Mail archive

[alpine-user] LAME unstable

From: Matthew Kaufman <>
Date: Mon, 7 Jan 2019 12:45:07 -0800

I am converting some transcoding tasks over to Docker, am
using openjdk:8-jdk-alpine as my base image. Unfortunately, I'm seeing
significant instability with the LAME MP3 encoder, with errors like these
(hard to reproduce, and never seen on my old CentOS platform):

Case 1:
Assertion failed: esv->header[esv->w_ptr].write_timing >= bs->totbit
(bitstream.c: putbits2: 166)
Case 2:
Internal buffer inconsistency. flushbits <> ResvSizebit reservoir error:
l3_side->main_data_begin: 480
Resvoir size:             552
resv drain (post)         3
resv drain (pre)          0
header and sideinfo:      104
data bits:                669
total bits:               776 (remainder: 0)
bitsperframe:             864
This is a fatal error.  It has several possible causes:90%  LAME compiled
with buggy version of gcc using advanced optimizations 9%  Your system is
overclocked 1%  bug in LAME encoding libraryAssertion failed:
gfc->bs.totbit % 8 == 0 (bitstream.c: format_bitstream: 972)
Suspecting I might be seeing a musl issue or a gcc optimizer issue based on
that last error message. I'll update if I learn more, but wanted to get
this out there for others to think about.
Received on Mon Jan 07 2019 - 12:45:07 UTC