IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Comparing August and February Nero AAC Encoder release
muaddib
post Mar 3 2008, 17:26
Post #1





Group: Developer
Posts: 398
Joined: 14-October 01
Member No.: 289



There were some complaints from user on HA about August version of Nero AAC encoder quality compared to February 2007 release.
But there were not much samples provided to back this up. Namely only guruboolez provided one sample. guruboolez is unfortunately not available since beginning of this year.
Can you please help us providing more examples where it is easily noticeable that February version produces worse results than the August one. Also please point me to any thread that I missed where such sample was provided.
Please provide very short samples (5-15 seconds) where the problem is evident, together with parameters used for encoding.
This information is needed to improve next release of Nero AAC encoder, so if you are interested in improving the quality please help us.
Go to the top of the page
+Quote Post
Squeller
post Mar 3 2008, 17:33
Post #2





Group: Members
Posts: 2351
Joined: 28-August 02
Member No.: 3218



I'm glad to see you're still working on the encoder. I'm very interested in the quality of the encoder. I reported this problem, but this is not august/february related. Maybe it's already solved, according to menno in that thread.
Go to the top of the page
+Quote Post
muaddib
post Mar 3 2008, 17:40
Post #3





Group: Developer
Posts: 398
Joined: 14-October 01
Member No.: 289



QUOTE (Squeller @ Mar 3 2008, 17:33) *
I'm glad to see you're still working on the encoder. I'm very interested in the quality of the encoder. I reported this problem, but this is not august/february related. Maybe it's already solved, according to menno in that thread.

This problem is solved, but we will not release encoder until we are sure that it doesn't introduce new problems and that is why users help is needed. To help us we need only samples that introduced new problems in August release compared to February release.
Go to the top of the page
+Quote Post
eevan
post Mar 3 2008, 22:48
Post #4





Group: Members
Posts: 537
Joined: 9-April 07
From: Belgrade, Serbia
Member No.: 42357



Here is my sample.
It is 9.887s long opening of „Flos florum", a motet by Guillaume Dufay.

Files in the archive http://rapidshare.com/files/96827886/flos.rar (3152 KB):
  • flos-orig.ape
  • flos-august.mp4 (neroaacenc.exe 1.1.34.2 -q 0.45)
  • flos-february.mp4 (neroaacenc.exe 1.0.7.0 -q 0.45)
  • flos-august.png (Adobe Audition screenshots showing the problem…
  • flos-february.png …and the difference in the two builds)
I hope it's helpful!


--------------------
If age or weaknes doe prohibyte bloudletting you must use boxing
Go to the top of the page
+Quote Post
muaddib
post Mar 4 2008, 10:56
Post #5





Group: Developer
Posts: 398
Joined: 14-October 01
Member No.: 289



QUOTE (eevan @ Mar 3 2008, 22:48) *
[*]flos-august.png (Adobe Audition screenshots showing the problem…
[*]flos-february.png …and the difference in the two builds)

Thank you very much for this contribution. However graphs are not needed.
Additional information that is needed is what difference do you hear? At which position?
We will not try to make graphs look better, we will try new version of the encoder to sound better (by making it have less differences from original that could be heard).
Go to the top of the page
+Quote Post
eevan
post Mar 4 2008, 13:02
Post #6





Group: Members
Posts: 537
Joined: 9-April 07
From: Belgrade, Serbia
Member No.: 42357



Of course, I know that the graphs aren't proving anything.
There is a problem in the syllable 'flo' of the word 'florum' (6–8s).

CODE
foo_abx 1.3.1 report
foobar2000 v0.9.5.1 beta 1
2008/03/04 12:09:58

File A: D:\TEMP\Tmp\flos-february.mp4
File B: D:\TEMP\Tmp\flos-august.mp4

12:09:58 : Test started.
12:10:10 : 01/01  50.0%
12:10:40 : 02/02  25.0%
12:11:15 : 03/03  12.5%
12:12:18 : 03/04  31.3%
12:13:08 : 04/05  18.8%
12:13:44 : 05/06  10.9%
12:14:30 : 06/07  6.3%
12:15:06 : 07/08  3.5%
12:15:49 : 07/09  9.0%
12:16:05 : 08/10  5.5%
12:16:46 : 09/11  3.3%
12:17:22 : 10/12  1.9%
12:18:08 : 11/13  1.1%
12:18:39 : 11/14  2.9%
12:19:02 : 12/15  1.8%
12:19:45 : 13/16  1.1%
12:20:02 : Test finished.

----------
Total: 13/16 (1.1%)


This post has been edited by eevan: Mar 4 2008, 18:48


--------------------
If age or weaknes doe prohibyte bloudletting you must use boxing
Go to the top of the page
+Quote Post
muaddib
post Mar 6 2008, 09:59
Post #7





Group: Developer
Posts: 398
Joined: 14-October 01
Member No.: 289



Thank you eeven!
The problem in the syllable 'flo' of the word 'florum' (6–8s) definitely exists, but is very subtle. I can not hear it, but I got confirmation from colleague, that has better hearing, that there is a subtle problem in February version.
This sample is definitely not an example of "easily noticeable", but it will be taken into account before next release.

Anybody else to provide a sample?
Go to the top of the page
+Quote Post
muaddib
post Mar 25 2008, 19:16
Post #8





Group: Developer
Posts: 398
Joined: 14-October 01
Member No.: 289



Would anybody with good pair of ears (capable of hearing small impairments:)) compare Feb and Aug version on famous bibilolo sample? Maybe check background noise from 4.5s to 7.5s rolleyes.gif
Thanx in advance.
Go to the top of the page
+Quote Post
Sebastian Mares
post Mar 25 2008, 20:33
Post #9





Group: Members
Posts: 3629
Joined: 14-May 03
From: Bad Herrenalb
Member No.: 6613



I could give it a try although I am not sure if I am able to hear anything. tongue.gif

This post has been edited by Sebastian Mares: Mar 25 2008, 21:20


--------------------
http://listening-tests.hydrogenaudio.org/sebastian/
Go to the top of the page
+Quote Post
IgorC
post Mar 25 2008, 22:03
Post #10





Group: Members
Posts: 1540
Joined: 3-January 05
From: ARG/RUS
Member No.: 18803



QUOTE (muaddib @ Mar 25 2008, 15:16) *
Would anybody with good pair of ears (capable of hearing small impairments:)) compare Feb and Aug version on famous bibilolo sample? Maybe check background noise from 4.5s to 7.5s rolleyes.gif
Thanx in advance.


If you will upload some encododed samples to abx some people will help you here (including me) smile.gif
Go to the top of the page
+Quote Post
muaddib
post Mar 26 2008, 12:54
Post #11





Group: Developer
Posts: 398
Joined: 14-October 01
Member No.: 289



bibilolo sample

Original can be downloaded from http://www.mp3-tech.org/tests/aac_48/samples/sample3.zip
(There is original in flac and some compressed formats inside zip)

Here is the sample encoded at 96kbps:
February version: Attached File  bibilolo__1.0.7_.mp4 ( 350.19K ) Number of downloads: 275
August version: Attached File  bibilolo__1.1.34_.mp4 ( 346.29K ) Number of downloads: 270
Go to the top of the page
+Quote Post
Alexxander
post Mar 26 2008, 15:49
Post #12





Group: Members
Posts: 457
Joined: 15-November 04
Member No.: 18143



QUOTE (muaddib @ Mar 26 2008, 12:54) *
bibilolo sample

Original can be downloaded from http://www.mp3-tech.org/tests/aac_48/samples/sample3.zip
(There is original in flac and some compressed formats inside zip)

Here is the sample encoded at 96kbps:
February version: Attached File  bibilolo__1.0.7_.mp4 ( 350.19K ) Number of downloads: 275
August version: Attached File  bibilolo__1.1.34_.mp4 ( 346.29K ) Number of downloads: 270

I've ABX'ed both mp4 against flac with Sennheiser headphone, cheap Sony amp and analogue out of onboard ADI1988B soundchip. If one listens to a small sound fading in at around 5.9-6.0 sec ABXing is very easy. I had really no problem matching samples so I quitted after 5 and 6 times. But I can say that v1.1.34 does the job much better than v1.0.7 (less distortion of that specific upcoming sound at around 5.9 sec).

CODE
foo_abx 1.3.1 report
foobar2000 v0.9.5.1
2008/03/26 15:25:10

File A: C:\Documents and Settings\Alexxander\Escritorio\bibilolo.flac
File B: C:\Documents and Settings\Alexxander\Escritorio\bibilolo__1.0.7_.mp4

15:25:10 : Test started.
15:25:47 : 01/01  50.0%
15:25:54 : 02/02  25.0%
15:26:00 : 03/03  12.5%
15:26:06 : 04/04  6.3%
15:26:14 : 05/05  3.1%
15:26:17 : Test finished.

----------
Total: 5/5 (3.1%)


CODE
foo_abx 1.3.1 report
foobar2000 v0.9.5.1
2008/03/26 15:26:46

File A: C:\Documents and Settings\Alexxander\Escritorio\bibilolo.flac
File B: C:\Documents and Settings\Alexxander\Escritorio\bibilolo__1.1.34_.mp4

15:26:46 : Test started.
15:27:08 : 01/01  50.0%
15:27:23 : 02/02  25.0%
15:27:30 : 03/03  12.5%
15:27:51 : 04/04  6.3%
15:28:02 : 05/05  3.1%
15:28:15 : 06/06  1.6%
15:28:16 : Test finished.

----------
Total: 6/6 (1.6%)
Go to the top of the page
+Quote Post

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: 30th July 2014 - 17:35