i meant the first generation of the mp3 codec, vs. mp3pro, not
reencoding, which i agree will degrade the data terribly.
on the vbr issue, it is my opinion that the relief of the sound may have
a tinny quality to it, which may be more or less detectable depending on
how high or low your bitrate is, etc., but i wouldn't want to sustain
this opinion in a discussion. there are so many other issues to talk
about besides, for example the number of instruments in a work.
i will also say that we sometimes get to a point on this topic where
discussion is really not advanceable without knowledge of engineering,
programming and mathematics which i simply do not posess. i have only my
intuition in reading into this topic as a tool. so i will be the first
to bow out of that part of the discussion. but i will suggest, in answer
to how "vbr might do that where cbr didn't", that a series comprised of
one repeated operation is more faithful than a series of different
operations, on principle of logic.
as far not accepting anything below 192k, i never said that. i was just
suggesting guidelines. i'll take what i can get, but i'll always shoot
for the best. i have had all sorts of experience with files of many and
diverse characteristics, and what i said is only a reflection of that. i
will not try to detail this lengthy catalogue of testing, etc. ... of
course, i have no idea of what you may think is a "better" file, and
even less so a "good" one.
what i liked about what you say is your argument for being able to
recommend better vbr settings for any given cbr settings someone may be
using. i think this is indeed possible, especially for a specific track,
although again i do not know what your idea of "better" is... for my
part i referred my arguments to the standards for psychoacoustics that
are a part of the mp3 codec; that is my idea or the idea i have accepted
of what sounds well. the idea of finding the optimum vbr settings is
nonetheless always tempting.
anyway, this was just to show what sense i think what i said may contain.
your reply, despite its being intelligent, was somewhat rushed and
abbreviates where it needed the most support for its arguments... (i
won't stoop to calling it "a crazy statement", or claiming that "it just
doesn't make 'logical sense' ", whatever that means, though). i guess
it's my fault, for not being brief, so i apologize. : )
On Thursday, January 31, 2002, at 01:50 , Michael Plump wrote:
quoted 32 lines On Thu, 10 Jan 2002, excm. prin. fuminaro konoye wrote:
> On Thu, 10 Jan 2002, excm. prin. fuminaro konoye wrote:
>
>> if you make a VBR file you are punching the midrange of the sound right
>> in the gut.
>
> I'm sorry, that's just a crazy statement. How could VBR do that where
> CBR
> didn't? So you refuse to accept that anything lower than 192kbps could
> possibly be good? Set your minimum compression level to that, and
> you'll
> have a better file.
>
> If you're using a good encoder and you show me a quality CBR file, I can
> show you setting that would've produced a better sounding VBR file.
> Think
> about what you're saying. It just doesn't make logical sense.
>
>> 192 is if you just want a "useful" mp3, but it must be accepted that
>> the
>> first generation mp3 must have a higher bitrate than that for real
>> quality.
>
> You really shouldn't ever, ever, ever re-encode your mp3s. All of a
> sudden you're taking the artifacts that you produced when you encoded it
> the first time, and forcing your encoder to try and encode those
> artifacts. You're going to lose a lot of quality.
>
> I've never tried this, but I would guess that an mp3 file re-encoded
> from
> 256->192 would probably sound worse than one just encoded originally at
> 128. Does anyone have any experience with this?
>
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at
http://mail.yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: idm-unsubscribe@hyperreal.org
For additional commands, e-mail: idm-help@hyperreal.org