Do 320kbps mp3 files really sound better? Take the test!

[Ra97oR]

Being unfamilar to the track and the headphones makes it more fun, but it wasn't too hard. Everything seems to hits with more energy and definition on the 320.

 

[fatcat28037]

I got it right, I keyed in on the cymbal. On the loudness test I was 1 for 3.

 

[Jespiir]

I got it right after listening to both for a couple of times. Some details where less "detailed" on the 128kbps...
 
Not the best test track though....

 

[WrxSTI]

Couple of days ago I converted a FLAC song I had for fun to 320 kbit mp3 then compared it to the FLAC version and thought "what, this is so easy, the MP3 sounds WAY worse, I can double blind this 100% right now" - and of course when I loaded this up in foobar's ABX plugin I had absolutely no idea which was which anymore even though I used the same section of the song ...

 
So I wanted to see what was the highest bitrate I could distinguish from FLAC and here are my results for 225 kbit mp3:
 
foo_abx 1.3.4 report
foobar2000 v1.1.6
2011/06/06 20:29:47

File A: C:\Documents and Settings\(username)\Desktop\5. LFObia.ape
File B: C:\Documents and Settings\(username)\Desktop\LFObia.mp3

20:29:47 : Test started.
20:30:45 : 01/01  50.0%
20:31:10 : 02/02  25.0%
20:31:45 : 03/03  12.5%
20:32:40 : 04/04  6.3%
20:33:15 : 05/05  3.1%
20:33:37 : 05/06  10.9%
20:34:22 : 06/07  6.3%
20:35:44 : 07/08  3.5%
20:36:13 : 08/09  2.0%
20:36:35 : 09/10  1.1%
20:36:53 : 10/11  0.6%
20:37:20 : 10/12  1.9%
20:37:34 : 11/13  1.1%
20:38:05 : 11/14  2.9%
20:38:27 : 12/15  1.8%
20:38:49 : 13/16  1.1%
20:39:11 : 14/17  0.6%
20:39:33 : 15/18  0.4%
20:40:05 : 15/19  1.0%
20:40:29 : 16/20  0.6%
20:41:07 : 17/21  0.4%
20:41:30 : 18/22  0.2%
20:41:58 : 19/23  0.1%
20:42:39 : 20/24  0.1%
20:43:29 : 20/25  0.2%
20:44:03 : 21/26  0.1%
20:44:47 : 22/27  0.1%
20:45:10 : 22/28  0.2%
20:45:44 : 22/29  0.4%
20:46:31 : 22/30  0.8%
20:47:15 : 22/31  1.5%
20:47:46 : 22/32  2.5%
20:48:20 : 23/33  1.8%
20:48:33 : 23/34  2.9%
20:49:02 : 23/35  4.5%
20:49:27 : 24/36  3.3%
20:49:43 : 25/37  2.4%
20:50:12 : 26/38  1.7%
20:50:36 : 26/39  2.7%
20:51:09 : 27/40  1.9%
20:51:49 : 27/41  3.0%
20:53:04 : 28/42  2.2%
20:54:49 : 28/43  3.3%
20:55:43 : 28/44  4.8%
20:57:49 : 29/45  3.6%
20:58:50 : 29/46  5.2%
20:59:55 : 30/47  3.9%
21:00:48 : 31/48  3.0%
21:01:21 : 32/49  2.2%
21:02:06 : 33/50  1.6%
21:02:09 : Test finished.

 ----------
Total: 33/50 (1.6%)
 
 
This was quite hard and took a really long time...I wouldn't really bother with trying any higher than that because I'm certain I could not do it.

 

[kalston]

Quote:

Listened to them about 3 times each, and I picked #2 . Listening again, the major difference I can tell, and would appreciate to have, is the 'spacing' around the vocals in #1, they are more clearly defined among the other instruments than in #2! They both sounded just great though, more than I can say for some of the 128k mp3s I have on my computer.
 
I find the main reason why 128k has gotten a bad rep is the audio source. I still have a few 128k AAC files from itunes, and they sound really good, better than higher bitrate encodes acquired from shady sources on the net
 
Hotaudio usb > Hotaudio Thunderbolt max > AKG K702
 
 

I'd say that 128k from LAME doesn't sound too bad

 

[ostewart]

i got it right, and that was with a crappy computer and some AKG K518DJ so yes there is a difference, but not huge.

 

[quintron]

Yes, 320kbps mp3 files do really sound better to my ears. May be that if you could listen the tracks using wasapi you'd be able to hear a larger difference.

 

[hakai]

Quote:

Currently at -33db, this is pretty hard. 

I failed the initial test posted in this thread, then I took the second one linked.
 
I'm not sure what the options mean (the ones before you take the test), or if they affect it at all, but I left it on the default (6" driver, Music T.Chapman), and selected 'headset' and 'no'.
 
I got all the way down to -45db before I got one wrong. On the second attempt at -45db I was incorrect, then it jumped me back to -36db, which I got incorrect as well. It then jumped me back to -27db, which I got correct, followed by -30db, -33db and -36db; all of which I got correct, at which point the test ended.
 
I've been reading these forums for a couple of weeks now, and this is my first time posting; so it's safe to say I don't really know what any of this actually means, but I thought I'd share .
 

 

[cuad]

I expect to fail this test cause I've been listening to Pandora and I think it sounds pretty nice even though it's supposed to be 128kbps or even less!
 
results: yeah they sound exactly the same to me. luls.

 

[LizardKing1]

Hey, I had taken the test once before, in my house, and got it right with just 1 listen of each sample.
 
Yesterday I was studying my girlfriend was going to meet me where I was, and when she arrived I had her try the test out of curiosity with the same setup I had used before (RE-Zero --> Fiio E7). She listened twice to each and guessed correctly. Nice =)
 
Then I went to listen again, and surprisingly I couldn't see so many differences. The track I knew to be 320 did sound better, slightly more detailed, but the difference was pretty redundant. If I had to take the test again without knowing the result, I would have probably guessed right again. But that I-have-no-doubt feeling of before was definitely not there, it was more of "I guess it's...".
 
Maybe there was a bit of noise around me, I need to try this again, but I was thinking, maybe my laptop isn't enough to power the RE-Zero? I know it's dumb, but it was the first thing I could think of, that instead of my headphones, I might be having trouble driving my DAC/amp, since before I had used it plugged into my desktop computer, which has a much better power source. Anyway, is this possible, or did i just have a bad moment?

 

[Quote]

I could tell the difference by the hi-hats mostly, but I also noted that the bass was clearer on the first one.

And this was just with M50s. :B

 

[zeroibis]

Got it only listening to them once. One of the things the algorithm will dump out is any static in the track. The second one appears to sound better becuase the static is less clear. This confuses people into thinking the quality is higher when in fact the higher quality one will preserve more of the static in the track. As the first clearly has more audible static it must be the one that is of the higher quality encoding. It is for this same reason that this test is flawed becuase most people will naturally think that when they hear less static==good thus skewing there judgement.

 

[Alghazanth]

The track is so slow that the extra bitrate matters little. I actually failed the "test". Then I went and converted one of my FLAC tracks to 320 and 128 kbps with latest version of LAME in Dbpoweramp. The difference was VERY noticeable.

 

[Alghazanth]

Quote:

The difference in dynamic range is quite noticeable.. this test is nothing compared to this

I got -45db, what do I win?

 

[drez]

Its fairly close but if you critically listen to each individual instrument there are definite differences.  texture on the guitar was one major giveaway, as well as extension and clean impact of the drums being more consistent in the 320.  The attack and decay of the vocals, as well as the bass guitar, are also much clearer on the 320 kbps.  Maraccas sound much more scratchy and dithery on the 128 compared to the 320.
What makes it difficult is that these differences don't seem to be constant - at certain points some of these differences are less obvious that at others - overall though if you listen for consistency of the music this is also revealing of which is which - the 128 is less consistent in its convincing reproduction - at times it sounds a bit botched in one or other instruments.