Help - Search - Members - Calendar
Full Version: ReplayGain Analysis Hangs Computer on FLAC Files
Hydrogenaudio Forums > Lossless Audio Compression > FLAC
costello
While batch scanning dozens of FLAC files for ReplayGain per-track analysis with foobar 0.72, my computer will ALWAYS freeze--usually in 5 and no more than 10 minutes. It never freezes immediatley (within the first several files analyzed). By freeze, I mean that it stops responding to all input, the video output does not change, and, if any audio was playing, it is stuck in stutter mode. The only way to correct this is to reboot.

ReplayGain is configured to skip files with existing replaygain values. It would appear that this includes only files with both replay_track_gain and replay_track_peak. And, it would appear that some/many of the files analyzed have the peak value but not the gain value.

Ironically, based on what you will read below, I have no other stability issues with foobar and FLAC. I have discovered that FLACCER can encode many dozens of FLAC files without any of system instability. And, I have never had an issue with system instability related to decoding FLAC files in either foobar or Winamp.

However... I have had exactly the same system lockups while encoding with both the FLAC CLI and dbPowerAmp FLAC encoder. Following is a note that I sent to Josh for that issue...

======================

I suffer constant system hangs/freezes in Windows XP. By that, I mean that the video is visible but doesn't change and the system doesn't respond to mouse/keyboard/usb hotplugging. I am running an Athlon XP 1700+ on an ECS K7S5A motherboard with 512mb RAM. Windows XP SP1 + all critical fixes.

There are no events reported in event viewer. There are no error messages thrown on the screen before or after rebooting. SiSandra 2004 reports no errors in system hardware. And I have had these problems for months despite rebuilding the computer's software from scratch.

Any _extended_ period of FLAC encoding will result in these hangs. Interestingly, FLAC decoding seems to have no effect on system stability. I can go for many days of uptime if I only listen to previously encoded songs.

Examples of ways to freeze while FLAC encoding seem to encompass both the CLI encoder and the libFLAC:

* Encode 10-20 WAV files using dbPowerAmp. The first few usually will be fine but at some point (not a fixed number of songs), the system will hang. Note that I can encode just one or two songs to FLAC without any significant chance of error. It doesn't seem to be any one song (i.e. an especially tricky sample)--as I can reencode the WAV file which was incompletely encoded to FLAC without error.

* Encode using Linco (a line in capture utility) to FLAC via stdin. My radio programs last between half an hour to 2 hours. At some point within 3 hours, I can all but guarantee that the system will hang. Yet, I can encode to LAME with Linco to stdin for days on end without issue.
Case
Sorry to say but sounds like hardware problem. Those ECS boards are quite often somewhat tricky to get working flawlessly, if you have two RAM chips you could see if removing one of them helps with stability. Other way to increase stability is to run it underclocked, reduce FSB speed to 100MHz and it should run fine. There are more tricks on this forum.
Garf
Your mainboard is pretty much guaranteed to be the problem. Welcome to the club of ECS K7S5A victims. Recommended tips:

a) run with only 1 ram chip. the board can't handle 2 (most important!)
b) look up the northbridge cooling tweak on google.com
c) downclock to 100Mhz FSB

Suggested testing utilities: BURNMMX and BURNBX.

Quite basically, this mainboard is pure garbage, but it's sold at garbage prices too.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2008 Invision Power Services, Inc.