WavPack Hybrid broken on (Arch?) Linux, non lossless and heavily distorted |
WavPack Hybrid broken on (Arch?) Linux, non lossless and heavily distorted |
Jul 5 2012, 18:47
Post
#1
|
|
![]() Group: Members Posts: 1147 Joined: 4-May 04 From: France Member No.: 13875 |
I tried to encode and decode several different WAV files from CD rips with WavPack (version 4.60.1) in hybrid mode, on Arch Linux, on two different machines (Intel Core i7 and Xeon, x86_64). I always got the same result: the files are heavily distorted (groups of samples repeating themselves), and obviously non lossless. See samples.
I ran the following commands: CODE wavpack -b400 -c -m hybrid.wav wvunpack -m hybrid.wv wvunpack outputs the following: CODE WVUNPACK Hybrid Lossless Audio Decompressor Linux Version 4.60.1 Copyright (c) 1998 - 2009 Conifer Software. All Rights Reserved. original md5: e180f2ec2c4c65fd214bc6c4f70c7988 unpacked md5: 2d5edeb0d3d7ba971282428b77fd41c6 MD5 signatures should match, but do not! I tried the win32 binary, which worked fine (lossless result). I don't have another Linux distro installed anywhere to see if this problem is distribution-specific. I also tried three different builds (including one built from source by myself), always with the same result. This post has been edited by skamp: Jul 5 2012, 18:59 -------------------- caudec -c lossyFLAC -q S *.flac
|
|
|
|
![]() |
Dec 2 2012, 10:51
Post
#2
|
|
![]() Group: Members Posts: 147 Joined: 18-May 10 From: Montana, USA Member No.: 80732 |
I'm a bit stymied as i've been using wavpack-4.60.1 since june 5 and i only ran into this problem for the first time last week (gentoo system).
Is there a patch available to fix this yet? Thanks for all the great work! -------------------- Music lover and recovering high end audiophile
|
|
|
|
skamp WavPack Hybrid broken on (Arch?) Linux Jul 5 2012, 18:47
plus I can confirm this bug exists on Debian Wheezy amd... Jul 5 2012, 19:21
2012 Fellow Archer here.
I confirm the bug. I got non-... Jul 5 2012, 20:28
Angenial The bug itself appears not to be architecture-rela... Jul 5 2012, 21:08
2012 Good catch Angenial.
Using memmove() instead of m... Jul 5 2012, 21:46
lvqcl Links about memcpy() behaviour:
https://bugzilla.... Jul 5 2012, 21:59
Kohlrabi QUOTE (lvqcl @ Jul 5 2012, 22:59) Links a... Jul 6 2012, 00:55
bryant Thanks guys! I have checked Angenial's pat... Jul 5 2012, 22:19
bryant I knew this sounded familiar!
http://www.hy... Jul 5 2012, 23:22
skamp Any ETA on a bugfix release? Sep 6 2012, 18:03
bryant I guess I’ve sort of dropped the ball on this. I a... Sep 7 2012, 07:25
Destroid QUOTE (bryant @ Sep 7 2012, 07:25) ...it’... Dec 2 2012, 12:57
Mr_Rabid_Teddybear QUOTE (BearcatSandor @ Dec 2 2012, 02:51)... Dec 5 2012, 01:33
BearcatSandor Thanks for the link to the SVN and instructions. I... Dec 7 2012, 01:26
bryant QUOTE (BearcatSandor @ Dec 6 2012, 16:26)... Dec 8 2012, 19:48
BearcatSandor Bryant,
Thanks for asking. It happens when using ... Dec 8 2012, 22:12
bryant Well, the first thing is do you get the same MD5 m... Dec 8 2012, 23:03
BearcatSandor Got it! Thanks Bryant and Mr_rabid_teddybear f... Dec 9 2012, 21:28
skamp I found another bug (still with version 4.60.1, Li... May 15 2013, 15:06
bryant Thanks, Skamp!
Yeah, the 64-bit Linux memcpy(... May 15 2013, 17:24
skamp OK, cool! May 15 2013, 18:28
skamp By the way, this kind of bug calls for generating ... May 25 2013, 10:09
bryant Yeah, the problem with this bug is that the issue ... May 26 2013, 01:51
skamp Another shameless plug: caudec SVN (upcoming versi... May 26 2013, 08:28
bryant QUOTE (skamp @ May 26 2013, 00:28) Since ... May 29 2013, 07:01
skamp Moreover, when asked to compute MD5 hashes, caudec... May 29 2013, 11:15![]() ![]() |
|
Lo-Fi Version | Time is now: 18th June 2013 - 23:45 |