opaqueice;279208 Wrote: 
> 
> Now, it's true that there is a connection between the smallest file
> size a given amount of information can ever be compressed into and the
> information content.  But a better way to address that would be
> something like this:  take a WAV file ripped from a CD and zip it, then
> compare that to a zipped 128 MP3 of the same track.  
> 
> -That- would give you a valid comparison, up to the limitations of the
> zip algorithm, and it will not give you the factor of 10 he mentions
> (it will give a factor that depends on the WAV file, for one thing).

Sorry, that does not make any sense - did you mean FLAC?  Zip is a
dictionary/window based algorithm which looks for duplicate strings of
bytes, or a non-random distribution of byte values. It might get some
amount of compression on a WAV file due to DC or low frequency content,
but this is a poor indication of the actual information content of the
WAV file. 

In any case, just because ZIP can't compress it does not mean it is
incompressible. You have to use a compression algorithm that is suited
for the data to be compressed. Conversely, FLAC would not be able to
compress a .EXE file, which might be highly compressible by ZIP.

Only a very poor compression algorithm (or one not suited for the input
data) would output data that can be compressed by further processing
using some other (lossless) compression algorithm.

A couple of trivial example: if you generated a file containing just a
triangle waveform, this would not be compressible at all by ZIP, but it
would be highly compressible by FLAC. Silence would be highly
compressible by ZIP, FLAC, or MP3. White noise would never be
compressible by _anything_, as it is 100% entropy by definition.


-- 
seanadams
------------------------------------------------------------------------
seanadams's Profile: http://forums.slimdevices.com/member.php?userid=3
View this thread: http://forums.slimdevices.com/showthread.php?t=44532

_______________________________________________
audiophiles mailing list
audiophiles@lists.slimdevices.com
http://lists.slimdevices.com/lists/listinfo/audiophiles

Reply via email to