Accented Characters in Tags: iTunes vs FLAC, etc.

[qblack]

I tag my classical music tracks consistently with the Composer field in the following style:

Blitzstein, Marc (1905-1964)
Dvořák, Antonín (1841-1904)

In the second case—Dvořák—I have one or two m4a tracks tagged in iTunes and several more FLAC tracks ripped and tagged from CD with XLD or tagged with the tagging program Yate.

When there are accented characters in the tag MinimServer then lists the iTunes and FLAC files separately. In other words, if I browse the list of composers, Dvořák, Antonín (1841-1904) is shown twice, though the letter rendering looks identical.

A text editor shows me that the accented characters do seem to be treated quite differently within the two types of music files.

Am I doing something wrong (apart from being pedantic!) or is there a fix?

Thanks,

Quentin

[simoncn]

(09-04-2015 17:27)qblack Wrote:  I tag my classical music tracks consistently with the Composer field in the following style:

Blitzstein, Marc (1905-1964)
Dvořák, Antonín (1841-1904)

In the second case—Dvořák—I have one or two m4a tracks tagged in iTunes and several more FLAC tracks ripped and tagged from CD with XLD or tagged with the tagging program Yate.

When there are accented characters in the tag MinimServer then lists the iTunes and FLAC files separately. In other words, if I browse the list of composers, Dvořák, Antonín (1841-1904) is shown twice, though the letter rendering looks identical.

A text editor shows me that the accented characters do seem to be treated quite differently within the two types of music files.

Am I doing something wrong (apart from being pedantic!) or is there a fix?

Thanks,

Quentin

What platform are you using to run MinimServer?

At the start of the MinimServer log, what platform default charset is shown?

Can you attach some screenshots from the text editor to a forum post so that I can see the difference?

[qblack]

Hi Simon,

MinimServer running on Synology NAS

From start of log:
MinimServer 0.8.3 update 64, Copyright © 2012-2015 Simon Nash. All rights reserved.
Java™ SE Runtime Environment (build 1.7.0_75-b13)
Java HotSpot™ Client VM (build 24.75-b04, mixed mode)
Platform default charset is UTF-8
Language setting is 'def'

Files for m4a and flac attached, I trust (new to this).

Quentin

[simoncn]

(09-04-2015 20:21)qblack Wrote:  Hi Simon,

MinimServer running on Synology NAS

From start of log:
MinimServer 0.8.3 update 64, Copyright © 2012-2015 Simon Nash. All rights reserved.
Java™ SE Runtime Environment (build 1.7.0_75-b13)
Java HotSpot™ Client VM (build 24.75-b04, mixed mode)
Platform default charset is UTF-8
Language setting is 'def'

Files for m4a and flac attached, I trust (new to this).

Quentin

I haven't seen encodings like this before. I suspect the text editor might be doing something strange. You might have a different charset setting on the machine that you're using to run the text editor (a Mac?)

Would you be able to upload these two files to my FTP server? If you're doing the upload from a Mac, you need to use the 'ftp' command from a Terminal window. If you have a Windows machine, you can use WinSCP to do the upload. I will PM you the server details.

[qblack]

Hi there,

Files posted to qblack folder (created) on your server as requested.

Quentin

[simoncn]

(09-04-2015 21:14)qblack Wrote:  Hi there,

Files posted to qblack folder (created) on your server as requested.

Quentin

Thanks! The .m4a file is encoded using "normal" UTF-8 accented characters. The encoding in the .flac file uses "combining caron" and "combining acute accent" UTF-8 characters to form the accented characters. How were the tags in the .flac file created?

[qblack]

(09-04-2015 21:28)simoncn Wrote:  

(09-04-2015 21:14)qblack Wrote:  Hi there,

Files posted to qblack folder (created) on your server as requested.

Quentin

Thanks! The .m4a file is encoded using "normal" UTF-8 accented characters. The encoding in the .flac file uses "combining caron" and "combining acute accent" UTF-8 characters to form the accented characters. How were the tags in the .flac file created?

My FLAC tags are initially created with the ripping software XLD. I've tried re-tagging with Yate but the result seems the same. I've also tried the metaflac command-line tool to create the COMPOSER tag, with the same result. Maybe the problem has its origins in the distributed FLAC libraries?

[simoncn]

(11-04-2015 10:38)qblack Wrote:  My FLAC tags are initially created with the ripping software XLD. I've tried re-tagging with Yate but the result seems the same. I've also tried the metaflac command-line tool to create the COMPOSER tag, with the same result. Maybe the problem has its origins in the distributed FLAC libraries?

Did you put the composer value on the metaflac command line? If so, how did you type the accented characters? Were you running metaflac on a Mac?

[qblack]

(11-04-2015 10:50)simoncn Wrote:  

(11-04-2015 10:38)qblack Wrote:  My FLAC tags are initially created with the ripping software XLD. I've tried re-tagging with Yate but the result seems the same. I've also tried the metaflac command-line tool to create the COMPOSER tag, with the same result. Maybe the problem has its origins in the distributed FLAC libraries?

Did you put the composer value on the metaflac command line? If so, how did you type the accented characters? Were you running metaflac on a Mac?

Yes, I'm on a Mac.

I think I've found the source of the problem: the way the Finder handles certain text in the Clipboard.

What I've been doing for some time (and this was to ensure consistency) is to press Cmd-C (copy) on a folder in the Finder which corresponds to the required Composer. It's there from discs already ripped. I then paste into the tag field (either in XLD, Yate or my own Applescript). In each case, the pasted name looks correct.

I have tested this now and see that it is only the above method that generates the 'combination' accent forms you describe above. I can make iTunes generate the same form in its m4a files by pasting a "folder" from the Finder with the appropriate diacritics in its name into a tag field.

I can avoid the problem by inserting the desired characters into a field the long way by using the Character Viewer or (in the case of simple accents) using the Mac's press-and-hold method.

The reason we've been looking at Dvořák is because the ř is more finicky to insert via the Character Viewer, so I was copying and pasting from the Finder. But this shortcut made an invisible change to the name.

Thanks for pointing me in the right direction and showing me that I was, indeed, "doing something wrong"!

Quentin

PS Minimserver rocks!

[simoncn]

(11-04-2015 12:53)qblack Wrote:  Yes, I'm on a Mac.

I think I've found the source of the problem: the way the Finder handles certain text in the Clipboard.

What I've been doing for some time (and this was to ensure consistency) is to press Cmd-C (copy) on a folder in the Finder which corresponds to the required Composer. It's there from discs already ripped. I then paste into the tag field (either in XLD, Yate or my own Applescript). In each case, the pasted name looks correct.

I have tested this now and see that it is only the above method that generates the 'combination' accent forms you describe above. I can make iTunes generate the same form in its m4a files by pasting a "folder" from the Finder with the appropriate diacritics in its name into a tag field.

I can avoid the problem by inserting the desired characters into a field the long way by using the Character Viewer or (in the case of simple accents) using the Mac's press-and-hold method.

The reason we've been looking at Dvořák is because the ř is more finicky to insert via the Character Viewer, so I was copying and pasting from the Finder. But this shortcut made an invisible change to the name.

Thanks for pointing me in the right direction and showing me that I was, indeed, "doing something wrong"!

Quentin

PS Minimserver rocks!

I'm glad to hear the mystery is solved. I have also learnt something about Unicode "combining" characters.