Well then it should be easy for you to pick the 128k versions here.
I began the test, but I noted first that even the reference sounds bad. Is it a true reference???
That is, is this really the original soundfile, before any compression? Than it must be of very low quality.
I found both tracks quite distorted, especially from 26-37 s (the instrumental part). The drum blends somehow with the other instrument and I found this annoying.
The intro (0 to 5 s) is quite simple, therefore 128 and 192 kbps shouldn't make any difference. This is definetely an example of how such a test SHOULD NOT be done. I recommend in the future to strip such portions from test files, because they do NOT bring any information.
From 4.8 - 7.3 s, both files sound again equally (low complexity), BUT the overall dynamic range of the output is somehow compromised. I tried various equalizer settings (external, NOT PC), but non could really compensate. Increasing the low frequency seemed to correct some of the problems, BUT it still had an artificial touch. (Though you can probably hear these differences only on a high end system; it happens that I have one.)
Another problem with this tune is, that one instrument is superposed on a backround hiss (you hear the drums more or less like a hiss, see 11-15.6 s, and up to 20+ s), so, even if there are differences, the brain is likely to process them as a hiss. Also, because of this hiss, the additional instrument (piano?) is overshadowed, so differences are again minimized.
21 to 26 s, again very LOW complexity, therefore NOT suitable for comparisons.
The real interesting part was that I consistently disliked the 27-36s sequence, because there were so many artifacts introduced.
I used an AA+ amplifier, and, for some of the tests, an external high-end equalizer.
Then I tried ABX sample 4 vs Original and the result were 5 of 5 (p=0.031). I used only the 28.60-33.28s interval for my guess. The test was difficult, but I was able to distinguish the 2 files. Though, this misses the point
. As stated previosly, I disliked even the original in this sequence.
And by the way, this thread misses the point. YOU NEED HIGH QUALITY originals, to be able to tell the difference. However, what one can ALWAYS tell, is IF the file sounds good or not
, and even a 320 kbps encoding might sound bad (because of poor original, or, on some difficult file, maybe due to the encoding itself). If 2 encodings sound equally bad, that doesn't make an 128 equivalent to 320 kbps. Basta.