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: EAQUAL results for OGG/Psytel/MPP/Lame (Read 2397 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

EAQUAL results for OGG/Psytel/MPP/Lame

OGG RC3
Bitrate   Setting      ODG
251   -q 7      +0.1060
202   -q 6      +0.0280
163   -q 5      -0.3053
132   -q 4.99      -0.5257

AAC 2.01 Pystel Build 04/01/2002
Bitrate   Setting      ODG
179   -normal      -0.2460

AAC 2.01 Psytel Build 07/01/2002
Bitrate   Setting      ODG
204   -archive      -0.1421
184   -extreme      -0.2553
163   -normal      -0.3100
137   -streaming   -0.4982

MPP 0.90o
Bitrate   Setting      ODG
236   -insane      -0.0716
202   -xtreme      -0.1187
170   -standard      -0.3667

Lame   3.91   
Bitrate   Setting         ODG
252   --r3mix         +0.0223
250   --alt-preset standard   +0.0748
241   --alt-preset fast standard   +0.0364
201   --alt-preset 200      -0.0730
125   --alt-preset 128      -0.7869

These are results from testing with a full 3:56 song Replica, which I often use to compare encoders.
A new build of Pystel was released as I was testing so I preformed all test with the newer version, I also included one of the older results for comparison.

The results show Lame beating AAC and MPP in the 200 kbit/s range, and OGG preforming considerably better than all the others. From my own oppinion listening to the files I would agree with the results.
Lame and OGG leave the tone of the song very much intact, OGG also has no audible artifacts, Lame has some slight artifacs.
MPC is fairly solid but it lacks warmth too the sound, sounding somewhat flat.
AAC sounds too bright and gives an unatural tone, sounding the least close to the original.

This particular song is a rather difficult sample, even lossless compression performs poorly reducing the file size to only 75% of its original size.

EAQUAL results for OGG/Psytel/MPP/Lame

Reply #1
Quote
Originally posted by Alex

Lame and OGG leave the tone of the song very much intact, OGG also has no audible artifacts, Lame has some slight artifacs.
MPC is fairly solid but it lacks warmth too the sound, sounding somewhat flat.
AAC sounds too bright and gives an unatural tone, sounding the least close to the original.


I would totally agree with your conclusion.

How do you think wether it is worth the bitrate jump at -q 5 compared with -q 4.99 with Ogg encoder RC3. Any audible quality difference or artifacts?

BTW, I strongly recommend that the lossy channel coupling feature be made optional even with high -q values because it seems that lossy channel coupling save bitrate a lot. I believe that's also one of Dibrom's phylosophy on LAME --alt-preset standard.

EAQUAL results for OGG/Psytel/MPP/Lame

Reply #2
Please read the:

http://www.hydrogenaudio.org/forums/showth...st&threadid=697

- EAQUAL can't be used as fully objective measure of perceived audio quality, as those two samples show how it misses some important events.

Only valid way of testing is double-blind listening tests, it would be great if someone is willing to organize them for MPC, Ogg, MP3 and AAC at variable bitrates.

EAQUAL results for OGG/Psytel/MPP/Lame

Reply #3
Quote
Originally posted by Ivan Dimkovic
Only valid way of testing is double-blind listening tests, it would be great if someone is willing to organize them for MPC, Ogg, MP3 and AAC at variable bitrates.


Hope there's a test soon.

But how to show a fareplay with each codec as they all are VBR files? If with similar averaged bitrates, Then what if the use of lossy channel coupling at high -q values of Ogg is showed not do harm to quality on the whole? In that case I believe Ogg win the champion.

 

EAQUAL results for OGG/Psytel/MPP/Lame

Reply #4
I find there is very little audible diffrence at -q 4.99 compared to -q 5, I find -q 5 sounds a little better but it doesnt appear to have anything to do with the change too lossless coupling.

The EAQUAL results show a diffrence between the two settings I would agree EAQUAL isnt a totally objective measurement, I just gave it a try and the results happen to correlate very closely too what my ears have been telling me for sometime.

I would also like to see a seperate setting to control channel coupling, It seems quite likely that it could offer some additional bitrate to quality improvement.
I think that with lossy coupling on -q 6 which produces files of average 200 kbit/s that could perhaps go down to around 170 kbit/s and sound just as good(Would have to try to find out).

I have been looking at the OGG source to try and enable lossy coupling at higher bit rates, but i have yet to locate the setting in the source. If I have any luck Ill be sure to post my results.

At the moment I tend to use -q 6 encoding from CD, I find at this quality I cannot here a diffrence between the original and OGG.

 
SimplePortal 1.0.0 RC1 © 2008-2021