A few notes concerning CEP2

regular mp3

There are now three mp3 settings within CEP: high quality, medium, and fast. In reality, they are all based on Fraunhofer's FastEnc codec, so they all encode very fast. The very slow "Alternate" codec appears to have been banished.

The fast setting does not lowpass, even at 128 kbit/s. The high quality setting is the one recommended by the program, and in fact, I don't see the need for the other two settings because normal FastEnc is already so fast.

The stereo collapse bug (which Cool Edit Pro 1.2 avoided, in any case) is now a feature, and may be useful for low bitrates.

Otherwise, the basic FastEnc codec does not appear to have been modified much, if at all. Pre-echo on castanets.wav sounds pretty much the same as it always has, and main_theme.wav still sounds as bad as it always has with this codec.

mp3pro

The interesting new addition to CEP2 is vbr mp3pro. At 100% quality, it apparently yields bitrates on the order of 130 kbit/s. If you play a 100% vbr mp3pro file through a regular mp3 decoder, the lowpass cutoff is 10 kHz. Otherwise, through an mp3pro decoder, the cutoff is > 20 kHz.

Newkid.wav using 100% vbr mp3pro sounds about the same to me as regular FastEnc cbr 128. Both have noticeable pre-echo. Castanets using 100% vbr mp3pro sounds worse (although not terribly worse) than FastEnc cbr 128. They both have similar (not too good) pre-echo, but the mp3pro has more "splattering" of sound outside the center where it should stay during the fast castanets section. This sounds like a last remnant of the "shaking rice" sound which is so terrible at 64 kbit/s.

Considering the compatibility issue of mp3pro, I probably wouldn't use vbr 100% personally. But perhaps people with very good high frequency hearing (i.e., above 16 kHz) might find it more pleasing than regular mp3 at 128.

ff123

hope you don't mind me translating and posting your conclusions on another forum

I remenbered that mp3 with no lowpass filter is not a good idea. So I want to know how's the fast setting with CEP2?

I'm reading this forum for very long time and I really believe you every word u say but I always trust my ears more!
I have just tested Mp3pro VBR 100% and it sounded to me very very but so very close to the original WAV file!!!
Before you say for my hearing I have to tell you that I can tell the deferens between original wav and –aps! Also I can recognize any Ogg at any bitrate because of the artifacts it produces!
Is there any possibility to overreact because mp3pro is not free?

I’m looking for so long a format that’s going to be a standard for the future and have much better quality than mp3 that mean that I’m afraid to use MPC because I don’t know if it’s going to die after some years. E.g. VQF

So I’m thinking that the only solution is AAC! There are already hardware devices that supports it. What do you think?

I think you are right.

Regards;

Roberto.

(Yes, I'm an extremely biased AAC lover)

Nomedor,

What exactly do you find that is wrong with mp3 using LAME VBR? I am using LAME with --alt-preset standard and it sounds very good indeed to me. I bet only very rarely will you find a file that is distinguishable from the original uusing --alt-preset standard. I have not tried the higher quality presets.

I suppose you could just encode with "--alt-preset standard" then if you find a song that is distinguishable from the original use (a) a higher quality LAME presets (b) lossless encoder or leave it as cd-audio. It depends what you want to play it on - if portable device go for mp3.

mp3 will be compatible with very many devices for a long time to come.

With the reduction in cost of media (soon DVD-R, etc will be commonplace) you could soon put enormous amounts of even lossless encoded audio on one disc.

RD.

To tell you the truth I’m generally satisfied with –aps is is really good that why I haven’t jumped to another format till now but I’m also only human (greedy)
So I want something to do the job better and at lower bitrates
AAC is a creation of fraunhofer and better than mp3, so it is the future. I believe that mp3 is still alive because the guys from fhg don’t want to give to the public something better than mp3. As far as I can see mp3 is a pain in the a** for so many companies because of the money the loose! They would be happy if the only solution we had was WMA which has so many protection for copies!
If people had an encoder than can produce really high quality audio files and also speared like mp3 they would only loose more money!
For now fhg is the queen of loossy audio compression and I’m just following!
I really hope that one day Ogg will become good and known enough so I can move to it permanently!

:confused:
Rjamorim can you please confirm to me that Psytel AACenc is compatible to FHG’s AAC project! Will I be able to buy a harware aac player tomorrow and play the files I created today!
Thanks in advance!

Eheh. FhG doesn't matter at all!

Psytel must be compatible with MPEG AAC project. It's MPEG that defines AAC bitstreams specs, decoder specs, goals...

Psytel AACenc is 100% compatible with MPEG AAC. FhG AAC is compatible too, according to the few test streams floating around created with FhG AACdemo 2.2.

And yes, if an AAC hardware decoder is ISO compatible, it will play streams created by Psytel AACenc.

But you should be aware that an ISO compatible decoder doesn't need to decode every kind of stream Psytel AACenc creates. One good example is Philips Expanium. It plays only MPEG2 AAC LC. If you try to play an MPEG4 AAC LTP, it will see the file as corrupt.

There are other ISO AAC hardware players around, like
egoman. No info on these, though.

Regards;

Roberto.

Hmm, I can't quite agree with that. Basically I find clear problems with almost all music I tested. Especially bad is pre-echo handling and cymbals. My settings are Mp3pro 100% VBR and allow mid/side. If you demand, I can provide some clips.
Hmm, don't take this wrong, but if you give the impression you can hear the difference between original and aps/any Vorbis, it should be very very easy to distinguish MP3pro 100% VBR. It fails badly much more often than aps or Vorbis at high bitrates.
No. However, I'm not signing that Vorbis is better than MP3Pro/MP3Pro VBR at low bitrates, like some people say. This may change after Vorbis RC4 is released. There will be new techniques to improve Vorbis low-bitrate coding (noise normalization).
Seems that AAC really has just started its journey. I don't think raw-ISO AAC will be popular with masses in Internet though (in p2p etc), at least anytime soon. Vorbis has a better chance for that.

IMHO sound of MÐÑ is better then AAC.
Even MÐÑ will die - it does not metter, course I
will be able to use MÐÑ always :^)

JohnV;
The reason I can tell the difference between the ogg and the original wav file is because Ogg produces some certain artifacts at any bitrate(a metallic sound)! I don’t tell that mp3pro is so perfect, maybe I’m not so sensitive to mp3pro’s artifacts! I didn’t make any blind test.
But I was really confused when I saw this post, because mp3pro with MS sounded really good to me! Especially @VBR128! Mp3@128 is far behind! Even –alt-present 160 is worst to me! –aps is better but its bitrates are ~200 kbps!
Maybe the reason is because I use mp3 for 7 years now and I can hear the artifacts it produces more easily!

I really like OGG’s idea of a free format but it’s not mature yet! Maybe after some time and hard work! I thank the people that made it and proved that not everyone wants to put their hand to our pockets! As I said before I’m waiting for the day that OGG will be ready for the masses!

I heard that digital radio will use AAC!!!! I don’t say I’m sure, I just heard it, if big companies are going to use AAC and send money for it maybe they know something more than me!

T_vitaly what is MÐÑ? Do you mean MPC

Yes. Digital Radio Mondiale ( www.drm.org ) and XM Radio ( www.xmradio.com ) use AAC+ technology (AAC+SBR) in their streams.

Menno is implementing DRM streams decoding in FAAD, and Ivan is implementing SBR in AACenc and FAAC.

Of course, many companies are going to use/are using AAC, because it's an ISO standard, and that means a lot in some kinds of business.

Regards;

Roberto.

If you are going to follow it then in essence you are more than likely physcologically always going to claim it's better than Vorbis without performing any blind listening tests as JohnV & Garf speak of below, why compare the listening test to Vorbis anyway if you are to make that general assumption? Perform a blind listening test than get back to us.


Time is always an issues, because many lead developers of Ogg Vorbis Codec don't work on it 24/7, I think that they have put a lot of hard-work into Vorbis especially over the last year, if they did anymore hard-work they might as well be a standards group like Fraunhoffer IIS , hopefully someday it will appeal to the masses which is what I know I want, but until then we will just to have sit back and relax and let "work" progress. In the meantime we could always always keep testing the physcoacoustics model against various PCM, RAW samples, etc. in order to see how well various intstruments perform on the CODEC.

Do you have a clip where this happens with -q6 or greater? I haven't heard this one before.



Bad, bad idea.

--
GCP

I strongly suggest you do blind tests. I never claim anything unless I've done at least basic blind testing.
MP3pro is just normal mp3 at the lower frequencies and totally artificially created SBR frequencies at higher frequencies. If anything, it should be much worse for people used to MP3 artifacts, because in addition to normal MP3-data there's artificial SBR-data. There's nothing "pro" quality wise in Mp3pro for higher bitrates.
Yes, Vorbis has some problems typical to it.

Anyway, I really urge you to do blind testing. Psychology plays a too big part even for experienced listener.