Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: AAC 48kbps test - results (Read 33532 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

AAC 48kbps test - results

The results of the public AAC 48kbps listening test are now available:
http://www.mp3-tech.org/tests/aac_48/results.html


Nero misses winning over 3GPP by a very small margin.
Overall it is safe to say that the tested Nero encoders are quite competitive against Coding Technologies ones, and 3GPP reference code is not that bad.

edit: there was initially a small processing error in the results, which is now corrected

AAC 48kbps test - results

Reply #1
Ouch... I sincerely hope Bond will like the results

AAC 48kbps test - results

Reply #2
Come on Ivan. Don't tease bond. 
Juha Laaksonheimo

AAC 48kbps test - results

Reply #3
It seems that you substancially improved your encoder, guys.
Now, please do not forget that you still have to release it within the (too big IMHO) Nero suite.

AAC 48kbps test - results

Reply #4
As conclusion  mostly PS is useless at 48 kbits.

AAC 48kbps test - results

Reply #5
Despite the low bitrate of these samples, it seems the encoders performed quite well. I'm very impressed. It's incredible how quickly and radically the efficiency of audio compression progresses.

Congratulations to the folks at Nero for providing such an excellent product. Keep up the great work!

UED77
UED77
wavpack 4.50 -hx3; lame 3.97 -V4 --vbr-new

AAC 48kbps test - results

Reply #6
Yep...congrats to the Nero devs and Gabriel to submit this test.


Now, we waiting the new core engine guys...

AAC 48kbps test - results

Reply #7
Excelent work of Nero. Hope there still room for good engineering to improve AAC.


AAC 48kbps test - results

Reply #9
Quote
"harashin_result15.txt" is located in sample16 directory.
Is it just a mistake when hosting files?

This was a processing error. It is now corrected, but changed the results: Nero can not be said anymore to be better than 3gpp (considering 95% confidence intervals).

AAC 48kbps test - results

Reply #10
Quote
Ouch... I sincerely hope Bond will like the results
[a href="index.php?act=findpost&pid=373109"][{POST_SNAPBACK}][/a]

hehe 

remembering all your promises i would have assumed that nero would have stumbed the reference encoder into the ground

anyways i hope those results will lead to that ct's encoder will not be that widely ignored anymore on hydrogenaudio compared to nero
(in fact after i tried to get guruboolez to care more about ct and not only about nero and the resulting flame war my mod status here was removed  )

anyways, who cares, great results for both ct and nero!
I know, that I know nothing (Socrates)

AAC 48kbps test - results

Reply #11
Quote
(in fact after i tried to get guruboolez to care more about ct and not only about nero and the resulting flame war my mod status here was removed   )
[a href="index.php?act=findpost&pid=373471"][{POST_SNAPBACK}][/a]


This assertion is false and the real reason was something else entirely:

Code: [Select]
Saved Moderator Logs

 Member Name  Actions Perfomed
 Peter                    3197
 [...]
 bond                        1

AAC 48kbps test - results

Reply #12
Note to all: whineposts about doom9 moderation are NOT welcome in this thread.


AAC 48kbps test - results

Reply #14
Quote
Quote
(in fact after i tried to get guruboolez to care more about ct and not only about nero and the resulting flame war my mod status here was removed   )
[a href="index.php?act=findpost&pid=373471"][{POST_SNAPBACK}][/a]


This assertion is false and the real reason was something else entirely:

Code: [Select]
Saved Moderator Logs

 Member Name  Actions Perfomed
 Peter                    3197
 [...]
 bond                        1

[a href="index.php?act=findpost&pid=373504"][{POST_SNAPBACK}][/a]

ok if thats the real reason, i just thought it was something else because the removal of the mod status happened at the same time as the clash with the nero dev crowd (hey, noone cared to drop me a message on why i was "demoded"...)
there just isnt much to moderate in the video section 

anyways, enough said about this topic...
I know, that I know nothing (Socrates)

AAC 48kbps test - results

Reply #15
Nero are preparing release 5.0 AAC with 2 pass support.  Maybe it will be possbile to introduce the feature autoselection between HE-AAC1 or HE-AAC2 basing on stastistic information of 1st pass.

I noticed that many times when LAME encoded samples with MS approx 90-95% or higher  then  PS is better than SBR for those samples.

AAC 48kbps test - results

Reply #16
For winamp 48000 bit/s != 48 kbit/s . 1 kbit = 1024 bits  != 1000 bits

So real 48 kbit bitrate for winamp will be 48*1024 = 49152 bit/s

Winamp CT in this test was tested at 48000 bit/s = 46,875 kbits/s

I just riped 1 cd  ( 4297 sec = 1 hour 11 mins 37 secs):

1. Nero 4.9.9.5 SBR vbr 3. Size 26498997 byte . Real bitrate 48,17 kbit/s
2. Winamp CT sbr --cbr 48000 . Size 25564613. Real bitrate 46,48 kbit/s
3. Winamp CT sbr --cbr 49000 . Size 26101657. Real bitrate 47,46 kbit/s
4. Winamp CT sbr --cbr 49500 . Size 26370201 . Real bitrate 47,94 kbit/s

Also riped 20:06 min:sec ac3 soundtrack . The same result.

AAC 48kbps test - results

Reply #17
Now I'm triyng to listen difference between CT and Nero and can't.

AAC 48kbps test - results

Reply #18
Quote
For winamp 48000 bit/s != 48 kbit/s . 1 kbit = 1024 bits != 1000 bits

Sorry, but we are talking about bits here, so 1k = 1000.
48kbps is 48 000 bits per second.

Practically, I encountered winamp's encoder to give 46000bps when asked 48000bps. This might be because they are taking the mp4 container overhead into consideration.

To check it, someone has to encode files at both 48kbps and 96kbps. If the difference is 2kbps in both cases, then it probably means that they are taking container overhead into consideration. If the difference is 2kbps in one case and 4kbps in the other one, it means that they have an issue understanding the required (by the user) bitrate.

AAC 48kbps test - results

Reply #19
Just a question, where may I obtain the Nero encoder used in this test? An earlier comment leads me to believe it's not available to the public, is that right?

AAC 48kbps test - results

Reply #20
Quote
Just a question, where may I obtain the Nero encoder used in this test? An earlier comment leads me to believe it's not available to the public, is that right?

Look at the bottom of this page:
http://www.mp3-tech.org/content/?48kbps%20...20public%20test

However, I just noticed that the encoder is not there anymore. I'll restore it this evening.

AAC 48kbps test - results

Reply #21
still not there

AAC 48kbps test - results

Reply #22
Yes gabriel that would be wonderful, thanks I'll download it when it's up

AAC 48kbps test - results

Reply #23
Sorry.
Nero AAC encoder is now re-uploaded.

AAC 48kbps test - results

Reply #24
Quote
Overall it is safe to say that the tested Nero encoders are quite competitive against Coding Technologies ones, and 3GPP reference code is not that bad.


!=

Quote
All the  contenders are statistically tied


whats up with that? (not that i really understand all this stats math.)

i imagine you could also say: just use the 3gpp reference code. (or maybe not?)
PANIC: CPU 1: Cache Error (unrecoverable - dcache data) Eframe = 0x90000000208cf3b8
NOTICE - cpu 0 didn't dump TLB, may be hung