AAC player/decoder ?


I'm looking forward to a AAC implementation (perhaps in the GP32 internal player) too. MP4 is the definite upcoming standard. MP3 is dead in my opinion. To have a good sounding MP3 file on your GP32 you need at least 160kbit-192kbit MP3 no matter if VBR or CBR and with this... a 128MB SMC is very soon full (with only about 70min playtime). Everything else is an sound artefact war if you transcode existing files (because you don't have the CD at hand or you downloaded an already tight encoded MP3)

AAC 96kbit CBR has the same quality as MP3 128kbit CBR and is 30% smaller than the MP3 file. Not to mention that it sounds way better. Good thing on it if it's implemented in a media player... it doesn't eat more CPU power than decoding a MP3 file.

Guess I really have to write K2Y and pray to the lord for a possible implementation, because no alternate player for GP32 was satisfying so far IMO (usage technically).
 
It's not about patents... it's about standards. MP3 and AAC (MP4) is a set standard by the MPEG association. OGG was just a hacked MP3 codec (and is a codec I don't support - sorry). AAC decoders are freely aviable everywhere on the net. Only AAC encoders aren't free.

Besides... thanks to the pals over at "Hydrogen Audio Board", they made in-depth tests with several codecs and AAC is under the top5 (quality technically), OGG is somewhere around place 10-15. Not to mention that WMA and OGG still eat more space on a SMC than an AAC file encoded in the standard LC-AAC profile that resembles the quality of the other codecs. And with the HE-AAC profile you can even go down to 64kbit and still have CD quality, by only a third space compared to a regular MP3 file.

That's why I want to see an AAC decoder implemented in a player. It's way more versatile.
 
Firstly there is only one ‘free’ source code decoder for AAC and that is FAAD that was written by Ahead (Who make Nero CD Tools) and GPL’ed.

The last time I looked at FAAD (Pre V2) it was using a lot of floating point code and was not really suitable for porting to integer only devices like the GP32 without a massive amount of work. I have not seen anybody working on an ARM port (the basis for it working on the GP32) as there are several commercial ACC decoders for ARM that a company could licence to add ACC to there ARM bases media player, WinCE device etc.

Secondly, there are still some grey issues over the use of ACC and its patents etc. that are hampering its growth in an open source community.

OGG Vorbis is not encumbered by patents or costly source licences and not a hacked MP3 codec :eek:. The newer versions of the Vorbis codec should match or surpass ACC in audio quality terms (or soon will) and OGG maintain a VERY good integer math decoder (Tremor) that is very easy to integrate into a player for something like the GP32.

Please don’t confuse free player plug-ins with open source decoders that could be used as the basis for a GP32 player. For reference most (all?) free AAC decoders are derived from FAAD and all require good floating point to work well.

Oh, and aren’t all these audio tests a little irrelevant, this is the GP32 we are talking about not my HI-Fi with a load of Burr-Brown DAC’s and costly interconnects ;).
 
Sorry if I say this, but this is just "blah!". <_<
I've seen the issue "you can't code this/that, because it isn't ASM code" or "the GP32 can't handle that, you have to rewrite it" etc so much on this board and still someone took it on and re-coded it, no matter what it was (for example the SNES emu, a GBA emu, perhaps soon even a DOS emu!). Why else would you think it is possible now to listen to WMA songs or OGG? You think the codes were already in ASM? So this is just another nonsense I've heard lately - sorry. The GP32 is capable of using AAC (theoretically), why shouldn't it if AAC is just an "enhanced MP3 with new psychoacoustic encoding matrix" (if we keep it simple). It even works on my Pentium 100 (sorry... don't have a 486DX-2 66Mhz up and running anymore) flawlesly without massive CPU usage.

AAC or other said MP4 is the definite future. It will tag off MP3 the next years. It's used by Apple (iPod for example), by Dolby (AOL Radio), and it's also well known in the audio scene and also massively used (webradios, online shops, a handful of music communities use them already too). So it isn't a greyzone either. It's only a matter of time until free-encoders pop up for AAC too, like they did as MP3 was "outstanding" and "new". And so what... a "Nero" encoding license doesn't cost a "kingdom", it's freely implemented in Nero6 and this is still one of the most spreaded burning software. Speaking about OGG... AAC might still be at the beginning (even though it is already something about one year old) and it's that good, no matter what you might say about "OGG" and new encoding matrixes... it will never touch AAC.


Besides... The FAAD isn't the only free decoder. Just go over to www.rarewares.org/aac.html and you not only find "decoder plugins" for Winamp, Sonique and other players I only heard of, but other stuff too. And they're not made of the FAAD one. Hell no.

If WMA is possible too (and this is a Microsoft only license!), then AAC is also possible. If I could code for GP32, I'd do it by myself. If you can't do/won't try it... then leave it be but don't tell me it isn't possible when the GP32 is just a small PC in pocket form. Hell it doesn't even have to be capable of using multichannels, stereo is more than enough. I'm sure someone can and will take on this task. But like I said... no actual free player can beat the original K2YAA one so far if we talk about layout and usage.

BTW: I already wrote K2YAA about the request of implementing AAC. Don't have an answer yet, but I'm sure that they're not against it.

We will see what happens in a couple of weeks... ;)


Speaking of audio tests:
Well... I never said something about DAWs, I said something about quality comparison. No matter if it's a PC or GP32... the codecs are still the same.
 
Compyfox,

You are free to go and write a GP32 player, all the SDK is available and you can build your own compiler using free tools (like mine or RobBrowns GCC guides) so go and do it and stop saying you won't/can't, you would soon find out what problems await your average GP32 developer :D.

It does not have to be ASM code to work on the GP32 but it does need to suit the GP32's architecture (ARM with no floating point) and please don't lecture me about OGG Vorbis, I am very aware of how it works as I ported it (and libMAD for MP3) myself for use in ScummVM and my own media player, this also means I have a good idea of the math involved, it's not lightweight stuff, it really messes with your head ;).

Oh, and you totally missed the point about 'free open source code' that a developer could use to make a player and 'free plug-ins (without any source)' :D most of the plug-ins you mention ARE derived from the FAAD or FAAC source, it even says so on the site you linked to ;).

For note, the WMA player is derived from ffMPEG (I think) and ffMPEG has a totally open WMA decoder in it that was re-engineered by hand by the developers without using ANY Microsoft tools, that is the only reason it is not covered by MS's licence (much to there disgust), the GP32 WMA player is a VERY cool utility but it's author did not reverse engineer Microsoft biggest codec on his own ;) (Though getting it working on the GP32 is no small achievement).

And I used the term 'grey' in a legal sense, not a community sense, open-source or not software still had to abide patent laws.

And as for codec’s, the quality might be an issue and your quite correct AAC is technically superior to most other codec’s currently but the GP32 has a certain level of hardware used for sound and no codec is going to improve that.

Don’t get me wrong, I think it is a great idea but there are a number of things that need to happen before a GP32 ACC player would be viable IMHO.
 
Compyfox posted on Apr 18 2004 said:
OGG was just a hacked MP3 codec (and is a codec I don't support - sorry).
Uhum, care to give any proof? Vorbis was developed from the ground up, it has nothing in common with MP3 apart from being a lossy audio codec.
 
Last edited by a moderator:
Back
Top