1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Testing the claim: "I can hear differences between lossless formats."

Discussion in 'Sound Science' started by music alchemist, Oct 15, 2014.
38 39 40 41 42 43 44 45 46 47
  1. cuiter23
    Theres a lossy vs lossless comparison thing on Tidal and if you pass you get 14 days free trial. I got 5/5 the two times I did the test. Not sure how reliable it is after all it is a marketing tool for Tidal to get you to subscribe.
  2. Music Alchemist
    Not really the topic of this thread...but as a general rule, you should convert the files yourself to isolate the variables. It has been said here in the past that 256 kbps AAC is indistinguishable from lossless.
    Also, you need at least 9/10 or 15/20 for your results to be statistically significant.
  3. castleofargh Contributor
    there is a topic about the tidal fake test  http://www.head-fi.org/t/743658/tidal-lossless-listening-test-whats-going-on-here
    where our brave fellow members took it up upon themselves to download and look at the files. and the conclusion was: yes there is a reason why we all succeed that fake test while failing all the ones we tried by converting the files ourselves ^_^.
    cel4145 likes this.
  4. Steve Eddy

    Do they say exactly what format and bitrate the lossy format was? Because Harman did something similar to market their "Clari-Fi" which claims to put back what the lossy codec removes. And how they did it should be criminal.

    First, although the subject was lossy data compression, throughout their video, they kept conflating this with dynamic range compression, which are two completely different things.

    But where they committed an absolute fraud was when the did a "comparison." This was done both audibly and graphically, by playing a snippet of music that played for a little while and was labeled "Uncompressed Audio," and then at a point in the clip they switch to "Compressed Audio." Here's a still capture from the video:


    The "Compressed Audio" starts where you see it highlighted in yellow.

    Looking at this, two things are evident. First, what they're showing is not the result of lossy data compression, but of dynamic range compression. Even if you apply data compression down to 96kbps and even lower, at the scale displayed here, the waveforms look identical. This is very severe dynamic range compression, if not hard limiting.

    Second, not only has it been dynamic range compressed, they significantly reduced the level. Normally, if you just did dynamic range compression, the "Compressed Audio" portion would sound louder than the "Uncompressed Audio" (dynamic range compression is what is at the root of the "loudness wars"). So they had to resort to reducing the level so that it will sound muted by comparison. You can easily hear a difference even over an iPhone speaker.

    Again, a fraud such as this should be criminal.

    kraken2109 likes this.
  5. sonitus mirus
    Yes they do.  The test is between their lossless (FLAC/ALAC @ 1411 kbps) and AAC 320 kbps subscription streams.  
    I have "Hotel California", which was/is one of their test samples. (they may have changed the test samples)  I attempted to ABX my own ripped AAC 320 and FLAC (using dBpoweramp) and failed.  I even attempted to ABX the Tidal FLAC sample with my AAC 320 version and failed that as well.  Yet, I can buzz through their test in about 5 minutes and pass 15/15.  I did this 3 times in a row and passed 100% (15/15) each time.  I can't even hear anything over 16 KHz and I passed their test with flying colors.  It is not even difficult.  Just listen for increased bass levels.  
  6. cuiter23
    i took a look on the link that was suggested to me and that is seriously sketch. Tidal streaming sounds great but for them to stoop that low is seriously worthy of a lawsuit. False advertisement.
  7. Steve Eddy

    Thanks. But after reading through the thread that CofA referenced, it looks a bit fishy. Harman dwarfs Tidal and look what they resorted to. And it even has the implicit approval of one of their head researchers, Sean Olive. A person for whom I previously had a lot of respect for, along with Floyd Toole who had previously retired from Harman. It's downright depressing what will be resorted to for the sake of making a buck.

  8. castleofargh Contributor
     CofA harman did a great job with the UN. kofi-annan1.jpg
    I will go now... [​IMG]
    earthpeople likes this.
  9. sonitus mirus
    Not sure if the intent is to misdirect the consumer.  I really think the differences fall into a similar realm of "Mastered for iTunes", where the lossy version is different, though not necessarily inferior.  
  10. Steve Eddy


  11. JWolf
    There could be a very valid reason why the same track sounds differently even though it's played in different lossless formats. It could be due to an issue with the codec used. Say there's a bug in the AIFF codec, then it could alter the sound.
  12. JWolf
    Do you think a 256kbps ACC file would be indistinguishable from a 24/96 lossless file? I think the answer would be no way it would be.
  13. JWolf
    I have heard some of these mastered for iTunes on an iPhone and they didn't sound all that good.
  14. hogger129
    The only way to know would be to do a blind test and see if the difference could be heard.  On my equipment, I can hear a difference between the CD and a high-bitrate AAC file when I know what I'm hearing.  Otherwise, I cannot.  I have also purchased some good 24-bit material from HDtracks in the past, downsampled it to CD-quality with Voxengo r8brain, and was not able to tell them apart.  It makes sense since the limit of human hearing is somewhere around 20khz.
    iPhones do not sound all that good out of the headphone jack.  Listen out of the dock connector, or use a better portable player.
  15. JWolf
    I cannot use the dock connector. It doesn't support my Fiio X3II. I get a message saying it draws too much power.
38 39 40 41 42 43 44 45 46 47

Share This Page