[MPlayer-users] Still no vorbis audio

Rolf Ernst rolf.ernst at silverlightning.org
Sat Jul 10 01:59:58 CEST 2010


On Fri, Jul 9, 2010 at 4:14 PM, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:

> Rolf Ernst <rolf.ernst <at> silverlightning.org> writes:
>
> > Having switched to libvorbis ...
>                     ^^^^^^^^^
>
> > I am using a very straightforward command such as
> >
> > mencoder "c:\a.ts" -o output.mkv -oac lavc -ovc lavc -lavcopts
> > vcodec=mpeg4:acodec=vorbis -of lavf'
>                       ^^^^^^
>
> Note that there is a difference, but don't worry, I don't think libvorbis
> is
> supported. I could ask again now why you are not using ffmpeg, but since
> you
> insist: Change the value in line 958 of libavcodec/vorbis_enc.c to
> something
> completely different (I just tested 30 successfully) to use libavcodec's
> native
> encoder: While some say it is bad, I think they just mean it doesn't
> support
> rate-control (as you found out).
>
> Feedback about quality problems - with sample(s)! - is probably welcome.
>
> Carl Eugen
>
> PS: The 0.01 is quite new there, I wonder what made us change the value...
>
> _______________________________________________
> MPlayer-users mailing list
> MPlayer-users at mplayerhq.hu
> https://lists.mplayerhq.hu/mailman/listinfo/mplayer-users
>


Carl, thanx, yes - that did the trick. Am I right to interpret it that atm
out of the box vorbis from svn is kaputt?

As for the value of venc->quality - what does that relate to? I can specify
abitrate but it is apparently ignored. I picked your 30 and it sounds good
to me. I haven't really calculated the effective bitrate (webm HD encoding
is a bit slow) and there is nothing in the container which would give me an
inherent clue. Does the 30 have any particular meaning? (vs. the 0.03 I had
in my code)?

I can upload something a bit later (tell me again where - do you want HD, SD
?). I have been playing with a number of the VP8 specific encoding
parameters and the clearly do something albeit I am not sure what. vbitrate
is acknowledged but also qmin and qmax although the max value according to
the book of 63 is not accepted - anything above 51 is rejected but I seem to
recall they same being true for ivfenc.

I think the quality rivals h.264. I have seen the various comparisons of
stills but I don't believe that tells the whole story. I think it'll be a
player in the long run ...

/re
-- 
/re

This is America. Huge multinationals choose our political officials here.


More information about the MPlayer-users mailing list