[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Source Code of Music (was: various opinions on Debian vs the GFDL)

Hi Edmund,

On Dienstag 03 Juni 2003 19:12, Edmund GRIMLEY EVANS wrote:
> Thomas Uwe Gruettmueller <sloyment@gmx.net>:
> > > You have a similar but less severe problem if A is a
> > > high-precision digital recording (with lots of random
> > > noise in the low bits) and D is a compressed version:
> > > clearly A is source of D,
> >
> > I would argue that D is an excerpt of A.
> If someone were to improve the compression algorithm then you
> would want to go back and recompute D, particularly if D has
> an audible artefact, so I would say that A is source.

The GPL defines the "source code for a work [to mean] the 
preferred form of the work for making modifications to it" (l. 
155f). If I take a wav file, compress it into the Vorbis format, 
and decompress it back to wav, I can modify this file the same 
way as I could do with the original one. There are differences 
in the files' contents, but there is no functional difference. 
(The resulting file will even have the same length as the 
original one, which is not true for a decompressed MP3.)

This is IMHO something entirely different than compiling and 
decompiling a program. In this case, the entire program would 
become less readable and thus much less modifyable.

> Of course, if we're talking about lossless compression
> (possibly preceded by throwing away the low bits)

With lossless compression algorithm (like flac), nothing is 
thrown away. It works more like gzip.

> then I might agree it's an excerpt, but I was thinking about
> something like Ogg Vorbis.

I was thinking about these cases:

A) I don't have enough bandwith or webspace to distribute the 
entire uncompressed work. So, in order to reduce the file size, 
I could reduce the duration of the recording, by only 
distributing the most interesting part. This would be clearly an 

B) If I wanted to preserve the sample's duration, I could also 
throw away details all across the sample, by reducing the sample 
rate, the bit resolution, and the number of channels. I would 
still consider this an excerpt.

C) I could also use a more advanced lossy compression codec like 
Vorbis. Like in A and B, I would throw away data, but the losses 
would be better selected. I would still consider this an 
excerpt, and an even more desirable one than the one in B, 
especially for modification purposes.

"Look! They have different music on the dance floor..."

Reply to: