Rockbox mail archiveSubject: Recording V2.4 vs. CVS 050828 (strange Xing VBR header - possible value overflow?)
Recording V2.4 vs. CVS 050828 (strange Xing VBR header - possible value overflow?)
From: Martin Borus <mborus_at_gmail.com>
Date: Mon, 29 Aug 2005 23:10:35 +0200
I've had some problems with long time recordings lately. Problems were
caused by the VBR header - the recordings played, but navigation in
them was not working properly.
To check this out I made two nearly identical recordings. One using
firmware 2.4, one using CVS 050828. Both are aprox 1.5 GB in size.
Recording settings were quality 5, 44.1 kHz, stereo, independent
See comparison screenshot
The 2.4 recording has no Xing header and works OK in Winamp,
mp3directcut and other programs. The CVS 050828 recording has a VBR
header, but the reported values are very strange. Average frame size
and Average bit rate values are useless. This causes rewinding of the
file to fail and ruins the navigation within the file.
Is it possible that the calculation of these values fails in Rockbox
for extremely large files?
If this can't be fixed, it might be useful to make the VBR header
optional in 2.5
Received on 2005-08-29