guruboolez, just read this comparison and again a great job, thx a lot!

now i have some proposals:

1) you are doing so many great comparisons, its hard to keep tracking. would it be possible for you to create some small webpage pointing to your tests, like the one of rjamorim, so its easy to see whats going on and compare the results aso... atm it seems more to me your results are vanishing in the depths of the forum
if you cant make a own page, maybe you can create some thread carrying that info?

2) imho this test once again showed that nero isnt able to beat apple, it can come close but i never really saw it to be clearly better. isnt it time to move on?
your he-aac comparison showed that helix and coding technologies aac outperformed nero also at the he-aac level clearly

wouldnt it be more interesting to compare helix/ct to apple in the future and not always nero over and over again? (i know nero is well established (again is used in sebastians public test), but this doesnt mean that there arent better codecs)

When a codec is still being actively developed, I see no reason not to compare it over and over again.
Even though there's already a consensus that it won't par so well against others.
Who knows, one day it might really meet people's expectation(s).

Moreover, if a certain codec isn't included on a test, some people might call the test 'flawed' or 'unfair'.

Yes, but to determine if the other encoders are better than Nero, you obviously have to include Nero. Comparing the results from a listening test with the results from another is not so simple because you most likely have different samples, different testers, etc.

well, i am not interested in how nero performs, but what the best aac encoder is, and as nero isnt really able to beat apple, i am interested in comparing the new codecs to apple and not nero

*pessimism detected*

Well, anyway, I don't think it's good to leave out Nero.
If Nero, like what you've said, won't be able to catch up w/ QT and/or Helix, then it's fine to make it a low anchor or something, no...?
And it's not like you're the one doing the tests so, let's just leave it to the tester's decision.

It's not that I'm a Nero fanboy or anything (heck, I rarely use AAC anyway...), it's just pretty subjective to tell that Nero won't be able, ever, to catch up w/ the other two.
Personally, I won't say an encoder's not worth mentioning if it's still under active development.
Yes, it's probably still far for Nero to catch up w/ the others and Garf's bickering on tests won't solve anything and that Nero's marketing's kinda misleading, but hey, I'm just trying to be objective.

the point is most people, except maybe nero devs and enthusiasts, are simply interested in knowing what the best aac codec is (eg at 128kbps), cause they want to make encodes with the best codec available

we know nero isnt able to beat apple at 128
so it would be interesting to see how other, till now pretty rarely tested, aac codecs, like winamp perform compared to apple

all i am proposing to guru is that if he is interested in it he might want to use winamp in a next test (and propably skip nero), as there might be a big potential in it

@Bond - There is a completely new AAC encoder branch (v3.1.0.x) used in current 128 kbps listening test, and I strongly believe this statement is not true - at worst, I am quite confident Nero and iTunes would be tied.

An AAC comparison without Nero would be rather incomplete. Besides that, you are vastly extrapolating on results from one person. No one can conclude so firmly (as you do) that Apple is better than Nero from this test. No one, but guruboolez.

Like stephan said, guru knows.
And yet, it's based on what's available at the time of the test; who knows about the coming months?


And you might as well skip a (probably) big potential in Nero's encoder.
Just keep in mind that Nero's AAC isn't as dead as FAAC.
It's gasping at the time, but still breathing nonetheless. =)

again all i am proposing is that guru might want to consider using winamp aac in one of his next comparisons.


-----------


ivan:
well i was summing up the tests done till now, where most tests show that nero isnt able to clearly beat apple

the nero encoder guru used in this test is now little bit more than a month old (he wrote "I used the very new encoder released two weeks ago") but suddenly there should be a brand new nero codec available being tied with apple? come on

but yeah sebastians test includes the new nero codec and we will see how it performs and if the nero devs's standard argument against worse performance ("if a newer version would have been used it would have performed much better") is not again invalid

Ok, let me get the facts straight:

Nero 7 encoder used on the launch time contained new SBR encoder, as well as some tweaks to the current (old) LC core.

Encoder used in the test contains brand new LC-AAC core, rewritten from the scratch - that simply wasn't ready for the N7 launch time. For the purpose of this test, it has been "rushed-in" and added for 44.1 kHz, LC, Stereo modes.

The reason is - these things need time - a lot of time, and I am sorry they cannot be delivered in the frames some might expect.

I am quite sure that Guru could find a difference in these two encoder branches if he is willing to test them



Exactly.

Look, I have no reasons to lie or make a spin here - purpose of my presence on this forum is to help and inform the people, and not something else. If that is hard to believe, oh well.

again all i am proposing is that guru might want to consider using winamp aac in one of his next comparisons, cause i am sure most people are interested in knowing and using the best codec available


no need to further discuss my views about how nero devs find excuses for the worse performance of their encoder compared to apple

And all I'm commenting is your 'other proposal' of not to include Nero's encoder on the next listening test.

Aa, but do "we" also know that Apple is able to beat Nero in the on going 128kbps listening test?
It will be interesting to see the results of the test, and then interesting to read your comments afterwards too.

You clearly haven't at least been following the discussion here, cause it has been said many times we have a completely new codec used in the on going public listening test. Hope you are doing a better job on D9..

The problem is that Nero has used that as an excuse for years now. "the next version" is always said to be a huge improvement, but then Apple's current stuff always beats it.

So?
Let them have their own excuses.

If their encoder is as good as advertised, then all is good.
If not, then I'll just use iTunes' or Helix'.

Simple as that. =)

Excuse me, but this is clerly an abuse of what I said:

- I claim that the Nero AAC Encoder 3.1.0.2 is completely rewritten (LC-AAC) and it does not use old LC-AAC codebase anymore for the 44.1 kHz, LC-AAC, Stereo mode (it uses it for other modes ATM, but this will change soon)

- I also claim that the Nero AAC Encoder 3.1.0.2 could be, at worst, tied to iTunes, and this is the result of the internal listening test - final proof (or proof that I am wrong) will be seen at the end of the public listening test - without this, we could just argue forever until someone comes up with some real scientific data

Now, I never use anything for "excuse" - it is a fact that iTunes won over Nero over previous few LC-AAC listening tests and this is something that cannot be excused or ignored - but franky it has nothing to to with what I said.

I would personally like if Guruboolez could check the latest encoder (3102) out - I am sure he has it, it is probably just matter of his own time and resources and willingness to conduct such a test.

Last but not least - I am not visiting and posting on HA as the marketing/sales person of the company I work for - I am member of this forum long before I joined my current employer (and, funny enough, long before some people actually abusing information I posted here), and I certainly do not enjoy someone (Not Busemann, I just quoted him for misunderstanding what I said) twisting my words, especially if that person's credibility falls to reploying "We all know that Company XYZ is evil"... sorry, but that is just not the level of discussion I am willing to be part of.

Just as a note for people curious about AAC and flash players, I encode stuff for my K750i using QT Pro 7.0.3, and 80kbps, 32khz, Stereo actually sounds really decent. iTunes doesn't allow you to set 32khz so you need to use QT player or something else. It's a CBR mode with a 15khz lowpass which seems to be a good match for the handsfree/earbuds that come with it. I think this is about as good as LC-AAC gets at this scale.

I experimented with the 96kbps 44khz VBR option earlier and a lot of tracks had really loud and obvious artifacting, even going up to 100kbps+ so i don't think they tuned that setting much if at all.

Can fit a good 90 minutes on the 64mb memory stick it came with. Unfortunately the solution works *just* well enough that i haven't bothered buying a new ipod yet to replace my last one which got stolen.

Just out of curiosity - was older versions prior to 3.1.0.2 based on your old psytel aac code?

And...



Thanks Guru for your quick test - so, @Bond - I guess this means I don't really talk crap, do I?

So Ivan, the Nero AAC encoder used by Sebastian in his test is newer than the encoder we have in Nero 7?

YES! Completely different, written from scratch, new codebase. BUT, it will be in the Nero7 in the near future.

@guruboolez
ic, well i just wondered because afaik there is no single more or less representative listening test comparing ct with apple, or am i wrong? and no updates doesnt mean imho that it shouldnt be tested at least once
so i just was proposing that this should be done at least once so we know for sure

well i will propose this too for the follow up of sebastians test, if its not done i will have to test it myself (cant be i am the only one wanting to see such a comparison)


of course you didnt talk total crap, still we will have to wait for sebastians test to finish

That's true. But when you perform a listening test there are several possible encoders that could be tested, especially for AAC. As tester, you must make some choice. For this test, I explained the principle: “I restricted the test to the most usable and interesting encoders”. There were no major update for CT LC-AAC since latest Roberto's listening test, and CT implementation [Real AAC] didn't performed very well two years ago. That's why I didn't include CT LC-AAC in this test, and prefered the new Nero Digital AAC encoder to any other stagnant contenders.


But I did it, this summer, at 96 kbps:
http://forum.hardware.fr/hardwarefr/VideoS...jet-84950-1.htm

very interesting thx
seems real aac is on par with nero + is freely available

couldnt find this comparison as you didnt post this on ha it seems? (thats why i propsed this central "guru's tests webpage"

“was” should correspond more to the exact situation. Nero Digital has been updated, not Helix.

Here is what independent group listening tests said

http://www.rjamorim.com/test/aac128v2/presentation.html





Looking from the confidence interval, I would say that Nero was statistically better, would you agree

That was age old Nero AAC version and the current one is for sure quite better... and I haven't heard that Real updated their LC-AAC codec (maybe they did) -

According to that test, FAAC performed a little better than it, and to use Bond's words "seems FAAC is on par with Real (actually it is a bit better than it) + is freely available "

How about that when we are about to quote tests? Or you prefer quoting ones you especially like

So that was the case last summer at 96kbps for you Guru, thanks. The new LC-core which will be released soon (and is in test in the current 128kbps public listening test) is imo very big improvement. And the improvement will be even bigger on lower bitrates like 96kbps, so it will be interesting to see your test results in the near future.

you are wrong, real updated the codec since then, so that test is outdated


may i quote yourself:

the facts are there is no new comparison comparing real with nero, so the currently known status is that real and nero are on par at 96kbps lc-aac, period

the rest is speculating based on promises by nero devs that new nero is sooo much better, but we already discussed that...

edit: seems johnv didnt follow the whole discussion till now bringing in again the old nero argument...

Which Guru just confirmed for both SBR and LC (there are big improvements) - and you seem to ignore that fact, don't you?

But please, keep going on - there are still few weeks more time to do it, and then we'll all have a good quality time laughing at some of the posts (hopefully not mine)

Bond: I haven't used the argument that we have a new LC-core ever before. Try to find one post from me where I say our LC-core has very big improvement (and it is actually in release state and under public testing), until very recently. You will get a prize if you will find such a message from me and paste the link here (goes for everybody else too).

Yes, and that's precisely the reason why I tested Nero Digital instead of Helix/CT. Nero Digital is the only encoder which may compete with iTunes

One precision: I noticed big improvements for the HE profile. For the LC one, I noticed improvements over aacenc32.dll v.4.2.1.0 (with tonal samples, harpsichord and low volume, which regressed compared to 3.2.0.20 “fast”), not the previous [nero 6] encoder. Between Nero 6 “fast” and the new generation encoder (why the hell don't you call it aacenc v.5!! ) I didn't noticed improvement for quality, but for efficiency (no bloated bitrate anymore on specific samples).

well again, we will see if nero will be clearly better, or at least clearly on par with apple, in sebastians test at 128kbps

for me the currently proven status is that nero is only on par with real at 96kbps, worse than apple at 96kbps and worse than apple at 128kbps in the lc-aac field

We will as soon as it is complete



Indeed - it is more about the semantics, However, if the quality is the same, and the bit rate is much lower - I would still call it a quality improvement since the quality power of the codec is actually measured in quality per bit rate, too

Bond: I doubt it's possible for any codec to be the statistical winner in this current 128kbps test. Many of the codecs are so good. Though you will see that Nero ABR uses much more controlled bitrates for the samples than some of the competition, so take that also into account if results end up pretty much even.

Nobody dispute this. But I don't understand. On one side, you seem to reproach to people to test Nero Digital (over other encoders) and on the other side you're waiting for listening test which could prove that Nero Digital is now as good or better than iTunes. Did I miss something?

i didnt do this

i proposed you to use coding tech in your next comparison. as you obviously cant test every codec on earth i meant the lc-aac codec to skip is nero, as its not better than apple

thats the main message (plus critising the "poping up like mushrooms nero encoders )

I don't understand the logic. One one side you have CT implementation, which is apparently inferior to Apple's one and similar to (older) Nero Digital. Then you have an updated Nero Digital encoder whereas CT's one stays the same. What you're proposing is to evaluate something which hasn't change (CT) over something which has changed (ND).

when i started to post i didnt know you had made a ct vs. nero comparison, i only knew that ct stumbed nero to the ground with he-aac

so that let me to think that ct is pretty good (potentially better than nero) and propose testing ct against apple

Guys, let it be. It's getting childish... Just wait for the test results. Or, Bond, do your own test.

So, now you know - and there you see, HA gets people educated

And regarding the bolded chosing of the words, this is not the first time I see a pattern - let's go to this doom9 post made by Bond:

http://forum.doom9.org/showthread.php?p=739912#post739912



Bond, I really don't know what is causing this quoted reaction (feel free to PM me) - but I usually do not think people with some kind of "mission" against product/technology are so much credible and all-in-all they usually start to damage their own reputation by writing something that is starting to lack professionalism - like this quote of yours.

I am sure there must be a way out for all of us here in the best possible way.

Look, I do think that Nero is a company probably most open to the communities (apart from the opensource devs) - communicating openly (like here) with the community, giving beta and test versions of the upcoming codecs as well as having free trials of their retail products

And, of course, listening to people skilled in the art, like Guruboolez and all other people doing listening tests for example.

I finally fail to see the reason towards your clear and obvious anti-Nero bias - IMO, this is just wrong, have ever you seen me being anti iTunes, or anti CodingTech, anti Microsoft - hell, I even have friends at some of those companies and I would personally find it completely wrong and pointless to use such words for the other products and technologies.

That said, I am kindly asking you to re-check once again the reasons why are you doing this - I am always available if there is something you think we are doing wrong, and I can for sure promise that I will listen.

So now that we *FINALLY* have completely new LC-core written from scratch, you want that in the future test this will be skipped. How surprising...
It will be interesting to get back to your message after the public listening test results are in. (I find it unlikely though that any codec can actually statistically win the test)

@JohnV - and to use codec that has been proven to be worse than Nero earlier (for starters, that encoder is based on FhG's Fast AAC encoder), and AFAIK not really updated - at least nobody actually confirmed this quality-wise.

I also fail to see the logic - but, there you go, Bond said that he didn't know of the CT vs. Nero tests in LC-AAC

Ah, finally.
I've been waiting for someone to utter that particular word : 'childish'.
I thought HA is all about objectivity, but some people just holding on to what they believe in even though they can't prove it right.

Well, things seem to be sorted out over here.
Can't wait to see the results of the 128kbps test...

After trying the test I believe this too. IMO only very experienced testers would be able to systematically differentiate the encoders. I think many of us can recognize only the low anchor with most of the samples.

Actually, this is very fortunate for the users. They can finally choose the encoder according to their practical needs and forget the quality issue. The old promises of perceivably CD quality 128 kbps digital audio are becoming true at last (at least for Joe and Jane Average).

yeah, I believe this 128kbps test will show that the next medium bitrate test should be 96kbps. Another interesting possibility would be to use only selected group of "golden ears" for higher bitrate test.
Guru's public tests are very valuable, but single person tests only show the hearing preferences and artifact estimation/rating of one person, so imo a "golden ear group tests" would be something to consider when trying to drive the development of 128kbps and higher bitrates of all codecs.

That would require some type of screening test, then. (I know it would count me out!)

That leads to the question, who chooses the listeners and to which criteria? if there were a large number of people, it is concievable that the results could be swayed by picking only 'golden ears' which highlight a detrimental element of one type of codec (not suggesting that such swaying is pre-planned, just that when selecting, it should be done carefully).

Well, two or three people who know they are roughly in the top 10% percentile here at HA.org (hmm... guru comes to mind) could select a set of say five samples with subtle artifacts that they themselves can ABX. Anyone who passes say all or 4/5 samples can be on the testing team...