Re: [Banshee-List] GSoC idea, finding duplicates

["gnomeuser gmail com" <gnomeuser gmail com>]

2012/3/13 Diego Fernandez <aiguo fernandez gmail com>:
> On Mon, Mar 12, 2012 at 5:15 PM, Samuel Gyger <samuel gyger at> wrote:
>> There is a similar Addon already in the Extension Repository, but…
>>
>> On Mon, Mar 12, 2012 at 21:15, Diego Fernandez
>> <aiguo fernandez gmail com> wrote:
>>> I've never done any work on a real project before, but now that I'm
>>> taking Software Engineering at school and I've learned more about
>>> developing in large projects, I feel it's time to finally start
>>> contributing to Open Source.
>> Try to fix some simple Bugs (at least one), just to find out if you
>> want to dig into this code base. At least my tip, and I hope you're
>> comfortable with C#.
>>
> True, I've already downloaded the code and started looking at it on MonoDevelop.
>
>>> So I have a fairly large music collection with about 22,000 songs.
>>> However, I have LPs, EPs, and Greatest Hits albums for many artists.
>>> This causes a lot of duplicate songs in my library, and with so many
>>> songs it's very hard to check manually.
>> For me there is always a problem that I also like to have complete
>> Albums, I've for example part of the discography of Johnny Cash and
>> there are a lot of duplicates over the years, but some a rerecording
>> etc. and some are even the same, but an album with a missing track
>> (after deleting duplicates) is also a little bit a problem. I don't
>> know if there is a way to solve this, or storage space is just to
>> cheap nowadays and don't care about this.
>> But what really makes me crazy is that if I add a song a second time
>> to the ipod it's copied there a second time, even if it's exactly the
>> same. At least this happens all the time at my ipod. This would be
>> something to tackle, but it's a bug and not whole gsoc project.
>>
> That's understandable, however I still think a good duplicate finder
> can be very useful for many cases.  If using fingerprinting, it's
> especially helpful for people with improperly tagged files.  You have
> to keep in mind there's also lots of people with individual songs
> downloaded from different sources who might want to clean up their
> library a bit.
>
> However, now that you mention it, the search could be expanded to
> include a domain for the search.  For example, one could search for
> duplicates within the whole library, specific folders, playlists, or
> external devices.  This could be more helpful as one might want to
> keep their library with full albums, but not want to waste space on a
> mp3 player with duplicate songs.
>
>>> The simplest, and fastest, is by checking
>>> things such as artist name, song title, and song length.
>> This is done by the extension at the current state, although not
>> always satisfying.
>>
> I hadn't installed the community plug-ins, but now that I have tried
> it I must say I didn't really get anywhere with it.
>
>>> A better, but longer, way would be using fingerprinting.  I've been
>>> looking at the Echoprint and Chromaprint (with Acoustid) which could
>>> be possible candidates.  Since these are also integrated with
>>> MusicBrainz, it seems like other possibilities could open up for
>>> Banshee - MB integration.
>> Definitely something to look into. But try to get in Contact also with
>> the Extension Author.

You might benefit from looking at the last.fm fingerprinter extension
and mirage for that kind of duplicate searching (in fact I recall that
mirage once did this but it is very CPU and time intensive).

>>
>> What is a little bit disappointing, in the past not so many Gsoc
>> projects have finally made it into Banshee, or at least from my view.
>> Perhaps I just wasn't looking careful enough.
>>
>> Samuel
> That's a shame.  I'd really like to work on this, as I believe that
> fingerprinting can bring a huge benefit to both Banshee and the music
> world in general.  I'd really like to hear back from possible mentors
> to discuss the possibility of this project.  I feel like if I were
> accepted and successful, there would be a lot of work I could continue
> to do with fingerprinting and Banshee.

We are in our defense getting better, the upnp work from last year was
just integrated and shipped. We also have one student lined up with an
exciting voice control project which I would personally love to see go
in as soon as humanly possible.

- David

>> _______________________________________________
>> banshee-list mailing list
>> banshee-list gnome org
>> http://mail.gnome.org/mailman/listinfo/banshee-list  (unsubscribe here)
>
>
>
> --
> Diego Fernandez - 爱国
> _______________________________________________
> banshee-list mailing list
> banshee-list gnome org
> http://mail.gnome.org/mailman/listinfo/banshee-list  (unsubscribe here)