IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
LAME equivalent in AAC world ?
chrizoo
post Mar 13 2010, 12:56
Post #1





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



hi everyone.
I am new to AAC and have read a lot about the technical advantages of HE-AAC v2.

No my question would be:
In the mp3 world the flagship encoder is commonly considered LAME. Is there an equivalent in the AAC world?

What is the preferred choice by the community here?
I think the most important consideration is the quality of the encoder, isn't it? So which one would you guys recommend?

many thanks.

(OS: Windows)

This post has been edited by chrizoo: Mar 13 2010, 12:58
Go to the top of the page
+Quote Post
lvqcl
post Mar 13 2010, 13:02
Post #2





Group: Developer
Posts: 3213
Joined: 2-December 07
Member No.: 49183



HE-AAC v2 is useful only for very low-bitrate files.
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 13:09
Post #3





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



doesn't HE-AAC v2 support up to 256 kbps (assuming that you don't consider this low?) ?
and why do you say HE-AAC v2 is useful only for very low bitrates (assuming that this statement holds true for the entire AAC family, not just HE-... and v2) ?

And even with your objection in mind (thank you for it BTW), my initial question still stands (just stretch it a bit).

This post has been edited by chrizoo: Mar 13 2010, 13:22
Go to the top of the page
+Quote Post
lvqcl
post Mar 13 2010, 13:22
Post #4





Group: Developer
Posts: 3213
Joined: 2-December 07
Member No.: 49183



HE-AAC v2 uses parametric stereo.

You can find something interesting in this (6 years old) thread: http://www.hydrogenaudio.org/forums/index....showtopic=18154
Go to the top of the page
+Quote Post
lvqcl
post Mar 13 2010, 13:26
Post #5





Group: Developer
Posts: 3213
Joined: 2-December 07
Member No.: 49183



QUOTE (lvqcl)
assuming that this statement holds true for the entire AAC family, not just HE-... and v2


I said "HE-AAC v2", not "entire AAC family". You cannot read?
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 13:36
Post #6





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



QUOTE (lvqcl @ Mar 13 2010, 13:22) *
HE-AAC v2 uses parametric stereo.

yes, that was one of the few things I was aware of.

QUOTE (lvqcl @ Mar 13 2010, 13:26) *
QUOTE (chrizoo)
assuming that this statement holds true for the entire AAC family, not just HE-... and v2

I said "HE-AAC v2", not "entire AAC family". You cannot read?

calm down. That's PRECISELY the reason why I was careful enough to use a hypothesis ("assuming that..."). Also note that I put forward a question, not an affirmation. From my prior readings, I thought I gathered that HE-AAC v2 was an umbrella term for standard AAC technology + additional technology (SBR+PS):



In this case, the statement would hold true for the entire AAC family, not just HE-AAC v2, wouldn't it? (hint: question!)


QUOTE
You can find something interesting in this (6 years old) thread: http://www.hydrogenaudio.org/forums/index....showtopic=18154

thanks, I will read it.
But being 6 years old, it would in no way answer my OP which concerns CURRENT implementations of AAC encoders. Encoders improve every year, see LAME.


Go to the top of the page
+Quote Post
Larson
post Mar 13 2010, 13:48
Post #7





Group: Members
Posts: 131
Joined: 27-March 09
Member No.: 68422



In AAC world there is not a "lame" equivalent,or better yet there is not a real and completely open source encoder. There used to be FAAC but that's history,many faac developers are into Nero now. We have more commercial implementations and also free. DivX is still in beta 1 since august so i wouldn't count on that for now,the two main encoders are Apple and Nero's,both high quality and pretty equal. If you're on Windows now you could also access Apple's AAC true VBR,thanks to nao,mac xld app creator (see thread about qtaacenc). A new AAC listening test is coming so that will maybe clear a bit about the "best" encoder by now but one thing for sure is that both Nero and Apple's encoder are high quality.

I do also believe that it's a shame we don't have a real open source like a Lame AAC,with true vbr settings and efficiency might be better to develop than mp3; I mean there is a lot of confusion about AAC encoding. I personally use Apple true VBR Q127 max because of its true vbr capabilities and fast decoding,at least as far as i've seen smile.gif

This post has been edited by Larson: Mar 13 2010, 13:49
Go to the top of the page
+Quote Post
lvqcl
post Mar 13 2010, 13:51
Post #8





Group: Developer
Posts: 3213
Joined: 2-December 07
Member No.: 49183



QUOTE
In this case, the statement would hold true for the entire AAC family, not just HE-AAC v2, wouldn't it? (hint: question!)

What statement? "HE-AAC v2 is useful only for very low-bitrate files"? Or "AAC is useful only for very low-bitrate files"?


QUOTE
But being 6 years old, it would in no way answer my OP which concerns CURRENT implementations of AAC encoders.

You can read there what parametric stereo is.
Go to the top of the page
+Quote Post
[JAZ]
post Mar 13 2010, 14:08
Post #9





Group: Members
Posts: 1710
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



I think i see now the reason of your confusion.


That graphic does not try to say that HE-AAC is a superset of LC-AAC and that HE-AACv2 is a superset of HE-AAC.

What it says is that to decode a HE-AAC file, you need an LC-AAC decoder to decode the part known to LC-AAC decoders, and an HE-AAC specific decoder to decode the HE-AAC specific part, and then generate the output using these two outputs.

Same applies for HE-AACv2 versus HE-AAC.



The FAQ in the wiki about AAC may clarify a few other things that you may not understand yet.
http://wiki.hydrogenaudio.org/index.php?title=AAC_FAQ
Go to the top of the page
+Quote Post
googlebot
post Mar 13 2010, 14:11
Post #10





Group: Members
Posts: 698
Joined: 6-March 10
Member No.: 78779



To cut a long story short, there is no LAME equivalent for HE-AAC (v2). The codecs have different goals. MP3 aims at transparency at sufficient bitrates, HE-AAC (v2) does not (even at very high bitrates).

There are equivalents for AAC LC, which targets transparency as MP3. The flagship encoders are Quicktime's and Nero's.
Go to the top of the page
+Quote Post
kornchild2002
post Mar 13 2010, 14:15
Post #11





Group: Members
Posts: 2043
Joined: 8-April 05
From: Cincinnati, OH
Member No.: 21277



QUOTE (chrizoo @ Mar 13 2010, 05:36) *
But being 6 years old, it would in no way answer my OP which concerns CURRENT implementations of AAC encoders. Encoders improve every year, see LAME.


Did you actually read the old thread? It contains information about parametric stereo that you could use. Being 6 years old doesn't mean anything. Uh-oh, I just read a paper from the 1930's saying that the Earth orbits around the Sun, humans breath oxygen in order to survive, and there is a remarkable discovery of Penicillin that can be used to treat infections. I guess that just doesn't hold up today since it is so old. That must mean that the Earth is the center of the universe, humans now breath helium, and Penicillin is now a major food group used to cure high pitched squealing. My point is that the age of the article doesn't matter. Someone gave you a link to the article to help you.
Go to the top of the page
+Quote Post
lvqcl
post Mar 13 2010, 14:24
Post #12





Group: Developer
Posts: 3213
Joined: 2-December 07
Member No.: 49183



If a decoder can handle HE-AAC v2 then it can also decode LC-AAC.

But HE-AAC v2 profile implies that SBR and PS are used to encode a file.

This post has been edited by lvqcl: Mar 13 2010, 14:38
Go to the top of the page
+Quote Post
[JAZ]
post Mar 13 2010, 14:45
Post #13





Group: Members
Posts: 1710
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



Adding some more info to this thread

http://www.ebu.ch/fr/technical/trev/trev_305-moser.pdf

Page 7 shows a graph at 48kbps for many formats and one for AAC between 18 and 48Kbps
Page 8 shows a graph at 160kbps multichannel (5.1) graph for he-aac, dolby and wma
Page 9 shows the quality level approximately of the three AAC modes up to 128kbps.
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 15:14
Post #14





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



QUOTE (Larson @ Mar 13 2010, 13:48) *
In AAC world there is not a "lame" equivalent,or better yet there is not a real and completely open source encoder. There used to be FAAC but that's history,many faac developers are into Nero now. We have more commercial implementations and also free. DivX is still in beta 1 since august so i wouldn't count on that for now,the two main encoders are Apple and Nero's,both high quality and pretty equal. If you're on Windows now you could also access Apple's AAC true VBR,thanks to nao,mac xld app creator (see thread about qtaacenc). A new AAC listening test is coming so that will maybe clear a bit about the "best" encoder by now but one thing for sure is that both Nero and Apple's encoder are high quality. I do also believe that it's a shame we don't have a real open source like a Lame AAC,with true vbr settings and efficiency might be better to develop than mp3; I mean there is a lot of confusion about AAC encoding. I personally use Apple true VBR Q127 max because of its true vbr capabilities and fast decoding,at least as far as i've seen smile.gif

Larson, that was exactly the kind of information I hoped to get. It's not necessarily what I hoped things would turn out to be, but this answered my questions in the most excellent way.


QUOTE (lvqcl @ Mar 13 2010, 13:51) *
QUOTE
In this case, the statement would hold true for the entire AAC family, not just HE-AAC v2, wouldn't it? (hint: question!)

What statement? "HE-AAC v2 is useful only for very low-bitrate files"? Or "AAC is useful only for very low-bitrate files"?

let's just drop the issue, this is getting out of hand. I was just wondering why you limited your assessment ("is useful only for very low bitrates") to HE-AAC v2, when - in my (apparently wrong) logic - this would automatically include all prior AAC versions, and thus the whole AAC family. If you want to clarify anything, be my guest, otherwise let's just drop the issue okay.


QUOTE
But being 6 years old, it would in no way answer my OP which concerns CURRENT implementations of AAC encoders. You can read there what parametric stereo is.

That's why I said, I will read it.
That's why I also said, that it wouldn't in any way answer my OP. Because I was looking for information about current AAC implementations. And Larson answered this perfectly. What is so difficult that you don't understand this?
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 15:15
Post #15





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



QUOTE ([JAZ] @ Mar 13 2010, 14:08) *

I think i see now the reason of your confusion.
That graphic does not try to say that HE-AAC is a superset of LC-AAC and that HE-AACv2 is a superset of HE-AAC.


really?!? now, that would mean I'm back to square 1 again sad.gif
In this case I don't understand anything anymore. Will have to read it all over again.


QUOTE (googlebot @ Mar 13 2010, 14:11) *
To cut a long story short, there is no LAME equivalent for HE-AAC (v2). The codecs have different goals.

right. I was unaware of this when writing my OP. I figured out by your answers here.

QUOTE
MP3 aims at transparency at sufficient bitrates, HE-AAC (v2) does not (even at very high bitrates).

Ok. And HE-AAC (v1) does? Or does not either?

QUOTE
[There are equivalents for AAC LC,

I'm sorry, but what do you mean by "equivalents of AAC LC" ?

QUOTE
which targets transparency as MP3. The flagship encoders are Quicktime's and Nero's.

OK, that's in a similar vein as Larson, and helpful for me. Thank you.


QUOTE (kornchild2002 @ Mar 13 2010, 14:15) *
QUOTE (chrizoo @ Mar 13 2010, 05:36) *
But being 6 years old, it would in no way answer my OP which concerns CURRENT implementations of AAC encoders. Encoders improve every year, see LAME.


Did you actually read the old thread? It contains information about parametric stereo that you could use. Being 6 years old doesn't mean anything. Earth, Universe, Penicillin, ETC. ETC.

see above
Go to the top of the page
+Quote Post
C.R.Helmrich
post Mar 13 2010, 15:36
Post #16





Group: Developer
Posts: 682
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE (chrizoo @ Mar 13 2010, 16:15) *
QUOTE
MP3 aims at transparency at sufficient bitrates, HE-AAC (v2) does not (even at very high bitrates).

Ok. And HE-AAC (v1) does? Or does not either?

Depends on who you ask. Some marketing people will say "yes", developers like me say "no, because SBR is a parametric tool. Ok, maybe a veeery high bit rates, but there you could do just as well without SBR".

Chris

This post has been edited by C.R.Helmrich: Mar 13 2010, 15:38


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
[JAZ]
post Mar 13 2010, 15:37
Post #17





Group: Members
Posts: 1710
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



QUOTE (chrizoo @ Mar 13 2010, 15:15) *
QUOTE ([JAZ] @ Mar 13 2010, 14:08) *

I think i see now the reason of your confusion.
That graphic does not try to say that HE-AAC is a superset of LC-AAC and that HE-AACv2 is a superset of HE-AAC.


really?!? now, that would mean I'm back to square 1 again sad.gif
In this case I don't understand anything anymore. Will have to read it all over again.


It is not a superset if we mean superset = everything the previous codec has, plus more.

HE-AAC and HE-AACv2 use LC-AAC to get most of the audio, but then generate a whole other part by their own.


Wrong thinking:
LC-AAC gets the pixelated image, and HE-AAC fills in the details.

Correct thinking:
LC-AAC is used for the center part of the image, and HE-AAC for the outside part.


The key point: HE-AAC (and v2) reconstructs a part of the audio which the LC-AAC encoder has not even seen.
Please, read again what Spectral Band Replication (SBR) is. (since I think you've already read about it).

This post has been edited by [JAZ]: Mar 13 2010, 15:40
Go to the top of the page
+Quote Post
C.R.Helmrich
post Mar 13 2010, 15:40
Post #18





Group: Developer
Posts: 682
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



QUOTE ([JAZ] @ Mar 13 2010, 16:37) *

Wrong thinking:
LC-AAC gets the pixelated image, and HE-AAC fills in the details.

Correct thinking:
LC-AAC is used for the center part of the image, and HE-AAC for the outside part.

Technically, your "wrong thinking" is actually the correct one. SBR reconstructs high frequencies.

Chris


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
[JAZ]
post Mar 13 2010, 15:44
Post #19





Group: Members
Posts: 1710
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



Yes, but that gives the impression that HE improves what LC does. While it is technically correct (since it expands the bandwidth), it can be interpreted in the wrong way.

That's why i've put the other example, clearly saying that LC is used for the most important part of the audio, and the SBR is used for the less important part.

LC could have been used for the less important part, but it is not in order to save bitrate.
Go to the top of the page
+Quote Post
C.R.Helmrich
post Mar 13 2010, 16:02
Post #20





Group: Developer
Posts: 682
Joined: 6-December 08
From: Erlangen Germany
Member No.: 64012



I still disagree. Why is the outer part of an image less important than the center part? If you look directly at the edges of an image, don't you want them to be artifact-free just like the center part?

SBR works on top of a downsampled "dull sounding" version of the audio to make it brighter. A downsampled image, scaled to its original size, is blurry, and SBR would "sharpen" it.

Chris


--------------------
If I don't reply to your reply, it means I agree with you.
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 16:14
Post #21





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



QUOTE (C.R.Helmrich @ Mar 13 2010, 15:36) *
Depends on who you ask. Some marketing people will say "yes", developers like me say "no, because SBR is a parametric tool.

thanks for throwing in your developer experience, C.R.Helmrich. And thank you, too, [JAZ].

QUOTE ([JAZ] @ Mar 13 2010, 15:37) *
It is not a superset if we mean superset = everything the previous codec has, plus more.

yes, from all my readings (different sources) that was what I learned. **

OK, concerning my OP, which really aimed at pratical issues (like what encoders are the community's favourites here, flagships, alternatives, etc.), has been perfectly answered by Larson and googlebot. (But I'm of course always interested in case of additional comments).

Concerning the theoretical background / underlaying technological structure etc. I really have to read all over again, there is no sense in bothering you with this because whatever knowledge I thought I had about AAC has turned into complete confusion by now. (But I'll sure read with interest if you guys continue your discussion).

** see e.g. Wikipedia:
QUOTE
High-Efficiency Advanced Audio Coding (HE-AAC) is a lossy data compression scheme for digital audio defined as a MPEG-4 Audio profile in ISO/IEC 14496-3. It is an extension of Low Complexity AAC (AAC LC) optimized for low-bitrate applications such as streaming audio. HE-AAC version 1 profile (HE-AAC v1) uses spectral band replication (SBR) to enhance the compression efficiency in the frequency domain. HE-AAC version 2 profile (HE-AAC v2) couples SBR with Parametric Stereo (PS) to enhance the compression efficiency of stereo signals. It is a standardized and improved version of the AACplus codec.


This post has been edited by chrizoo: Mar 13 2010, 16:19
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 16:34
Post #22





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



QUOTE (chrizoo @ Mar 13 2010, 15:15) *
QUOTE ([JAZ] @ Mar 13 2010, 14:08) *

I think i see now the reason of your confusion.
That graphic does not try to say that HE-AAC is a superset of LC-AAC and that HE-AACv2 is a superset of HE-AAC.


really?!? now, that would mean I'm back to square 1 again sad.gif
In this case I don't understand anything anymore. Will have to read it all over again.



quoting from the .pdf kindly provided by [JAZ] here:

QUOTE ([JAZ] @ Mar 13 2010, 14:45) *

adding some more info to this thread
http://www.ebu.ch/fr/technical/trev/trev_305-moser.pdf


QUOTE (trev_305-moser.pdf)
Thus, HE-AAC v2 is a superset rather than a substitute for the AAC core codec
and extends the reach of high-quality MPEG-4 audio to much lower bitrates.


This post has been edited by chrizoo: Mar 13 2010, 16:35
Go to the top of the page
+Quote Post
saratoga
post Mar 13 2010, 17:17
Post #23





Group: Members
Posts: 4718
Joined: 2-September 02
Member No.: 3264



QUOTE (chrizoo @ Mar 13 2010, 10:14) *
** see e.g. Wikipedia:
QUOTE
High-Efficiency Advanced Audio Coding (HE-AAC) is a lossy data compression scheme for digital audio defined as a MPEG-4 Audio profile in ISO/IEC 14496-3. It is an extension of Low Complexity AAC (AAC LC) optimized for low-bitrate applications such as streaming audio.



That seems pretty clear to me.

QUOTE (chrizoo @ Mar 13 2010, 10:14) *
Thus, HE-AAC v2 is a superset rather than a substitute for the AAC core codec
and extends the reach of high-quality MPEG-4 audio to much lower bitrates.


Again that seems crystal clear. It adds some additional features to AAC-LC that enable better performance at very low bitrates. Hence it is a superset of AAC-LC that works much better at low bitrates.

Which part of that exactly is confusing?
Go to the top of the page
+Quote Post
chrizoo
post Mar 13 2010, 17:22
Post #24





Group: Members
Posts: 345
Joined: 25-March 08
Member No.: 52274



In posting #22 and #21 I highlighted in red the parts that made me - erroneously according to [JAZ] - fall for the "superset" explanation and I highlighted and posted them because I felt they might contradict [JAZ] when saying "That graphic does not try to say that HE-AAC is a superset of LC-AAC and that HE-AACv2 is a superset of HE-AAC.". That's all.

But again,
QUOTE (chrizoo @ Mar 13 2010, 16:14) *
Concerning the theoretical background / underlaying technological structure etc. I really have to read all over again, there is no sense in bothering you with this because whatever knowledge I thought I had about AAC has turned into complete confusion by now. (But I'll sure read with interest if you guys continue your discussion).
Go to the top of the page
+Quote Post
halb27
post Mar 13 2010, 17:54
Post #25





Group: Members
Posts: 2414
Joined: 9-October 05
From: Dormagen, Germany
Member No.: 25015



chrizoo, you're about to swim in an ocean of information in this and other threads, and you try to understand every single word and want to figure out what it exactly means resp. argue about it no matter how little importance it has (see your last posts in your joint stereo thread). Why do you bother so much? As far as I can see you get answers for all your questions that would be sufficient to 99% of HA members I guess.

Just a suggestion:
Find out what's most important to you at the very moment, and why you want to know about it.
If it's thirst for knowledge about AAC (or whatever), search the net, try to understand what you'll find, and after some days of own deep studies come back with those questions that can complete your understanding here.
If it's about audio codec practicing, you don't need to know a lot, just choose from the suggestions you get. Here you just need to know: for usual AAC usage take the LC profile, if you have to use a bitrate way below 100 kbps use HE-AAC, and if you have to use a bitrate below say 50 kbps use HE-AACv2. Don't bother too much about the exact limits - there's no science about it, it's more of a matter of taste.
Try not to bother too much in general. Life's a lot easier.

This post has been edited by halb27: Mar 13 2010, 17:56


--------------------
lame3100m --bCVBR 300
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 20th April 2014 - 14:50