[MPlayer-users] Re: NO audio with mencoder

pl p_l at tfz.net
Tue Dec 18 18:29:07 CET 2001


On Tue, Dec 18, 2001 at 04:24:54PM +0100, Sven Hartge wrote:

> Arpi <arpi at thot.banki.hu> wrote:
> 
> >>> mp3lame compiled with bad compiler or bad cflags ? rtfm?
>  
> >> Here is the output of the 'make test' of a fresh compiled (gcc version 2.95.3 
> >> 20010315 (SuSE)) lame 3.89:
> 
> >> cmp -l testcase.new.mp3 ./testcase.mp3 | wc -l
> >>     415
> 
> > have you read lame docs? this number must be <=20
> 
> You have to completely turn of _any_ optimization-flags (no -march= or
> -mcpu= and eve no -O2 or -O1, you have to force -O0) to achieve something
> below 20. (It does not matter, if you use gcc-2.95, gcc-3.0x or egcs
> 1.12)

I've just checked (read Makefile and search test:) and it was 30 in 3.89beta.
#  Because of compiler options and effects of roundoff, the 
#  number of bytes which are different may not be zero, but
#  should be at most 30.

In the latest CVS of lame:
# The differences depend on the used processor architecture, the used
# compiler and the used options for the compiler, so make test may
# show some differences. You should only be concerned if you are a
# LAME developer and the number of differences change after you
# modified the source.

So I guess it's not probably most critical but doing 'make test' does
not cost much and permits to early detect a broken executable.

-- 
Best regards,
  pl




More information about the MPlayer-users mailing list