Here's the summary answer:
The MPEG spec for audio allows for compression based on psychoacoustic
studies or some such thing. Masked frequencies and other "redundancies" are
removed to allow for the compression. For various reasons, MP3 works really
well for overcompressed pop music (i.e. top 40 stuff); for a lot of
electronic and classical music, it kind of blows in general.
BUT--it makes nice, small (distributable) files that are useful for
demonstration purposes.
It's worth noting that the major mp3 sites (e.g. mp3.com, besonic) accept
only 128 bitrate files--they've standardized so that file sizes will be
reasonable.
So yes, the alteration of sound quality is a general problem. Incidentally,
you'll find the same problem (though slightly different) with the MiniDisc
format. This uses Mtrac compression, which is (I've heard) quite similar to
the MPEG spec.
I notice the most problems with flanging-type activity in the mid-high
frequencies, hi-hats especially (as someone mentioned). Oddly (and Paul
Wagorn will agree with me here), mp3 compression can actually create some
desirable changes, much as magnetic tape does. It adds a certain sense of
warmth that can sound really good in the mid-low area for certain sounds.
But if you're looking for precision, forget it. Dave--I'd be interested in
seeing (though I could do it myself at home) an oscilloscopic print of
square wave via .wav vs. square wave via .mp3. Maybe I'll do a test at
several frequency points. I've never considered doing that.
Myself, I'm used to what mp3 does to my music. It messes it up, but I still
mix to full-scale specs. I ∗do∗ sometimes "master" my tracks for mp3's sake,
because until now I've been distributing my music for free online via
mp3.com. But it's difficult to do and involves a lot of trial and error.
Best thing to do is suck it up, recognize the imperfections, and applaud the
wonderful file size reduction capabilities. We're only moments away from
more spectacular results, and I think within a couple of years, another
flavor of this style of compression, far superior, will be standard. Either
that or anyone who cares will just use such a codec by default, and it will
be easier to find the tools to encode files for high fidelity.
Regarding layer 1, 2, 3--there is a lot of technical information about these
online. Basically, the audio standard that most use is layer 3. I have heard
people say that these are successive iterations, each better than the last,
but I'm pretty sure that's not true. For a good read (for those of you who
are good at math and physics--not my strong suits), check out the
Fraunhoefer Institute, who developed the basic codecs. They have a site out
there somewhere that goes into excruciating detail about the various MPEG
codecs. Incidentally, I talked with a couple of the guys from the institute
at last year's AES and they're totally appalled that the mp3 format has
taken off the way it has--they're years ahead of it by now, and they're
realizing that they opened Pandora's box. Their feeling is "but we have such
better stuff now! It sounds so much better!" But of course, no one cares,
and the de facto standard, at least for the time being, has been set.
Have a great weekend everyone!
David.
> -----Original Message-----
> From: Dave Bradley [mailto:daveb@...]
> Sent: Friday, June 30, 2000 8:45 AM
> To: motm@egroups.com
> Subject: [motm] fun with mp3 converters
>
>
> I just downloaded the encoder suggested below by David. This is my first
> time playing with them. I took Larry's .wav file as input, and played with
> all the different codecs and bitrate settings.
>
> Each and every one of them altered the sound audibly. It was plain bad
> (gargling) until I bumped the bitrate up to 256 or 320. Even at
> the highest
> rate, it changed the sound of the VCOs from more of a pulse wave in the
> original .wav file, to a hollower, more square wave sound in the
> .mpg file.
>
> Is this alteration of sound quality a general problem with mpg
> encoders? How
> do you seasoned mp3'ers deal with this?
>
> Dave Bradley
> Principal Software Engineer
> Engineering Animation, Inc.
> daveb@...
>
> > -----Original Message-----
> > From: dbivins@... [mailto:dbivins@...]
> > Sent: Thursday, June 29, 2000 10:06 PM
> > To: motm@egroups.com; Paul Schreiber
> > Subject: Re: [motm] DIY - interval switch, audio sample
> >
> >
> >
> > Ech, sharpen your teeth first on the very FREE Electronic Cosmo. Go to
> >
> > http://ec2000.xperiment.net/projects.html
> >
> > and you'll find a dead-simple mp3 encoder for Windows (only,
> > sorry if that's a drag), including 3 codecs, plenty of options,
> > rivalling ones that cost money. Save yer bux fer MOTM!
> > I like it better than the Xing program anyway.
> >
> > David.
> > -----Original Message-----
> >
> > I ∗highly recommend∗ the Xing Audio Catalyst program. It's like $29, you
> > give them credit card and
> > then you download it. 1 yr of free updates. Works like a dream.
> >
> > Paul S.
> >
> >
> >
> > ------------------------------------------------------------------------
> > Special Offer-Earn 300 Points from MyPoints.com for trying @Backup
> > Get automatic protection and access to your important computer files.
> > Install today:
> > http://click.egroups.com/1/5667/6/_/529958/_/962332079/
> > ------------------------------------------------------------------------
> >
> >
> >
> >
> >
> >
> >
> >
> > ------------------------------------------------------------------------
> > Free @Backup service! Click here for your free trial of @Backup.
> > @Backup is the most convenient way to securely protect and access
> > your files online. Try it now and receive 300 MyPoints.
> > http://click.egroups.com/1/5668/6/_/529958/_/962334360/
> > ------------------------------------------------------------------------
> >
> >
>
>
> ------------------------------------------------------------------------
> Need a credit card?
> Instant Approval and 0% intro APR with Aria!
> http://click.egroups.com/1/6034/6/_/529958/_/962376438/
> ------------------------------------------------------------------------
>
>