IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
What is the best quality of these 2 spectrums?, Spectum analysis
Bahamut2
post Jan 2 2012, 02:51
Post #1





Group: Members
Posts: 11
Joined: 26-February 10
From: Alicante
Member No.: 78551



Hello, i think its my first post here, im not very professional but i know that the spectrum analysis tells about the quality of a song. Well i just wanted to have the best quality for my MP3 files and i downloaded a MP3 from the internet that is supposedly 320KBps (it really is as it sounds perfect), but then i wanted to encode the same file by meyself (using a FLAC source) and the last version of LAME so i could get the BEST MP3 quality at the moment.
My surprise is that both spectrums are not the same, and it even looks that the downloaded file from internet is a "bit" better, at least the "Hz" are better with that file, the programs tells me it was encoded using LAME 3.93. But my file encoded with latest lame and 320 CBR is slightly lower quality? (i cant hear the quality difference, i just want to know about the spectrum analysis). Why is this happening?

IS it supposed that latest lame has the best encoding quality right?

MP3 CBR 320kbps Downloaded from internet (it uses LAME 3.93):

(Auciadity shows between 21k Hz and 22k Hz)

MP3 CBR 320Kbps Encoded by me (Slow Encoding for best quality) from a FLAC file with 100% peaks/100% accuracy, using dbPowerAmp and LAME 3.99.3:

(Auciadity shows between 20k Hz and 21k Hz, slightly lower...)

Anybody knows wich one has the best quality?
Go to the top of the page
+Quote Post
mjb2006
post Jan 2 2012, 03:00
Post #2





Group: Members
Posts: 706
Joined: 12-May 06
From: Colorado, USA
Member No.: 30694



You've made a number of assumptions, like "spectrum analysis tells about the quality of a song" and that the last version of LAME = best quality. I doubt you can ABX these encodes, so your question is moot; until you post ABX logs proving otherwise, for you, the quality is the same.

This post has been edited by mjb2006: Jan 2 2012, 03:03
Go to the top of the page
+Quote Post
db1989
post Jan 2 2012, 03:02
Post #3





Group: Super Moderator
Posts: 5141
Joined: 23-June 06
Member No.: 32180



Please read (1) some of the many previous discussions on the invalidity of visual methods to evaluate audio quality, both in general and with specific reference to LAME and the lowpassing behaviour of its current version, e.g. this thread; and (2) #8 of our Terms of Service.
Go to the top of the page
+Quote Post
Bahamut2
post Jan 2 2012, 04:45
Post #4





Group: Members
Posts: 11
Joined: 26-February 10
From: Alicante
Member No.: 78551



Ok db1989 im sorry for not reading the rules, now that i read them i can understand this, but i read some days ago that with the spectrum analysis you could determine the quality of a song by the high frequency (i read you can detect false 320kbps mp3 with just the spectrum) so i did this test to see if the downloaded MP3 was a false 320 or a real one. By this test i can appreciate it is a real one so this concluses that both are the same quality. It is just that the last version of LAME has a different lowpassing behaviour (and i didn't knew that because im a n00b at this) so now i can understand why the spectrum is not the same and i feel better now, i jsut thought my encode had a worst quality because of this, but it seems LAME has changed so quality may be the same as the downloaded one.

Is it wrong to ask if and old version of LAME (3.93) encodes better/worst than the actual version? (3.99)
Should it be the same quality? (or even improved... because thats all about new versions)
Go to the top of the page
+Quote Post
Ron Jones
post Jan 2 2012, 07:01
Post #5





Group: Members
Posts: 412
Joined: 9-August 07
From: Los Angeles
Member No.: 46048



My advice would be to not worry about which version of LAME to use. If memory serves, there have been regressions in certain samples in the past, but this isn't what you should concern yourself with. Major point releases are safe.

3.99 is good. 3.98 is good. 3.97 is good. Etc., etc., etc.

I would also suggest that you not perform any spectrum analysis or any other visual analysis on your music. Listen to it and enjoy it smile.gif
Go to the top of the page
+Quote Post
Bahamut2
post Jan 2 2012, 07:37
Post #6





Group: Members
Posts: 11
Joined: 26-February 10
From: Alicante
Member No.: 78551



QUOTE (Ron Jones @ Jan 2 2012, 07:01) *
My advice would be to not worry about which version of LAME to use. If memory serves, there have been regressions in certain samples in the past, but this isn't what you should concern yourself with. Major point releases are safe.

3.99 is good. 3.98 is good. 3.97 is good. Etc., etc., etc.

I would also suggest that you not perform any spectrum analysis or any other visual analysis on your music. Listen to it and enjoy it smile.gif


Very useful, thanks for your reply smile.gif
Go to the top of the page
+Quote Post
slks
post Jan 3 2012, 11:57
Post #7





Group: Members
Posts: 374
Joined: 31-March 06
From: Houston, Texas
Member No.: 29046



As to which version you use, for CBR I don't think there have been many changes between 3.93 and 3.99. However, I know that VBR has undergone extensive tuning since 3.93, so if you plan on using VBR at any point, you might do better with a newer version.


--------------------
http://www.last.fm/user/sls/
Go to the top of the page
+Quote Post
lvqcl
post Jan 3 2012, 12:03
Post #8





Group: Developer
Posts: 3208
Joined: 2-December 07
Member No.: 49183



QUOTE
for CBR I don't think there have been many changes between 3.93 and 3.99.

From LAME history:

LAME 3.99 beta 0
* All encoding modes use the PSY model from new VBR code
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: 18th April 2014 - 07:09