Matthew Aidekman

In the past, I have asked about and discussed adding metadata to 

sound files. No one really came up with much. Among the easy and 

useful things I found out how to do via the included link, I found 

that you can use UNIX (the shell object) to add spotlight attributes 

http://www.matisse.net/OSX/darwin_commands.html

I know I'll be working on a system to derive, store, search, and 

In the past, I have asked about and discussed adding metadata to  
sound files.  No one really came up with much.   Among the easy and  
useful things I found out how to do via the included link, I found  
that you can use UNIX (the shell object) to add spotlight attributes  
to files and run spotlight searches

I know I'll be working on a system to derive, store, search, and  
recall metadata for sound files..


possible-metadata-solution

Hrm. The spotlight metadata store is from my understanding not adding 

metadata to the file, but simply to metadata stores that the OS 

If you add metadata to an AIFF under OS X, move the file to a Windows/ 

Linux/Solaris/OS 9 box, and then back, your metadata doesnt follow 

For what its worth, there are a few really nice front ends to xattr 

http://arstechnica.com/reviews/os/macosx-10.4.ars/7)

manually and create custom smart folders for the Finder.

Ive also used the kick ass open source getID library for PHP to 

handle metadata parsing for websites (its modular, and supports 

almost every container format and codec) (the mostly finished content 

But again the issue is you can add all the metadata you want in your 

system, but it doesnt walk with the file (usually) ... fucking pain...

Hrm. The spotlight metadata store is from my understanding not adding  
metadata to the file, but simply to metadata stores that the OS  
associates with the file.

If you add metadata to an AIFF under OS X, move the file to a Windows/ 
Linux/Solaris/OS 9 box, and then back, your metadata doesnt follow  
the file.

For what its worth, there are a few really nice front ends to xattr  
(http://arstechnica.com/reviews/os/macosx-10.4.ars/7)  to do this  
manually and create custom smart folders for the Finder.

Ive also used the kick ass open source getID library for PHP to  
handle metadata parsing for websites (its modular, and supports  
almost every container format and codec) (the mostly finished content  
management system I developed at http://sharecms.sf.net has the code  
in there that uses it).

But again the issue is you can add all the metadata you want in your  
system, but it doesnt walk with the file (usually) ... fucking pain...

*hangs head in disappointment and shame*


Dont do that! Its a difficult problem and no one has a really good 

solution except for introducing new file formats with room for 

metadata and static fields. If you are building stuff for yourself 

and friends and can share the metadata stores, its perfectly useful. 

Dont do that! Its a difficult problem and no one has a really good  
solution except for introducing new file formats with room for  
metadata and static fields. If you are building stuff for yourself  
and friends and can share the metadata stores, its perfectly useful.  
What did you have in mind?

Why not use something like that mxj SQL databas that was posted by

Why not use something like that mxj SQL databas that was posted by
Nick Rothwell some time ago?

On 12-Jun-2006, at 21:27, matthew aidekman wrote:

> I know I'll be working on a system to derive, store, search, and 

To this reader the obvious (and entirely Mac-centric, with possibly 

only a limited lifetime) approach would be to use Resource Forks.

Specify a custom resource type, add your data to it, and go. Data 

stays with the file on copy, on disk images, there are dozens of 

tested solutions for transferring double-fork files from Mac file 

I know (yes, I really do know) that Apple is encouraging developers 

not to use resource forks, but this is one legacy part of the Classic 

Mac OS that still seems to be supported quite thoroughly on OS X, 

Peter Castine +--> Litter Power & Litter Bundle for Jitter

iCE: Sequencing, Recording & |home | chez nous|

Interface Building for |bei uns | i nostri|

On 12-Jun-2006, at 21:27, matthew aidekman wrote:
> I know I'll be working on a system to derive, store, search, and  
> recall metadata for sound files..

To this reader the obvious (and entirely Mac-centric, with possibly  
only a limited lifetime) approach would be to use Resource Forks.

Specify a custom resource type, add your data to it, and go. Data  
stays with the file on copy, on disk images, there are dozens of  
tested solutions for transferring double-fork files from Mac file  
systems to Unix, MS-DOS, etc.

I know (yes, I really do know) that Apple is encouraging developers  
not to use resource forks, but this is one legacy part of the Classic  
Mac OS that still seems to be supported quite thoroughly on OS X,  
even up to Tiger.

--------------    http://www.bek.no/~pcastine/Litter/    -------------
Peter Castine             +--> Litter Power & Litter Bundle for Jitter

I always liked this resource idea (and still have to get rid of them in 

my externals), but I met so many users complaining about "sound files 

that don't work on my PC" that I wouldn't encourage this anymore...

Although it isn't as convenient as with resources, it is allowed to 

create custom data chuncks in WAVE and AIFF files, as long as their 

identifier don't interfere with the standard ones (the rule is to 

ignore data chunck with an unknow identifier). This is also the reason 

why these formats (and expecially .wav) are so chaotic.

Coding this in C or Java is quite easy. Check RIFF specs.

I always liked this resource idea (and still have to get rid of them in 
my externals), but I met so many users complaining about "sound files 
that don't work on my PC" that I wouldn't encourage this anymore...

Although it isn't as convenient as with resources, it is allowed to 
create custom data chuncks in WAVE and AIFF files, as long as their 
identifier don't interfere with the standard ones (the rule is to 
ignore data chunck with an unknow identifier). This is also the reason 
why these formats (and expecially .wav) are so chaotic.

honestly, the discovery of the spotlight thing rekindled my interest 

in the project. I think I might just use a file by the same name 

with a txt extension. the files ability to be moved around and 

retain their metadata is imperative. I find it odd that mac os x 

appeared to be embracing metadata and in a way rejected it by 

creating its own database. I have a email in the spotlight- 

developers list. I wish I knew more about this stuff. It really 

seems that this would be a big step forward.

As for the wav spec, I'll look into working with it but I've never 

ANYTHING like reading /writing file formats before.

sorry to keep annoying the list with the same old junk.

honestly, the discovery of the spotlight thing rekindled my interest  
in the project.  I think I might just use a file by the same name  
with a txt extension.   the files ability to be moved around and  
retain their metadata is imperative.  I find it odd that mac os x  
appeared to be embracing metadata and in a way rejected it by  
creating its own database.   I have a email in the spotlight- 
developers list. I wish I knew more about this stuff.    It really  
seems that this would be a big step forward.

As for the wav spec,  I'll look into working with it but I've never  
ANYTHING like reading /writing file formats before.

sorry to keep annoying the list with the same old junk.
-matt


Well, the OS X way to keep multiple files together is the Bundle.

To the file system, a Bundle is a subdirectory hierarchy with any 

I don't know that Bundles are any more MS-DOS/Windows-friendly than 

Resource forks, but they work well on OS X.

To the file system, a Bundle is a subdirectory hierarchy with any  
number of files.

I don't know that Bundles are any more MS-DOS/Windows-friendly than  
Resource forks, but they work well on OS X.

bundles are indeed a great idea. I found the apple guide to 

"document packages" apparently, I would need to alter the Info.plist 

for every application that wanted to use these .wavs. unfortunately 

http://developer.apple.com/documentation/CoreFoundation/Conc

CFBundles/Concepts/DocumentBundles.html#//apple_ref/doc/uid/ 20002128

bundles are indeed a great idea.  I found the apple guide to  
"document packages"  apparently, I would need to alter the Info.plist  
for every application that wanted to use these .wavs.   unfortunately  
it doesn't seem that max has one

-matt
reference:
 http://developer.apple.com/documentation/CoreFoundation/Conc eptual/ 
 CFBundles/Concepts/DocumentBundles.html#//apple_ref/doc/uid/ 20002128


i am looking for a way to build a spotlight-like-live-search for 

Am 12.06.2006 um 16:27 schrieb matthew aidekman:

> In the past, I have asked about and discussed adding metadata to 

> sound files. No one really came up with much. Among the easy and 

> useful things I found out how to do via the included link, I found 

> that you can use UNIX (the shell object) to add spotlight 

> attributes to files and run spotlight searches

i am looking for a way to build a spotlight-like-live-search for  
folder content in max.

> thought someone might find this useful
>
> In the past, I have asked about and discussed adding metadata to  
> sound files.  No one really came up with much.   Among the easy and  
> useful things I found out how to do via the included link, I found  
> that you can use UNIX (the shell object) to add spotlight  
> attributes to files and run spotlight searches
>
> http://www.matisse.net/OSX/darwin_commands.html
>
> I know I'll be working on a system to derive, store, search, and  
> recall metadata for sound files..


possible metadata solution