Hugo

Rather depends on licence fees (per unit, and up-front to get a suitable decoder for ARM); these look like they may be steep, so I'd say the chances were very slim.

Just use high bit rate MP3s?

Hugo

I thought it was the lower frequencies only that you would hear through the normal decoder. There are two streams, one is a low freq mp3 stream and the other carries the info that allows reproduction of high frequencies. Higher frequencies are very expensive in terms of bits to encode with the mp3 codec. Thus it makes sense that this is what mp3pro could save space on and make the files sound ok at 64k.

I got my info from here: http://www.digit-life.com/articles/mp3pro/index.html

Sean

Yes, that's what I said: you get a 22kHz file (ie, max theoretical freq 11kHz) - mp3 pro then adds the high frequency stuff using its own magic frames (the other 50% of size of the file)

Hugo

Does anyone know if there's arm implementation for MPC (MP+) ? This seems like a worthwhile format to look into.

I haven't come across one. If there are any c programmers here who want to give it a try, the source is here: http://www.uni-jena.de/~pfk/mpp/#decoder-source I can't tell if the code is using floating point or not.

Sean