OT - Headphone (Can) Shopping

Poormanq:  I'm sorry, but you're simply wrong here.  Lossless compression doesn't remove or clip anything.  Yes, _LOSSY_ compression, such as that used in mp3s or Ogg files, DOES use perception based encoding.  LOSSY compression does remove or clip information.  LOSSLESS compression does neither.  FLAC is LOSSLESS.  FLAC does not remove or clip information. 

I'm not talking about mp3s or any other lossy scheme.  I'm talking about lossless compression.

Not all compression removes information.  Zip files are an example.  If a zip file removed information we'd never be able to open the binary files contained within once compressed.  FLAC actually works very very similarly to zip compression.  Indeed, the MAJORITY of compression schemes use lossless compression.  This text, for example, is compressed and decompressed at multiple layers between the server and your computer.  Nothing is lost.  I've written compression algorithms in my past life as a programmer.  At the core, and very basically, most lossless compression schemes focus on reducing the length of code used to represent redundant pieces of data.  Just because the representation has changed does not mean that the data itself has changed!!   See this introduction to compression techniques:  http://www.faqs.org/faqs/compression-faq/part2/

i.e.  (adapted for plain english)  why store " programmer ", including the spaces, 100 times in a document about a programmer?  Instead, why not define "asdf" as " programmer " and store that ("asdf") 100 times instead?  Instead of storing 12 characters you're now storing 4.  Multiply this savings by 100.  You're saving nearly 800 characters.  It takes time to put the document together, and time to reconstruct the document, but once done it is exactly the same.  Actual compression algorithms take this further and determine the probability of encountering patterns before creating the corresponding codes.  The most frequent patterns are assigned the shortest codes.  So, while " programmer " might have been a relatively frequent pattern, something like, " and " is probably much more frequent and might be assigned the more simple code "asd" in the lookup table instead. 

Anyway, I hope you see that not all compression schemes are lossless. 

lol - "supposed lossless"

You seem like a smart guy, I'm frankly shocked that the concept of lossless compression is such a novel concept to you.  Perhaps you need to do some more basic reading.  Here:  http://computer.howstuffworks.com/file-compression.htm

Our computers use lossless compression ALL THE TIME.

What?  Now you aren't making any sense at all.  At the core even .wav files are strings of 1s and 0s.  How is this any different than a normal text file?  You're right, music isn't at repetitive as text, but this simply effects the percentage of compression.  It surprises me that you can't see this. 

Any "loss" over the original music is inherent in the digital version of the music itself.  i.e. I agree that analog to digital loses some resolution in the initial process of converting sound waves to 1s and 0s.  However, once digital, it is certainly possible to compress the information in a .wav file without losing anything.  It would be absolutely foolish to believe that the fact the data represents music in the final stage makes it "special" in any way.

You are a smart guy - you ask questions and challenge your understanding and the understanding of others.  In my opinion that process is more valuable than actually knowing the underlying fact.

Your instinct was right.  Music is not as compressible as most programs or text, but it is still compressible.  :-)

Now - back to the behavior that Haemphyst described - whether the actual FLAC algorithms correctly take the pieces apart and put them back together again might be another story.  I'm interested in hearing his view.  However, based on a light reading of the information available on the flac page, the resulting sound from flac and from the wav should be identical.