Re: [Muine] Problems with UTF-8 on ID3 tags

[Carlos Laviola <carlos laviola org>]

On Fri, Jun 25, 2004 at 11:14:47PM -0700, Brian Kerrick Nickel wrote:
> On Sat, 2004-06-26 at 00:28 -0300, Carlos Laviola wrote:
> > On Fri, Jun 25, 2004 at 06:37:18PM -0700, Brian Kerrick Nickel wrote:
> > > I think the diagnosis may be the opposite. Muine is functioning fine
> > > while the tagging software is operating incorrectly. Unless your tags
> > > are formatted with ID3v2.4 (Most taggers use 2.3), your tags can't be in
> > > the UTF-8 format. The earlier versions of ID3v2 only supported ISO-8859-
> > > 1 and UTF-16. Here is probably what happened. You entered your UTF
> > > characters into your tagging software. Your tagging software then
> > > (incorrectly) stored the frame with the ISO-8859-1 flag rather than
> > > converting it to UTF-16 and storing it with the UTF-16 flag. When
> > > Muine's metadata function loaded, it recognized that there was a ISO-
> > > 8859-1 frame and converted it to UTF-8 for rendering, thus causing the
> > > problem. If this is the case, a bug should be filed against the tagging
> > > software.
> > 
> > It's unlikely, since both id3v2 and grip, the program I originally used
> > to rip my stuff, use id3lib (http://www.id3lib.org), which is pretty
> > standards-compliant.  But in case it's really the tagging software's
> > fault, do you know how I can check for these flags?
> 
> Sure, Take a look at tagged file in a hex editor like ghex.
> 
> The ID3 structure is pretty simple. You're going to want to search for a
> string TIT2 (maybe TIT1). It will be followed by 6 bytes you can ignore,
> then one byte which is the encoding byte.
> 
> If set to 0x00, it is ISO-8859-1 and must not contain UTF characters
> information.
> If set to 0x01 or 0x02 it should be a form of UTF-16.
> If set to 0x03, it is indeed UTF-8 and Muine's ID3 decoder is in error.

You were right.  It was set to 0x00.  Changing it to 0x03 makes it work
flawlessly.

Any suggestions as to how I might do this change to all of my songs at
once, or at least name a program which does proper encoding checks?

Thanks,
Carlos.