Date: Thu, 31 Jan 2002 21:02:31 +0100 (CET) From: BOUWSMA Beery <freebsd-misuser@dcf77-zeit.netscum.dyndns.dk> To: Alexander Leidinger <Alexander@Leidinger.net> Cc: larse@ISI.EDU, multimedia@FreeBSD.ORG, gmh003532@brfmasthugget.se, MP3 Encoder List <mp3encoder@tuhs.org> Subject: Re: BCP: CD ripping Message-ID: <200201312002.g0VK2Vb00358@beerswilling.netscum.dyndns.dk> References: <20020128191513.77360a03.gmh003532@brfmasthugget.se> <200201291052.g0TAq1s03902@Magelan.Leidinger.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Moin, moin!
%s wrote on %.3s, %lld Sep 1993
> On 28 Jan, Martin Faxér wrote:
>
> > I personally don't think that you will have a problem with the output
> > quality of GOGO, but if you do, I guess you can check out LAME (although
> > GOGO is supposed to be based on it).
> GOGO is based upon an older version of LAME, I don't know if they ported
> recent quality improvements of LAME (maybe someone on the mp3encode
> mailinglist is able to tell more, please don't forget to readd
> multimedia@freebsd.org to the CC).
At the time I was actually doing something with `lame' beyond compiling
and using it, which was more than a year ago, with machines slower than
today's by something like an order of magnitude, I did some speed and
quality comparisons of `gogo' vs `lame', and what surprised me was that
lame had gotten much faster to the point where simple options given to
it made it nearly as fast as gogo on the hardware I used.
The home page for gogo refers to `petit' which is apparently based on
3.8x lame, but does not seem to be available as a port, and I haven't
looked at it at all. The documentation on the page for gogo gives its
base as lame 3.29 plus some improvements of the 3.5x series, so it looks
like the more recent improvements are probably not in gogo, but `petit'
may be worth a look, for those who are interested or maybe need speed.
> ${PORTSDIR}/audio/lame) of LAME has a lot of quality improvements
> compared to the previous stable release (3.7x), it's easy to hear them
> even for untrained ears).
Even back then, I could easily hear that lame (3.6wozzit?) gave a much
higher quality signal, being only slightly slower.
On the other hand, while I had one machine encoding three or so real-
time high quality streams with CBR and lame, I was running another
machine at max to encode a VBR stream, and not at highest quality.
It is possible to use different types of encoding that require much
more horsepower than `gogo' but deliver quite tangible improvements,
such as VBR and ABR encoding, or to select encoding options that give
more quality at the expense of CPU power.
Today I'm using ABR encoding, easy enough as fast as throwaway PCs
are today, to do in realtime with plenty of room to spare. This may be
five or ten times slower than `gogo' was, but it's worth it. The idle
time of machines today is ... something to be put to use.
It could perhaps be that I didn't see a tremendous difference in speed
between `gogo' and `lame' since I wasn't compiling them optimally, or
my hardware couldn't offer all possible improvements, and naturally my
comments are based on long-dated k0deZ, if gogo has received substantial
updates in the last year or more.
barry bouwsma, netscum
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-multimedia" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200201312002.g0VK2Vb00358>