[zhenya]

Quote:

There's a difference between talking about science and actually doing scientific research. This forum is a great place to do the former, but limitations inherent to Internet forums make the latter extremely difficult, if not impossible. Ignoring those limitations is in direct opposition to the scientific method.
That's not my attitude. My attitude is that it's pointless to use logs as a prerequisite to posting a personal conclusion, since the logs are very fast and easy to fake. They won't weed out false claims.
Of course it matters that some people fake their results. That's why these threads are entertainment and not science. That's a limitation of the format (Internet forums), and there's nothing that can easily be done.

Nobody here has claimed that we are doing scientific research, where you got that idea god only knows. 
 
The primary purpose of these tests is to check your own abilities.  Sharing the results here is done primarily as a community gesture, and to spur conversation among people of similar interest.  Again, the bar that we are choosing to set is that if you want to post an opinion, post your logs.  The rest of head-fi is open to posting any opinion you want with no limitations at all (well, so long as you stay mostly positive when discussing sponsors, but I digress).  If someone wants to fake the tests, that's up to them.  It does nothing to minimize the rest of the discussion because this isn't a contest. 
 

 

[bowei006]

guys...i can't belive this. so i got back home. booted up the comp. moved the two files to the fast drive. disabled all of windows whatevers. not that this might make much difference. and did the test again. this time i did it with hidden results cause i felt like i got rushed or perusaded the first time by it >_<
 
 
foo_abx 1.3.4 report
foobar2000 v1.1.11
2012/02/17 14:29:49
 
File A: K:\A.wav
File B: K:\B.wav
 
14:29:49 : Test started.
14:31:31 : 01/01  50.0%
14:32:00 : 01/02  75.0%
14:32:14 : 02/03  50.0%
14:32:37 : 03/04  31.3%
14:32:48 : 04/05  18.8%
14:33:13 : 05/06  10.9%
14:33:44 : 05/07  22.7%
14:34:02 : 06/08  14.5%
14:36:21 : 07/09  9.0%
14:37:05 : 08/10  5.5%
14:37:35 : 09/11  3.3%
14:37:46 : 10/12  1.9%
14:38:53 : 11/13  1.1%
14:39:33 : 11/14  2.9%
14:39:47 : 12/15  1.8%
14:40:03 : 13/16  1.1%
14:40:45 : Test finished.
 
 ---------- 
Total: 13/16 (1.1%)
 

 
i don't really care who calls me out..but let me tell you this..i don't have any photoshop or editing skills at all. and yes i run a dual monitor that i unpluged. 

 

[Jaywalk3r]

Nobody here has claimed that we are doing scientific research

So, you recognize that no meaningful conclusions can be drawn from these blind test threads, that they are good for entertainment purposes (and bragging rights) only, right?

The primary purpose of these tests is to check your own abilities.

Exactly. Which is why requiring anyone to post logs instead of just a summary of results (e.g., 17/20, p = .001) is wholly pointless. It adds no relevant information or credibility.

Again, the bar that we are choosing to set is that if you want to post an opinion, post your logs.

Again, that bar is completely meaningless. It has already been demonstrated that such logs are easily faked. I posted such a fake log, and other users chimed in with other methods that could be used to obtain fake logs that appear real.

What about people who use ABX testing applications other than FoobarABX? For example, ABXTester simply provides the percent correct. It is up to the user to calculate the p-value. Logs aren't even provided. Are the results from those users automatically invalid simply because they can't meet your bar for posting?

 

[killkli]

Well..... my vote is much the same as most of the guys'.
The 320 file is very difficult to tell from lossless file.
No good in picking the two from the sound's of instruments.
But the 320 file's vocal is a little grainy. Therefore I decided which was which.
 
 
But it's still verl small difference and doesn't effect my feeling, so 320 is good enough I think.
 
 
128k is another matter, it's pure distortion  ^^||

 

[bowei006]

Quote:

Well..... my vote is much the same as most of the guys'.
The 320 file is very difficult to tell from lossless file.
No good in picking the two from the sound's of instruments.
But the 320 file's vocal is a little grainy. Therefore I decided which was which.
 
 
But it's still verl small difference and doesn't effect my feeling, so 320 is good enough I think.
 
 
128k is another matter, it's pure distortion  ^^||

at first i felt one of the vocals were a bit grainy. but really couldn't put my tongue on it. so i just used the overal depth of the feeling to choose..and with hidden results..i performed very good. im very proud of myself
 
 

 

[silverbox]

No difference. I also converted B to 128kbps and heard no difference. All the same to my ears. Meh

 

[bowei006]

Quote:

No difference. I also converted B to 128kbps and heard no difference. All the same to my ears. Meh

that's good..in a way. the ability to not hear it sometimes..is good in a way as you know you can just listen to it all and not feel biased

 

[Mkubota1]

I'm definitely no expert in ABX or even statistical analysis.  And I completely believe all of the ABX results posted here- I don't think George Costanza is on this forum.  But I think getting good marks on just one test isn't good enough.  I could take the test once without even listening to the tracks, and get 18/20 by pure luck.  The chances are pretty slim, but it's possible- isn't it?  However, the collective results here and ones I've seen from the same members on other threads in similar tests lead me to believe that I can put part of my OCD psyche to rest if I just encode my CDs at 320k CBR, and the stuff I get from iTunes is perfectly fine and enjoyable!  Although, for complete albums I usually buy used from Amazon.
 
Edit:  I need to start hanging out in this forum a little bit more!  

 

[adamlau]

The primary purpose of these tests is to check your own abilities.  Sharing the results here is done primarily as a community gesture, and to spur conversation among people of similar interest....because this isn't a contest.
 

Exactly. If anything, consider it a contest between you and your own system. The sample was pulled from what is widely considered the best mastered and thusly, the best digital version of the song. So sit back and enjoy .
 

 

[onlyjoekin]

Mediafire doesn't seem to be working...

 

[stv014]

Quote:

Mediafire doesn't seem to be working...

They tend to remove anonymously uploaded files after a while. My DAC blind test files are still there, but I guess will not remain available for a long time.
 

 

[purk]

I tried this as well and go it right.  I was able to detect that the lossless file has more consistent instrument placement as well as more depth.  Did not take me long though, thanks to my Shure SE535.

 

[Deltaechoe]

I can't really tell, I have decently good ears but when it comes to lossless to "high quality" mp3, the differences are so small its difficult for me to tell unless i'm in a perfect listening environment (like the studio I record in)

 

[skamp]

They tend to remove anonymously uploaded files after a while. My DAC blind test files are still there, but I guess will not remain available for a long time.

Try Dropbox.

I tried this as well and go it right.  I was able to detect that the lossless file has more consistent instrument placement as well as more depth.  Did not take me long though, thanks to my Shure SE535.

I find it annoying when people who claim success don't give any details. Did you run an actual ABX test (not "A/B")? Did you set the number of tries beforehand? What was your score? What was your p value?

 

[purk]

Quote:

Try Dropbox.
I find it annoying when people who claim success don't give any details. Did you run an actual ABX test (not "A/B")? Did you set the number of tries beforehand? What was your score? What was your p value?

Then enjoy being annoy.  I'm happy with my results and even though it was not a full blown ABX test, I still manage to identify the difference b/w lossless and MP3 320KBS.  I didn't say it was easy, so stop criticize others if they are able to identify it.