Has anyone done any amount of testing at 48k Parametric Stereo between Nero 7 and AACPlus v2?

I am interested in providing Internet streaming/downloads of .m4a files at 48k Parametric Stereo, and would like to decide if I should use either the latest CT AACPlus v2 encoding codecs (as found in either Poikosoft's CD-DAE 9 or WinAmp 5.12) or else Nero 7 for encoding.

Which sounds better to people's ears, Nero 7 or AACPlus v2 at 48k Parametric Stereo?

Any listening test results that have been done would be greatly appreciated.

I am strongly in favor of doing multiformat listening test @48 kbps after the current 128 kbps one - My idea was to include following codecs in the test:

- Nero Digital (HE-AAC (v2))
- Coding Tech's HE-AAC/v2 trademarked as aacPlus (v2)
- mp3Pro
- Ogg Vorbis
- WMA Pro

This test would be a very good check of the state-of-the-art in audio coding, since I believe there has been huge improvement in the 48 kbps area.

I would be interested by an AAC (HE/PS allowed) @ 32kbps. Perhaps an aac@32 would be interesting BEFORE a multiformat @ 48kbps.

Contenders would be:
*Apple, with whatever they can provide @ 32kbps (potentially only LC)
*Nero
*CT commercial
*3gpp reference code

Perhaps other contenders, I do not know who else has AAC encoders that could compete (Real? Dicas?)
Considering the target bitrate, I think that every competitor should be using CBR.
Regarding anchors, I'd suggest Lame in cbr, 32 and 96 kbps.

If you are interested, we could start a thread about this.

I personally don't think that the developer of a competitor should conduce a listening test.

I never did any testing but the winamp encoder sounded markedly worse than the one that's hosted on the dbpoweramp site (i think it may be the 3gpp reference encoder) - enough that i haven't bothered with it since anyhow. Maybe i stuffed up the settings but honestly there's not much that can go wrong...

guru's test showed that winamp's ps-aac at 48kbps came close to 64kbps winamp/helix he-aac on various music (and was slightly worse on classical)

winamp/helix he-aac clearly outperformed often used nero he-aac at 64kbps


edit:
i for myself would love to see an aac comparison at 128kbps between
- helix aac (btw freely available)
- winamp aac
- apple
- nero

as helix/winamp (both maybe coding technologies, but its not prooven) performed so good at low bitrates they might be able to beat apple too at 128?


edit2: just grabbed the helix aac encoder dlls and in the readme it says:


also it talks everywhere about aacPlus and not he-aac which also points to being ct's codec

Try it with the Nero 7 encoder instead and then we'll talk again

meh, nero is always promising that a new version is so heavily better, still...

He never said that HE will conduct the test.

I definitely won't - exactly because of the credibility problem, but there is another idea that could actually make sense (will explain in the next few weeks)



Nero 7 HE-AAC encoder is heavilly better than N6 HE-AAC, and it could be easily tested even by completely untrained ears.

Like I said, I'd like that someone (Sebastian?) performs a new listening test @48 kbps which would be much more credible than a) marketing and b) test performed by one person

Ah, seems I misunderstood the "doing multiformat listening test". Well, after the 128 kbps test is over, I think I can run a poll or something for what is to be tested next. Either an extension to the 128 kbps test including ATRAC and MusePack, or a new 32/48/64 kbps listening test with various formats or only a listening test featuring AAC encoders.

well i today sniffed through all sorts of listening tests regarding aac and pretty much all show that nero isnt able to clearly beat apple
and i also saw that nero devs tend to argue as an answer to these tests that their new things will perform sooo much better, still the listening tests hardly prooved this to be correct, or better said nero still was not able to clearly beat apple

now we have the example where guru used in his last test a version of nero that is now ~ 1 month old but suddenly there should be a new nero encoder out there that performs sooo much better? cant be...

but we will see how sebastians tests turns out....

@Bond, before replying further please read my reply on the listening test topic - I am afraid I am not able to debate this furthermore because I have not too much else to say.

Let's wait for the end of the current listening test, and then continue, ok?

Bond,

You are basing your "bashing" based on listening tests that are clearly out-of-date. I am dismayed as to why you don't select a half a dozen music samples and have a listen for yourself. As has been suggested, at these bitrates you don't require "golden ears" to make a determination between these 3 He-aac encoders. The differences are so obvious here, but I will let you determine what sounds best for yourself.

Indeed - almost everybody could actually do a 48 kbps listening test - artifacts should not be so hard to spot, and clearly the differences would be audible to the ordinary people, hence I don't understand anyone "waiting" for someone to do a test for him at this bitrate.

But, I am clearly in favor of well organized, independent listening test, where there are at least 20 or more people taking part in exactly because of credibility and wide reliability

Therefore - my vote is clearly for the 48 kbps listening test

I'm feeling like organising one.

Readying my samples now. If enough people ask for a follow up, I'll submit them.

I would advise for organising such a test after the current 128 kbps one because it would be clearly too much tests at the same time - which would also have impact on number of people taking part.

All right, but I can still start planning it, right? Could anyone who can help me set up ABC/HR files with keys, process data files, etc, contact me by pm or email please?

I'd also need to know
a)How do I check what versions of each encoder I use?

b)What settings should I use (eg, on nero AAC, winamp AAC, itunes AAC, oggvorbis and lame)

And dont forget about AACplus2 of CD-DA extractor it has an updated AAC+2 by codingtechnolies as in the last Winamp 5.12. The lowpass now is 20.5 khz and not 16-17 khz as before in old versions. Quality was improved.

I thouht it was clear for all new Nero7 AAC+2 is very better than Nero6 AAC+ at all bitrates not only at 48 kbps PS.

We defeantely need a good test at 48 kbps

I could confirm that current Nero Digital HE-AAC profile is much better than the previous one I tested two months ago (less SBR artefacts/sandy noise). It's now clearly competitive with Coding Tech. encoder. I can't say from my short evaluation which one is better, but there isn't obvioulsy a big difference anymore.
A listening test would be nice.

Ivan> WMAPro at 48 kbps is difficult to get

I'm surprised that you guys have waited so long to push for this shootout, but I'm also surprised that the proposal is for 48 kbps rather than 64 kbps. At so low a bitrate the only possible contenders are differently tweaked HE-AAC implementations.
In fact, I believe Nero 7 so good that I went back to my FLACs to convert everything to this newer implementation after having used the one from Coding Technologies via Magix (and, subsequently, the Winamp one).
That said, I think there will have to be many blindfold safeguards to dispel doubts about the tests' fairness; it may be assumed that a few of the most prominent contributors to Hydrogenaudio were consulted and paid in the work producing Nero's new codec. Mind you, I'm not saying that that's wrong, since there aren't all that many audio codec mavens in the whole world. But just bend over backwards to assure that people will credit the tests as impartial.
I think the slowness of moderators to approve as a news item the introduction of Winamp 5.10 with CT HE-AAC injured HA's reputation.

I wouldn't call a 2-3 months old listening test "clearly out-of-date"
I personaly admit that my test is now outdated (because I noticed the progress between the version I tested few months ago and the version released with Nero 7), but Bond has the right to be perplexed. Generally speaking, after each listening test (personal ones or collective as well) a new updated Nero Digital encoder claimed to be better pop up like mushrooms after a rainy night. This can perplex people and it's not necessary "bashing" to be doubtful. People have the right to be critical and to not blindly believe what developers and/or marketing people are saying. It's a minimum we could expect from an "objective" community.
What really matters is to evaluate the reality of the claimed progress. It's a minimum we could expect from an "enthusiast" community And in this case (HE profile between Nero 6 and Nero 7), progress are not only real: they're big (in my opinion).

More than it would have hurt to approve news about leaked software?

my words thx

now lets stop this discussion and wait for sebastians test to finish

Thanks for all your comments to this thread I started. I look forward then to a future 48k Parametric Stereo listening test.

I understand that the 128k test is first, but I am glad to see such support for a 48k parametric AACPlus v2 vs. Nero 7 test, because I don't have a clue which one is best to use for creating my Internet .m4a file song downloads.

I am a music publisher and need to know which is the best one to distribute reasonably good sounding 48k bitrate songs to my users. Bandwidth is an issue, so I want to go with 48k rather than 64k HE-AAC/AACPlus. Thanks!

Winamp 5.13 will feature improvements to AAC+ encoding. Not sure if it will be out in time for this test, though

Timetable for approximate Winamp release? I sure would like to see the latest AACPlus v2 encoder (CT) used. I think Poikosoft has the new encoder already in their app. I guess Winamp just updated the decoder in the 5.12 release and not the encoder.

If the timetable is reasonable, about January 1st, for the new Winamp release with updated AACPlus encoder, then I hope the 48k test can wait until then. Also by then Nero may have an updated Nero 7 HE-AAC encoder also.