This was posted almost a year ago and enjoyed a decent conversation in the comments [0]<p>I'll repeat my comment from the time:<p>"For what it's worth, the presence of seemingly significant signal in the difference between the original and compressed tracks does not necessarily mean that significant sonic/perceptual loss has occurred. Operating correctly, the encoder is designed to cut not just sounds that the human ear cannot hear in general (e.g. sounds above 22kHZ) but also sounds which may not be perceptible in context (e.g. the quietest signals in a loud section). So if you find something beautiful about the ghost tracks (and I think there is something beautiful to find), don't immediately jump to concluding that mp3 is awful for cutting these sounds—they might be hardly perceptible when added to the mix.<p>Of course, at high-compression rates mp3 does begin to significantly degrade fidelity.<p>Edit: all of this is not to put down the project—I still think it's pretty cool as art and as a demonstration of the encoder, I just didn't want people to think that this was some sort of massive failing of mp3."<p>[0]: <a href="https://news.ycombinator.com/item?id=7955917" rel="nofollow">https://news.ycombinator.com/item?id=7955917</a>
The information present here isn't really "lost" as much as it can't be heard in the context of the other sounds in the original recording. These forms of audio compression take advantage of auditory masking[1] which means those sounds likely wouldn't be heard in the original.<p>[1] <a href="http://en.wikipedia.org/wiki/Auditory_masking" rel="nofollow">http://en.wikipedia.org/wiki/Auditory_masking</a>
A quick scan didn't reveal to me whether he's time aligning the signals during the subtraction. I've played with listening to the wav-mp3 signal before, and I seem to recall that the mp3 encoder would introduce a little delay.<p>I added a transient pulse in front of the music so that I could (visually) time align the signals before subtracting them.
The file we are watching is lossily compressed. So we are watching the lossy compression of a delta between original and lossy compression.<p>How good is the lossy compression at capturing that delta?
If this kind of thing interests you, I highly recommend the book "MP3: The Meaning of a Format" by Jonathan Sterne [0]<p>[0]<a href="https://www.dukeupress.edu/MP3/" rel="nofollow">https://www.dukeupress.edu/MP3/</a>
How does he get the information lost in compression? Does he put the compressed and uncompressed version on two different tracks with one phase-flipped?