128 vs 320 kbps Test - post pics!
May 24, 2014 at 1:46 AM Thread Starter Post #1 of 13

Pandaclocker

New Head-Fier
Joined
May 18, 2013
Posts
47
Likes
23
On this cool site:  http://www.mp3ornot.com/  you can take a 100% random test that will show if you can hear the difference between 128 and 320 kpbs audio or not.

There are 3 different music samples, and they will then repeat if you continue. It's however totally random each time whether clip X is the 320 or 128 kbps one.

I would say that if you can get over 90% correct in a sample of 12 tests in a row (means at least 11/12 correct), then one can conclude that you definitely can hear a sonic difference. If you get much lower than 90% correct in a sample of 12 tests (thus less than 11/12 correct), one can probably conclude that you can't reliably hear any difference.
 
To make it interesting I therefore suggest everyone interested to take 12 tests in a row and then post their result as a picture here in this thread.

I'll start: 7/12 correct (58%)



Yep, I can't hear any clear difference, mostly guessing from my side.

Excited to see if someone can get 12/12 (or 11/12). 

Remember to take your time and listen carefully to both listed samples in step 1 before actually listening (step 2) and guessing (step 3) on "clip X".
 
May 24, 2014 at 2:27 AM Post #2 of 13
I had a hard time with the female singer with just the guitar... both recordings have artifacts in them the 128kbs was just slightly worse... might have fared better using my HD800's but tonight was a Grado night.
 
 

 
May 24, 2014 at 2:33 AM Post #3 of 13
  I had a hard time with the female singer with just the guitar... both recordings have artifacts in them the 128kbs was just slightly worse... might have fared better using my HD800's but tonight was a Grado night.
 
 


Impressive enough, you can clearly hear the difference.  Yes it's probably somewhat easier to hear the difference when using more open and airy headphones (like the HD800).
 
May 24, 2014 at 2:38 AM Post #4 of 13
 
Impressive enough, you can clearly hear the difference.  Yes it's probably somewhat easier to hear the difference when using more open and airy headphones (like the HD800).

I was impressed with myself... I figured that PS500+Glenn OTL amp would hide some little details. So far I've been very impressed with this amp, its warm and dynamic, but very detailed. Kinda feeling I need to move on from my AudioGD 11.32 to eek out that last bit of audio information, and finally have a balanced amp/DAC combo.
 
Heck it even kinda says something about the amount of detail.information I got out of my Grado's!
 
May 24, 2014 at 3:11 AM Post #5 of 13
It's interesting to see how other people did based on the Facebook comments. Usually ABX tests are done with at least 10 trials and 9 are required to be correct in order to say you can tell a difference. A lot of the comments have less than 8 trials. XD
This test isn't totally accurate since the X song often didn't play for me, even when using no add-ons on my internet browser, and it remembered its playback location unlike the A and B files.

I think I took this test before or something because I've heard the same 3 test songs from a similar test.
 
Dec 30, 2014 at 3:35 PM Post #6 of 13
Definitely not a test to try with my laptop's built-in speakers :)
 
Dec 30, 2014 at 3:42 PM Post #7 of 13
I should try this with my emotivas! See if my near field setup is up to snuff!
 
Dec 30, 2014 at 5:08 PM Post #8 of 13
Song 1: "searching for all my dreams to realize"; sounds metallic/distorted on 128 kbps compared to 320 kbps
Song 2 Instrumental that starts with drums and guitar: on 320 kbps the drums at the very start sound more lively (more high frequency presence and arguably better reverb) than the 128 kbps version.
Song 3: "I just can't find a loving man"; on 128 kbps it seems to me that "just" is pronounced different than on 320 kbps; exremely minor difference to me, I can barely tell after repeated (AXABXB)^n listening.
Edited image for 12 tests, didn't read OP properly.

 
 
All songs have very minor differences to me, but that instrumental song is the most tricky for me to pick apart.
Anyhow, I would never ever listen to music like this. I doubt I can tell apart 320 kbps from 192 kbps at all. So I don't think going to lossless formats is for me, I'll just stick with >192 kbps MP3/AAC/OGG.
 
Dec 30, 2014 at 5:57 PM Post #9 of 13
6 right out of 12. :cry:
 
Jan 3, 2015 at 6:49 PM Post #10 of 13
These tests between digital formats are pointless. Brickwall-limit and apply makeup gain to the 128kb version or smiley-face EQ the 320kb if you want to hear a 'difference'.
 
Mar 6, 2015 at 7:13 AM Post #11 of 13
Why do you need to get 90%?
What's the rationale behind the 90% figure?
 
Mar 6, 2015 at 9:21 AM Post #12 of 13
Why do you need to get 90%?
What's the rationale behind the 90% figure?

 
Some people are claiming that the difference between FLAC and 320kbps, for example, is as plain as day and night. If that were the case, then they should have no problem in achieving at least 90%.
 
It is also statistically less likely to be chance to achieve such a high score with such a low number of tests. If you want to set a pass mark of 80% with 10 tests, then the chance of someone achieving that just by chance is around 10%. Quite a few head-fiers would then claim to hear the difference, even if they were guessing every time. If the pass mark is 90%, then the odds of achieving it with 10 guess is around 1%. Significantly lower but we could still get one or two guessing their way to a pass.
 
I'd prefer 90% with at least 20 tests. Then I would agree that it is very unlikely the subject was not guesssing.
 
Mar 11, 2015 at 7:52 PM Post #13 of 13
Some people are claiming that the difference between FLAC and 320kbps, for example, is as plain as day and night. If that were the case, then they should have no problem in achieving at least 90%.

It is also statistically less likely to be chance to achieve such a high score with such a low number of tests. If you want to set a pass mark of 80% with 10 tests, then the chance of someone achieving that just by chance is around 10%. Quite a few head-fiers would then claim to hear the difference, even if they were guessing every time. If the pass mark is 90%, then the odds of achieving it with 10 guess is around 1%. Significantly lower but we could still get one or two guessing their way to a pass.

I'd prefer 90% with at least 20 tests. Then I would agree that it is very unlikely the subject was not guesssing.


Nice to see this backed up with some statistics and logic.
Based on that, I would go with 8 out of 10, but that's just me.....
 

Users who are viewing this thread

Back
Top