BLIND TEST: 128kbps mp3 vs Lossless
Feb 6, 2012 at 11:06 AM Post #46 of 180

stv014

Headphoneus Supremus
Joined
Jul 17, 2011
Posts
3,493
Likes
271
Quote:
Originally Posted by milosz /img/forum/go_quote.gif
 
1. Absolute Sound recently published a series of articles that claimed there would be clearly audible differences, and in fact the unmolested WAV file would sound better, even though they were bit-identical...

 
I assume I guess right that it was a sighted test ?
 
 
Feb 6, 2012 at 4:41 PM Post #47 of 180

borizm

Head-Fier
Joined
Dec 19, 2011
Posts
98
Likes
18
Location
Gdańsk, Poland
I have done such a test before - it was obvious difference, so I have swap to FLAC ~6 yeas ago (I have swap to FLAC from CD, not from mp3, or ogg).
Even without both format to compare, 128kbps mp3 sounds bad for me. Of course filtering algoritm of mp3 may sometimes improves bad recorded CD :wink:
 
Feb 6, 2012 at 7:50 PM Post #48 of 180

juman231

100+ Head-Fier
Joined
Dec 11, 2010
Posts
488
Likes
18
Took me a while to hear any difference, but I thought the guitar in the beginning sounded more articulate and louder on B. But I don't think my hearing is acute enough for me to worry about getting my music in flac... maybe just 320kbps is enough for me!
 
 
 
Feb 6, 2012 at 8:57 PM Post #49 of 180

MohawkUS

1000+ Head-Fier
Joined
Jul 13, 2010
Posts
1,036
Likes
52
Location
Connecticut
It was obvious to me even with my $50 portable headphones. A sounds a little compressed and the vocals reverb slightly, while B sounds more natural and clear.
I listen to a wide range of stuff that isn't always available above 128kb, and A exhibited the exact same shortcomings I find in my other lossy music.
 
Feb 6, 2012 at 9:30 PM Post #50 of 180
Joined
Jan 30, 2011
Posts
13,315
Likes
24,307
I'm guessing I just have tin-ears ....... or maybe I should say normal ears (as my wife who has excellent hearing couldn't hear a difference either).
 
I wonder how many people who state the difference as "being obvious" would actually ABX the two tracks in Foobar (with replay gain applied to equalise volume), and post the results.  Lets say 20-25 tests ..... should be enough, and wouldn't take too long as obviously to the majority it is "really obvious" (the difference).  I do note that the only one in this thread who posted an actual test result was Deadly Lover (post #11) - and I know his ears are really good.  You'd need to post an actual screenshot of the ABX results like Deadly did.
 
I'd just be curious to see how many would actually do the test and post a result to back up the 'omg it was easy' claims .......
 
BTW - OP - the track you used was obviously designed to make it difficult - heavily compressed.  Would like to see something else well recorded with enough dynamic range to show a contrast.
 
Feb 6, 2012 at 10:46 PM Post #51 of 180

Jupiterknight

Headphoneus Supremus
Joined
Aug 17, 2009
Posts
1,989
Likes
160
Location
USA/OH


Quote:
I'm guessing I just have tin-ears ....... or maybe I should say normal ears (as my wife who has excellent hearing couldn't hear a difference either).
 
I wonder how many people who state the difference as "being obvious" would actually ABX the two tracks in Foobar (with replay gain applied to equalise volume), and post the results.  Lets say 20-25 tests ..... should be enough, and wouldn't take too long as obviously to the majority it is "really obvious" (the difference).  I do note that the only one in this thread who posted an actual test result was Deadly Lover (post #11) - and I know his ears are really good.  You'd need to post an actual screenshot of the ABX results like Deadly did.
 
I'd just be curious to see how many would actually do the test and post a result to back up the 'omg it was easy' claims .......
 
BTW - OP - the track you used was obviously designed to make it difficult - heavily compressed.  Would like to see something else well recorded with enough dynamic range to show a contrast.

 
Agree, this could be a great way to set up a new test, only allowing votes to count if posting screenshot results from ABX in Foobar. I believe the result of such a test could be different and slightly more reliable.
 
 
 
Feb 7, 2012 at 6:13 AM Post #53 of 180

xnor

Headphoneus Supremus
Joined
May 28, 2009
Posts
4,092
Likes
224
I wonder how many people who state the difference as "being obvious" would actually ABX the two tracks in Foobar (with replay gain applied to equalise volume), and post the results.  Lets say 20-25 tests ..... should be enough, and wouldn't take too long as obviously to the majority it is "really obvious" (the difference). 




Code:
foo_abx 1.3.4 report
foobar2000 v1.1.11
2012/02/07 12:09:21

File A: D:\temp\A.wav
File B: D:\temp\B.wav

12:09:21 : Test started.
12:09:43 : 01/01 50.0%
12:09:50 : 02/02 25.0%
12:09:55 : 03/03 12.5%
12:10:00 : 04/04 6.3%
12:10:11 : 05/05 3.1%
12:10:15 : 06/06 1.6%
12:10:18 : 07/07 0.8%
12:10:22 : 08/08 0.4%
12:10:27 : 09/09 0.2%
12:10:33 : 10/10 0.1%
12:10:39 : 11/11 0.0%
12:10:44 : 12/12 0.0%
12:10:48 : 13/13 0.0%
12:10:56 : 14/14 0.0%
12:11:03 : 15/15 0.0%
12:11:10 : 16/16 0.0%
12:11:28 : 17/17 0.0%
12:11:35 : 18/18 0.0%
12:11:43 : 19/19 0.0%
12:11:50 : 20/20 0.0%
12:12:19 : Test finished.

 ---------- 
Total: 20/20 (0.0%)
 
Feb 7, 2012 at 6:33 AM Post #54 of 180
Joined
Jan 30, 2011
Posts
13,315
Likes
24,307
Nice - thanks.  Now maybe some of the others .....
 
I'll even have another go in the weekend - and post a result (probably a fail) 
wink.gif

 
Feb 7, 2012 at 7:42 AM Post #55 of 180

xnor

Headphoneus Supremus
Joined
May 28, 2009
Posts
4,092
Likes
224
Btw, it's not "really obvious" unless you know what to listen for. Give it some time and try to compare single instruments, preferably high frequency stuff. Also pay attention to where the instruments are placed / where the sound is coming from. Once you got the hang of it it shouldn't be hard to get 10/10 or maybe even 20/20. Good luck. :wink:
 
Feb 7, 2012 at 8:14 AM Post #56 of 180

RPGWiZaRD

Headphoneus Supremus
Joined
Jan 31, 2010
Posts
6,348
Likes
385
How do you enable it, put it my components folder but doesn't show up in dsp manager.
 
Hate it when it seems impossible to get some documentation on some of the foobar components.
 
EDIT: Nevermind got it, first install from options, then select the two tracks and right click and go from the "utilities" menu.
 
Feb 7, 2012 at 8:37 AM Post #57 of 180

Naim.F.C

1000+ Head-Fier
Joined
May 20, 2008
Posts
1,146
Likes
276
Imo this test very much depends on the recording. A lot of my old collection was 128kbps and every now and again one would sneak in. Without even checking for it, sometimes I'd be listening to my music and thinking, why does this not sound very good, or some ways from perfect, and sure enough most times it would be 128kbps. Other times it's just a bad recording. But even at 128kbps, a good recording can still sound commendable. I don't really know what constitutes it, but from listening, I think 128kbps files have a bit more grain, noise, shrill etc. Like a very slight echo or reverb like presence. They don't sound as smooth or natural as 320kbps, but the differences are quite subtle.
 
Then again, you get some 128kbps files that sound surprisingly good. I guess certain sonic attributes just favour poorer recording qualities better than others.
 
Feb 7, 2012 at 10:55 AM Post #58 of 180

Bill-P

Headphoneus Supremus
Joined
Jan 24, 2012
Posts
2,112
Likes
343
I think I can hear the difference quite easily with my MacBook Pro (Fidelia) -> M50S LE. The cymbal was obviously smoother and more natural in track B. The same cymbal sounded grainy in track A, with the echo sounding very distorted. It sounded more like noise than cymbal.
 
Using Fiio E7 as DAC cleaned the sound up further, and then the violin also sounded more distorted in track A.
 
Sound stage was also different.
 
Feb 7, 2012 at 6:59 PM Post #59 of 180

gamereric22

New Head-Fier
Joined
Feb 5, 2012
Posts
11
Likes
0
Location
Canada, Alberta
For me i dont have a high end set up yet, but i was still able to tell. My setup currently is entry level at best...
ASUS G74SX Notebook (w/ EAX Advanced HD 5.0 & THX TruStudio Technologies) > Munitio SITi IEM

A=MP3
B= Lossless

What gave it away for me was the guitar and bass guitar was less distorted in B.

I should get my Shure SE535s and FIIO E7 Amp in a day or two so i will post back if i see a significant difference between test files with that setup.

Thanks for the test files!
 
Feb 7, 2012 at 8:09 PM Post #60 of 180

FieldEffect

100+ Head-Fier
Joined
Feb 3, 2012
Posts
134
Likes
10
The most obvious part of the track for me was in the first few seconds. It becomes a little harder to distinguish afterwards (at least for my ears).
 
It depends a lot on the particular music as to whether I will hear much of a difference. If the sample was Norah Jones or Astrud Gilberto, I'd be able to tell in an instant. But with busy rock music, the difference isn't so prominent for me.
 

Users who are viewing this thread

Top