4 ms·

I picked 4 out of 6 and never picked 128kbps. I am pretty sure I can reliably weed out 128Kbps, but WAV vs 320Kbps is more or less luck. Jay-Z was the easiest a

I picked 4 out of 6 and never picked 128kbps. I am pretty sure I can reliably weed out 128Kbps, but WAV vs 320Kbps is more or less luck. Jay-Z was the easiest as the chiptune sample is a rectangular wave and hence has a very high frequency component that actually makes sound less palatable, not more. "Hissiness" of strings helps on classical and Neil Young, but pop music (Katy Perry and Coldplay) with its flat wall of sound is much harder - I was listening for "z"s and "s"s and high hats. Acapella (Suzanne Vega) was the hardest for me and I think I just lucked out.

cogman10 5y ago
Same results. I'd love to see a similar experiment with newer codecs like Opus.
dsr_ 5y ago
Suzanne Vega ought to be the hardest thing in the world to find an MP3 artifact in, since "Tom's Diner" was the primary litmus test used in the development of the psychoacoustical model. https://en.wikipedia.org/wiki/Tom%27s_Diner#The_%22Mother_of_the_MP3%22 https://en.wikipedia.org/wiki/Tom%27s_Diner#The_%22Mother_of...
- nr2x 5y ago
  I also found that really hard, and now I know why - that's amazing!
- fogihujy 5y ago
  Great. I was feeling good about myself for managing to pick the uncompressed one, and now I keep wondering if I just got lucky. :D
- dusted 5y ago
  I'd argue the opposite, "voice is the hardest to get right so we chose a vocal track" -> voice is hard to get right, so they got it least right.. For me, Tom's Diner was the easiest of the lot to pick out.
  - blowski 5y ago
    I was the opposite. The classical was easiest because it had so much texture that was easier to spot tiny flaws in that texture.
    - alanbernstein 5y ago
      Can you explain what exactly "texture" means here?
      - blowski 5y ago
        At any point, there are many different types of instrument playing a variety of pitches and melodies.
superseeplus 5y ago
I have a similar experience with being easily able to distinguish between uncompressed and 320kbps MP3s with certain kinds of music. Music with a lot of emotion in the singer’s voice or with both instruments and singers spread over a large area (Bollywood music from the late 90s) are easier to distinguish. Of course, I tested this with two different streaming services, so it may just be differently processed files. Modern pop songs on the other hand sounds the same to me irrespective of the level of compression.
saltcured 5y ago
I did the quiz once and also got 4/6. I alternated through all the samples around 4 times before making each choice. I messed up on the only two where I felt baffled by the choices, so at least I know what I don't know. :-) I chose the 320 kb/s for Katy Perry but was vacillating between that and the actual WAV. The one I completely screwed up was Jay-Z where I picked 128 kb/s. However, I knew I was lost with so much electronic glitching and distortion. While I heard differences, I had no idea what was a compression artifact versus a synthesis/mixing artifact. I also had no prior listening experience for either of these two, so no preconception for how it should sound. For the other recordings, I think I could hear the masking of the psychoaccoustic models. Essentially, in the WAV I heard a "fuller" mix with overlapping sounds, while in the compressed versions I heard dropouts where a dominant sound was present in what felt like isolation. For Suzanne Vega this wasn't so easy, but I felt her voice and the reverb were more clear and natural in the WAV. This was just with my old Sony MDR-NC22 noise canceling ear buds driven by my Thinkpad audio jack. My idea of a hi-fi environment is a Yamaha receiver and reasonable tower speakers like my Polks.

cogman10 5y ago

Same results. I'd love to see a similar experiment with newer codecs like Opus.

dsr_ 5y ago

Suzanne Vega ought to be the hardest thing in the world to find an MP3 artifact in, since "Tom's Diner" was the primary litmus test used in the development of the psychoacoustical model. https://en.wikipedia.org/wiki/Tom%27s_Diner#The_%22Mother_of_the_MP3%22 https://en.wikipedia.org/wiki/Tom%27s_Diner#The_%22Mother_of...

nr2x 5y ago
I also found that really hard, and now I know why - that's amazing!
fogihujy 5y ago
Great. I was feeling good about myself for managing to pick the uncompressed one, and now I keep wondering if I just got lucky. :D
dusted 5y ago
I'd argue the opposite, "voice is the hardest to get right so we chose a vocal track" -> voice is hard to get right, so they got it least right.. For me, Tom's Diner was the easiest of the lot to pick out.
- blowski 5y ago
  I was the opposite. The classical was easiest because it had so much texture that was easier to spot tiny flaws in that texture.
  - alanbernstein 5y ago
    Can you explain what exactly "texture" means here?
    - blowski 5y ago
      At any point, there are many different types of instrument playing a variety of pitches and melodies.

superseeplus 5y ago

I have a similar experience with being easily able to distinguish between uncompressed and 320kbps MP3s with certain kinds of music. Music with a lot of emotion in the singer’s voice or with both instruments and singers spread over a large area (Bollywood music from the late 90s) are easier to distinguish. Of course, I tested this with two different streaming services, so it may just be differently processed files. Modern pop songs on the other hand sounds the same to me irrespective of the level of compression.

saltcured 5y ago

I did the quiz once and also got 4/6. I alternated through all the samples around 4 times before making each choice. I messed up on the only two where I felt baffled by the choices, so at least I know what I don't know. :-) I chose the 320 kb/s for Katy Perry but was vacillating between that and the actual WAV. The one I completely screwed up was Jay-Z where I picked 128 kb/s. However, I knew I was lost with so much electronic glitching and distortion. While I heard differences, I had no idea what was a compression artifact versus a synthesis/mixing artifact. I also had no prior listening experience for either of these two, so no preconception for how it should sound. For the other recordings, I think I could hear the masking of the psychoaccoustic models. Essentially, in the WAV I heard a "fuller" mix with overlapping sounds, while in the compressed versions I heard dropouts where a dominant sound was present in what felt like isolation. For Suzanne Vega this wasn't so easy, but I felt her voice and the reverb were more clear and natural in the WAV. This was just with my old Sony MDR-NC22 noise canceling ear buds driven by my Thinkpad audio jack. My idea of a hi-fi environment is a Yamaha receiver and reasonable tower speakers like my Polks.