Nero AAC-HE sounds terrible - just me?

[Richard Berg]

Even at the highest VBR setting (~80kbps stereo), there's tons of distortion plus a constant loud "whoosh" plus very bad sibilance whenever somebody talks. The only good thing about it is that lower bitrates don't sound much worse. Does anyone actually use this?

Nero's AAC-LC seems to be decent. Haven't tried Vorbis recently since MeGUI doesn't support it.

[Kurtnoise]

Yes...some others people here and there are noticed the same thing. According to Ivan Dimkovic, one of Nero Digital Audio engineer, the next version of Nero should be more powerfull for sbr.

[Caroliano]

The nero encoder don't seen as good as other HE-AAC encoders. But it is the only that suports 5.1 audio...

http://foobar2000.net/divers/tests/2005.09/AACHE/01/plots.png

Source: HE-AAC v.1 & v.2 comparison, Winamp vs Helix vs Nero Digital (http://www.hydrogenaudio.org/forums/index.php?showtopic=36868)

[Anima123]

I'd wait for the next release of he-aac implementation of Nero.

[[)370|\|470!2]

Dunno, guys, how you are doing your measurements, but i've just tested aacplus from winamp, and it sounds just awful. No idea either, how did guruboolez got his results, though, imo aac+ is only could be used for speech encoding atm.
Also nero lc aac encodes sounds way worse towards itunes. I can upl some samples, if
any1 is interested.

[Leo 69]

It is not tuned for transparency. But nothing sounds as good as AAC+v2 at 48k. It's true.

[[)370|\|470!2]

So, if it tended to use in cellphones then ok. But using such lame quality in dvdrips makes
a very poor sence to me. I've never look deep into it, but is mp3+sbr sounds as crappy as aac+? Since i remeber they've(tomson or whoelse's dev it) claimed that mp3+sbr @ 64 kbps sounds
as good as regular mp3 @128kbps... Is it really so? Is sbr algorithm can be improved or it's already
have reached its limits..?

[Raziel6969]

Hello, here are more results from well done listening test:

Results of Multiformat at 128kbps public Listening Test
http://www.rjamorim.com/test/multiformat128/plot18z.png
http://www.rjamorim.com/test/multiformat128/results.html
Winner: Vorbis aoTuV (OGG) & MPC

FINAL TEST: AAC vs Vorbis vs WMA vs MP3 ~ at 80kbps
http://foobar2000.net/divers/tests/2005.07/80/80TEST_PLOTS_06.png
http://www.hydrogenaudio.org/forums/index.php?showtopic=35438&hl=listening+test
Winner: Vorbis auTuV

Results of 64kbps Public Listening Test
http://www.rjamorim.com/test/64test/plot12.png
http://www.rjamorim.com/test/64test/results.html
Winner: AAC-HE

HE-AAC v.1 & v.2 comparison, Winamp vs Helix vs Nero Digital
http://www.hydrogenaudio.org/forums/index.php?showtopic=36868
Winner: AAC+V2 = AAC+SBR+PS. Nothing better at 48kbps. Don't use Nero Digital yet...

Bye :)

PD1: To Richard, as you can see, at 80kbps, is clear winner 'Vorbis aoTuV'.

PD2: To '[)370|\|470!2' if you are aware of the quality, as you can see at the pictures, better use Vorbis aoTuV (OGG) or MPC at 128kbps, not mp3, if you 'need mp3', use Lame to encode. mp3pro_64kbps never was as good as mp3_128kbps, only marketing.

[Rockaria]

Beside the each different listening test methodology, the encoder versions(release date) seem to matter a lot. In fact, I have a big problem understanding the last two results above.
I get pleasing sound from nero HE-AAC 5.1ch @ vbr 150~200kbps(-internet ~ -streaming quality, presumed around 64kbps for stereo), about 1/8 size of the original wave.

Among all the current codecs, flac, wma pro, real multichannel, nero aac and wavepack seem to be supporting reasonable multichannel encodings(yeah terrible vorbis-aoTuVb4 6ch even with the channel mapping using ffdshow or avisynth).
Another person's fingers crossed for enhanced nero AAC for any profile and cbr/vbr combinations.

[edit]
peasing sound->pleasing sound : more than a simple type-error :rolleyes:

[Raziel6969]

Hello Rockaria,

You're right at this point, but
the encoder versions(release date) seem to matter a lot.
But if you want to know which version is used on a listening test, follow the url bellow the pictures. The general differences at new version of the codecs here are minimal (Nero Digital has post in forums that will make important changes to his AAC+V2 encoder in the near future to better), because of that the listening test are a good guide to all interested in audio quality. If you are insecure, make your own 'public listening test'.

In fact, I have a big problem understanding the last two results above.Do you reffer to listening test 64kbps and 48kbps?
64kbps: Winner: AAC-HE
48kbps: Winner: AAC+V2: Coding Technologies (like Winamp5.1 version or CLI) ~ Helix
if not, please add detail.

I get peasing sound from nero HE-AAC 5.1ch @ vbr 150~200kbps
(-internet ~ -streaming quality, presumed around 64kbps for stereo),
about 1/8 size of the original wave.(My first language is not english) did you want to write peacing (keep silent or quiet)? do you try to describe a deffect? please add detail.

At these codecs:
flac & wavpack (wavepack) are loseless, and seem to me pointless to have audio files bigger than 'original' like AC3 in DVD, in that case i prefer maintain the AC3. If i need smaller size, i choose depending on size, 128kbps, 80kbps, 64kbps, 48kbps

at 64kbps wma pro, real multichannel Has lowest audio quality including QT AAC.

nero aac seem to be supporting reasonable multichannel encodings
(yeah terrible vorbis-aoTuVb4 6ch even with the channel mapping
using ffdshow or avisynth).
I don't understand what are you trying to tell...

Another person's fingers crossed for enhanced nero AAC
for any profile and cbr/vbr combinations. AAC is not a solution for audio encoding to all bitrate/mode, and never will be, the fight that AAC is fighting is at the lowest bitrate (under 64kbps).

Note: You can always post your audio questions at Hydrogen Audio forums (is like doom9 but to audio)
http://www.hydrogenaudio.org/

Bye

[Rockaria]

Well, I correct the mistyping. I meant 'good enough'. So I regard it as settled down. Thanks for pointing out.
If you are insecure, make your own 'public listening test'.
Do you reffer to listening test 64kbps and 48kbps?
64kbps: Winner: AAC-HE
48kbps: Winner: AAC+V2: Coding Technologies (like Winamp5.1 version or CLI) ~ Helix
if not, please add detail.

I was referring these last two results comparison, both for 64kbps.

http://www.rjamorim.com/test/64test/results.html
Winner: AAC-HE

HE-AAC v.1 & v.2 comparison, Winamp vs Helix vs Nero Digital
http://www.hydrogenaudio.org/forums...showtopic=36868

Simply check the Mp3 pro and nero HE AAC classic music tests.
3.48/3.68 -> x.xx/2.20
So the x.xx can be projected like 3.48 / 3.68 * 2.20 == 2.08
And another comparison between helix and nero he-aac :
x.xx/3.68 -> 3.40/2.20 :: x.xx = 5.68 helix quality projected to the www.rjamorim.com test!!( I think the last relative result has some problem)

And check other results(I will link it later). Where the bias comes from?
If you are not doing the test in a same condition, it's almost useless to compare the numbers directly from other test results. You need to adjust the bias(This was the problem I was referring to).

flac & wavpack (wavepack) are loseless, and seem to me pointless ...
The wavepack infact supports both lossy & loseless encoding. If you are not certain, go read the spec or do the download & encoding test.
Also, they are useful to me for perfect transcoding or faster intermediate result with less size.

I don't understand what are you trying to tell...
AAC is not a solution for audio encoding to all bitrate/mode...
If you are not interested in multi channel, it's quite normal.
The best multi channel encoding solution for me has been nero HE-AAC so far. And I am not interested in transcoding the mp3 files to other formats in favor of the size sacrificing the quality.

Your suggestion, Raziel6969?

[Raziel6969]

Hello, i'm gonna clear my thoughts a little more...

http://www.rjamorim.com/test/64test/results.html
this test is multiformat at 64kbps and is 'official' at audio community like HA. And shows that AAC-HE is the winner.

http://www.hydrogenaudio.org/forums/index.php?showtopic=36868
this test is about only AAC at 64kbps and 48kbps, and shows that
Nero implementation of AAC-HE at 64kbps is worst than AAC+V2 at 48kbps
and AAC-HE implementations of Helix & Winamp are better (as should be).
Conclusion, now is not audio quality wise use Nero AAC at 64kbps. Your right:
it's almost useless to compare the numbers directly from other test results
but the test has different direction, one is multiformat, and the other is only one format, you can't compare results between both.

To me, mp3pro is a dead format, because there are any updates years ago,
and is restricted only to a plugin. Any player can play those files, but play
only the 'mp3 standard' part that is half the bitrate and sounds annoying.
None open source decoder to this (is like VP7, but this change in the near future :P ). The support at players is BAD, and to AAC is promising. At 64kbps i'll choose AAC-HE, then Vorbis, and never mp3pro. To me, if i can't play the audio format in foobar, is useless :P .

I did use wavpack & flac only at loseless, when i've made some audio captures to compress a lossy format later like MPC. Your right wavpack has a hybrid mode (lossy+loseless) that i never saw a file in that mode nor benchmark yet. Seems to me pointless re-encode an already lossy format like AC3 (like DVD ones). Flac & Wavpack are ok to temporal files... but the discussion here is about lossy at very-low bitrate.

If you are not interested in multi channel, it's quite normal.
The best multi channel encoding solution for me has been nero HE-AAC so far.

Your right, till now, i never did an encode in multi-channel (5.1?) , Maybe Nero
is easy to use for multi-channel.

And I am not interested in transcoding the mp3 files to other formats in favor of the size sacrificing the quality.
Your wrong, if your final format is AAC-HE you ARE sacrificing the quality, that codec is only good bellow 64kbps (the quality difference is audible), above there are better options, if you are afraid of quality with a lossy format, never encode nothing bellow 160 or 192kbps with Vorbis or MPC (I recognize AAC-HE 64kbps from original easilly); and if you are 'insane' about audio quality, stay loseless.
Your right at no re-encode from lossy formats, is EVIL, mp3 => AAC-HE is BAAD, i think is called 'generation loss' and is applicable to images too.

The good news is that dimzon is working to do a tool to encode multi-channel with the AAC version of winamp :) .

The PS part of AAC+V2 is only for bellow 48kbps, don't expect nothing new above this, from Nero. They are gonna "tune" AAC encoder to better results at 64kbps.

Bye

[Rockaria]

Raziel6969, I appreciate your efforts editing in various colours! :D
As we are looking for better quality anyway, I add some more in response to your instructions.

To me, mp3pro is a dead format, because there are any updates years ago, And when mp3pro & nero he-aac looked promising, the helix was busy in the labs only.

but the test has different direction, one is multiformat, and the other is only one format, you can't compare results between both.
The comparison of any combination of result(s) also affects your decision to choose the ideal codec & encoder & tool for your each purpose.
The problem is how you interprete the results when they are all relative to some extents.

but the discussion here is about lossy at very-low bitrate. So I cannot talk about the quality? The quality is also affected by the channels reduced from the sources...

Your wrong, if your final format is AAC-HE you ARE sacrificing the quality, that codec is only good bellow 64kbps (the quality difference is audible), above there are better options, if you are afraid of quality with a lossy format, never encode nothing bellow 160 or 192kbps with Vorbis or MPC (I recognize AAC-HE 64kbps from original easilly); and if you are 'insane' about audio quality, stay loseless.
Your right at no re-encode from lossy formats, is EVIL, mp3 => AAC-HE is BAAD, i think is called 'generation loss' and is applicable to images too.
Would you elaborate some more? Why should I be 'wrong' 'insane' 'afraid' 'EVIL' and instructed for you?


The good news is that dimzon is working to do a tool to encode multi-channel with the AAC version of winamp .
That is the real part of the programming I am also interested in. Look vorbis.

The PS part of AAC+V2 is only for bellow 48kbps, don't expect nothing new above this, from Nero. They are gonna "tune" AAC encoder to better results at 64kbps.
Again, no thanks. I haven't given you any clue to learn from you this way.

[Raziel6969]

Hello

I like colors because is easier to read :)

I don't think to use mp3pro is a good idea, mainly for:
http://www.hydrogenaudio.org/forums/index.php?showtopic=20108
as posted by Garf, and Transcoding from one lossy format 2 another by picmixer
http://www.hydrogenaudio.org/forums/index.php?showtopic=34575
http://www.hydrogenaudio.org/forums/index.php?showtopic=11884
http://www.hydrogenaudio.org/forums/index.php?showtopic=33661

Note: I did find this with a simple search of mp3pro at: Hydrogranaudio http://www.hydrogenaudio.org

At: http://www.rjamorim.com/test/64test/results.html
shows that between many formats, AAC-HE performs better, later,
http://www.hydrogenaudio.org/forums/index.php?showtopic=36868
shows that between AAC-HE implementations, Nero is worst, including
than a 48kbsp AAC (AAC+V2). There are no confussion to me, the second is more specific only.

About transcode between lossy formats (this is related to 'evil' and 'bad'):
http://www.hydrogenaudio.org/forums/index.php?act=ST&f=16&t=10702
http://www.hydrogenaudio.org/forums/index.php?showtopic=32700
http://www.hydrogenaudio.org/forums/index.php?showtopic=8606
http://www.hydrogenaudio.org/forums/index.php?showtopic=32440

but the discussion here is about lossy at very-low bitrate.
I wrote this because of the title of this thread: "Nero AAC-HE sounds terrible - just me?" Just to stay focused, nothing more.

About audio quality:
Lossy codecs bellow 128kbps has clear different audible issues, more low bitrate => worst sound, then the difference between orginal and encoded at low-bitrate is easy to recognize, then, the people at listening test try to find the less annoying sounds, to you AAC-HE at 64kbps of Nero can be enough, but to many people don't (like at HA forums). There you can post about all your audio quality concerns between 'masters' :P .

At: http://www.hydrogenaudio.org/forums/index.php?showtopic=36868
Nero 7 is announced for October. I guess that the new generation of
Nero Digital HE-AAC will be released with the new Nero package and
I’m sure that Ivan had built a competitive encoder.

About 'insane':
is a quality preset at MPC audio codec, and reflects that if you don't gonna process the audio or re-encode, is pointless to decode an AC3 (is lossy) to WAV to leave a WavPack that has Bigger size and same quality as the good old AC3. If you want to reduce the size, or process (normalize, limiting, denoise, etc) is good to work loseless before to re-encode. If you have the audio recorded from soundcard is good to work loseless and if you are an audiophile, stay loseless.

About dimzon:
i see that you already posted about this in other threads...

Bye

[Rockaria]

Let me point out one big problem in this conversation.
You are full of assumptions and want to believe you are born to teach stupid lives.

For an example :
I like colors because is easier to read

I don't think to use mp3pro is a good idea, mainly for:
http://www.hydrogenaudio.org/forums...showtopic=20108
as posted by Garf, and Transcoding from one lossy format 2 another by picmixer
http://www.hydrogenaudio.org/forums...showtopic=34575
http://www.hydrogenaudio.org/forums...showtopic=11884
http://www.hydrogenaudio.org/forums...showtopic=33661

Note: I did find this with a simple search of mp3pro at: Hydrogranaudio http://www.hydrogenaudio.org

I cannot find anywhere stating myself using mp3pro nor using it a good idea. What is your problem?

I really don't understand why you are so desperately advertising yourself and hydrogenaudio.org again and again over the known facts?.
I seriously advise you to look for some genuine kids to teach in your own thread or in hydrogenaudio.org. :confused:

[[)370|\|470!2]

There's huge difference between "lossy" when the "loss" is almost virtually inaudible and "evil lossy"
or, even more precise — "sux lossy" as mp3pro and neroHE and aac+. When i'm listening to
mp3 encoded at 192 kbps, it sounds acceptable to me (and 99% of ppl will agree since there's
only 1% of mankind has an ultrasonic hearing capability xD)

Furthermore, it seems completely absurd imo to create 5.1 sound with less than 64kbps per channel...
(384kbps overall) It's same as encoding 1080p hd-dvd with 700 kbps(1 cd rip)... Sounds a bit silly, huh?

And, btw, old and "outdated"
FhG encoder creates much preferable results than lame just because it sounds way clearer @ 192 kbps.
At fewer bitrates lame performs a bit better.

[Rockaria]

Furthermore, it seems completely absurd imo to create 5.1 sound with less than 64kbps per channel...
(384kbps overall) It's same as encoding 1080p hd-dvd with 700 kbps(1 cd rip)... Sounds a bit silly, huh?
I thought like that like many normal people before trying it. Recently I tested encoding vbr he-aac @170kbps 5.1ch and 64k stereo(not mono) by downmixing from the same source. The 5.1ch was actually better than the stereo. I suppose it's because the stereo contains much more mixed signals in two channels compared to the each seperate channel simpler signal in 5.1 compressing.

Now I understand why people are encoding the 5.1 music to he-aac vbr -streaming quality. I don't say it SILLY at all.

[[)370|\|470!2]

Now I understand why people are encoding the 5.1 music to he-aac vbr -streaming quality. I don't say it SILLY at all.

Yea right. You may also buy teh highest def plasma tv available and use it to
watch vhs tapes... Im certain that some people out there doing so...

[Raziel6969]

Hello :) again,

I'm just giving URL from a site dedicated to audio from where i got the litening test (that is already mentioned at this forum in other threads), and appears to me informative, and that is the way i want to stay, and some little experience at audio encoding (not video). I never meant to makes feel bad to anyone :( .

I think like you '[)370|\|470!2'

Bye ;)

[Rockaria]

Yea, certainly there were the days believing everything mine is the best before a preacher dropped by one day and began pushing the long 'ideal life speech' and his sister in a law ceaselessly affirming by the big mouth.

I just pointed out few things :
. Be careful when interpreting the test results not to have the biased belief.
. Experience it personally before simply adopting common belief and criticizing others
. multichannel to stereo is pretty much like stereo to mono.

And the person began all the irrelevant listing of the (personally unverified) knowledge just to be loud, which is not a conversation any more.
In this internet era, interpreting the data is more important than producing and parroting, to make it a useful information, which 100% of reasonable people will agree.

[[)370|\|470!2]

Rockaria
Heheh, point taken xD
Though, i've just did a few testings by my own, and must say:
1)True, aac+ creates a bit more pleasant results over nero-he
2)5.1 channel aac+/nero-he may sound better than fe audiotape ^^
3)both cannot compete with any format at 128kbps(at least)

PS.: multichannel to stereo is pretty much like stereo to mono
Nah... Not really... Just an example: we have only 2 ears, not 6, so mono is closer to
5.1 than stereo :DDD

[Rockaria]

In fact, I am not still convinced with the current sound solution(upto 7.1) in the market.

It may cover the two dimensional sound somehow. But I don't think the current technology can capture & reproduce the three dimensional position without considering the height(up/down origin), which people with normal TWO ears can recognize.

[Rumbah]

I have to say that I agree with Rockaria about 5.1 encoding. Nero HE-AAC produces very good reults at 170kbps. But I can only speak for normal film soundtrack, so a lot of dialogue and noises with some background music. You also have to consider that one channel only has very low frequencies (subwoofer), one nearly only speech (center) and the back speakers are often only quiet atmospherical sounds (rain, wind, some birds, ...).
In addition to that you do a 5.1 test with speakers, not headphones, so some small artefacts might not be audible that would annoy you on headphones.

[[)370|\|470!2]

Sometimes when travelling i watch movies on laptop, with headset, and then, all these
"Almost inaudible artifacts" turns out very annoying.

[Rumbah]

Then you should perhaps add a 2 channel soundtrack at 128kbps :P
Or just stay at the original AC3.
For me, I don't hear arctifacts that easy (yes, for music I hear the difference between 64kbps and 128 kbps ;) ), that's probably because I'm not well trained to hear them. And I will not start to practice it, because I'm already too well trained for video artifacts. It's horrible, you watch TV and see artefact over artefacts and wonder why they choose that low bitrates ... And I don't want to hear the audio artefacts in addition to that ^^