Hey guys,

Just thought I'll drop another killer sample here for you to check it out.
I have also tested this sample on older versions of Lame as well and also tried the -m s mode but the result was always the same.
I could easily ABX this sample on a very cheap headset (10 EUR).

Here is a log for the latest version of Lame 3.98b8 using -b 320 switch:

foo_abx v1.2 report
foobar2000 v0.8.3
2008/05/20 20:29:25

File A: file://E:\abx\living on my own.mp3
File B: file://E:\abx\living on my own.wav

20:29:26 : Test started.
20:30:14 : 01/01 50.0%
20:30:24 : 02/02 25.0%
20:30:30 : 03/03 12.5%
20:30:36 : 04/04 6.3%
20:30:45 : 05/05 3.1%
20:30:51 : 06/06 1.6%
20:30:57 : 07/07 0.8%
20:31:03 : 08/08 0.4%
20:31:10 : 09/09 0.2%
20:31:16 : 10/10 0.1%
20:31:24 : 11/11 0.0%
20:31:28 : Test finished.

----------
Total: 11/11 (0.0%)

The problem is the pre-echo or smearing but you will hear it for yourself.

Do you notice the same problem with -V0 instead of -b320 too?

Yes.
I tried the -V0 and also the --preset 250 and --preset insane and the problem was the same.

So can anyone confirm my finding?
By the way, I find this sample to be just as easy to ABX as the EIG sample which I had no problem ABX-ing at 320Kbp/s as well.
I realize this is a native MP3 format limitation which can't be fixed.
Something to do with Window Size if I'm not mistaken.

Anyways, I've also tried the LAME 4.0a14 also using -b320 switch and the problem is even more obvious as there is an audible glitch right before the sample starts.
Regardless of that, the same smearing/pre-echo occurs starting at cca. 1-2s and then goes on until the hi-hats fade-in (that's where I no longer can hear pre-echo on kick-drum (or the lack of kick) because of hi-hats masking it).
The only good thing about LAME 4.0a14, though, was there was no clipping detected.
I'm aware this LAME version is dating back to 2006 and its development has stopped.

Just listened to the sample and I must say it was very hard to ABX at higher bitrates. I got v398b8 -V 0 13/15 but -b 320 was too tough for me (it might be due to fatigue after doing V2 and V0). I focussed on high hats at two points: at about 15.7 sec and 25.7 secs. I could not ABX anything at the first part of the sample.