[Home] MediaTags specifications clarification

This was converted from e-mail / html to markdown with html2text.py

Speed, Completeness, and Identification are the three most important qualities of this application.


which will use MediaTags on 30,000+ files.


Ex: "@art": "Az48tks9cC...."

* --extract-binary-tags-to=path/to/attachments-dir -- placed into the

folder and referenced in JSON

Ex: "@art": "./path/to/attachments-dir/my song.m4a.@art.jpeg"

* Binary tag extraction will usually be a post-processing feature and

should be off by default

Identification - Checksums of the "stream" part of a file


TagLib - Let's use taglib if

Mutagen - Probably not a good fit

Libexiv2 - yes

Exiftool - probably not a good fit

Type Detection - GNU file is too slow!

What does Unix Filter Class mean?


While I was waiting my friend created prototypes for outputting mp3 and m4a media metadata which I am using for now.

The most important thing that I need right now is to be able to checksum the data stream.

It's okay to rearrange some of the other things if it's better for your workers' workflow,

but I would like the checksum-ing ability first.

Once the --literal-tags is done I'll know better what the --normalized-tags should look like

  1. Stream (not file) checksums -- { "stream": { "sha256sum": "ae68f......" } }
    1. jpegtags --without-metadata --with-sha256sum ./my-file.jpeg
    2. mp3tags --without-metadata --with-sha256sum ./my-file.mp3
    3. aactags --without-metadata --with-sha256sum ./my-file.m4a
  2. JPEG Media metadata --literal-tags
    1. exivtags ./my-file.jpeg
    2. xmptags ./my-file.jpeg
    3. iptctags ./my-file.jpeg
  3. Stream metadata
    1. jpegtags
    2. aactags
    3. mp3tags
  4. Media --literal-tags
    1. m4atags
    2. id3tags
  5. Media --verbose-tags
    1. m4atags
    2. id3tags
    3. exivtags
    4. xmptags
    5. iptctags
  6. eBook/pdf tags
    1. more information about what information is stored and can be extracted is needed

Before --normalized-tags I first want to see the outputs of the stream and meta-data --literal-tags

I've pushed --binary-tags to be a future consideration

General Clarifications

Meta-data organization

I want to make it clear that there are three types of meta data that I am particularly interested in.

Media (tag) metadata

Stream (data) metadata

File (data + tag) metadata - not necessary to analyze at this time





**The most important parts of the organization are this

A possible organization

**Future Considerations


Ideas to consider, but not to implement yet.

Binary Tags

**Streaming Input


blog comments powered by Disqus Updated at 2010-11-30