Re: mp3 encoding - lame?
jurriaan wrote:
> Script started on Sun 21 Aug 2005 02:49:45 PM CEST
> INTEL :cat test.in
> file /usr/local/bin/lame
> r --vbr-new -V 0 "18 US Forces.wav"
> bt
> INTEL :rm *mp3; gdb < test.in
> <snip>
> Program received signal SIGSEGV, Segmentation fault.
> 0x000000000045224b in quantize_x34 ()
> (gdb) #0 0x000000000045224b in quantize_x34 ()
> #1 0x0000000000456cdb in VBR_noise_shaping ()
> #2 0x000000000044d888 in VBR_iteration_loop ()
> #3 0x0000000000443abb in lame_encode_mp3_frame ()
> #4 0x000000000040f126 in lame_encode_buffer_sample_t ()
> #5 0x00000000004104e7 in lame_encode_buffer_int ()
> #6 0x0000000000402a6e in lame_encoder ()
> #7 0x0000000000403763 in main ()
>
> This is with my own gcc-4.0 compiled lame. Sometimes it hangs, most of
> the time it crashes on this file. AFAIK the .wav file is OK, it plays
> and is a decompressed .flac file. It's 5.8 MiB, btw.
Are you saying that your lame behaves differently if you run it multiple
times on the same file? If that is the case you almost certainly have
faulty or overheating hardware.
- If you're overclocking, don't do so as much.
- Check to make sure your system is being cooled adequately.
- Run memtest86 and see if it likes your RAM.
For what it's worth, my machine has no problem with your test file and
"--vbr-new -V 0". You can try my deb if you want, but I suspect your
problems are hardware-related.
http://fatooh.org/files/lame_3.96-0.1_amd64.deb
-Corey
Reply to: