[mb-devel] Future-proof FP design sketch

[Peter Brooks]

On 20/05/07, Juha Heljoranta <juha.heljoranta at iki.fi> wrote:
>
> Hi,
>
> The test results which I mentioned earlier where fine. Except for the
> codebook which is pretty darn difficult to get right.
>
> I also decided to rewrite it in Java (mild compromise between R and C).
> The port to C or some other faster language will happen after we have
> get it right.
>
The source code and binaries (they're in same .zip) are available from
> http://www.fsfe.org/en/fellows/juha/fpfpf


Good to see some nice development, I will look into your code after next
week (exams end of) and talk about it then.


Crash course:
>
> "The utility reads 16 bit 8192 Hz signed audio data (little endian) from
> stdin and prints audio fingerprint in intervals of 512 sample (62.5ms)
> to stdout."
>
> $ # process three second audio clip and print the finger print
> $ sox sample.wav -t raw -r 8192 -s -w -c 1 sample.dat avg resample \
> -ql bandreject 0 220
> $ java -jar dist/lib/Fpff-20070520.jar < sample.dat
> QQQQQmAAAIdIIBQQQQQQQQIcAAAZdd
>
>
> See the README file in the archive for details.
>
> All comments/suggestions/questions/etc are welcome. Patches are even
> more welcome :D
>
>
> Cheers,
> Juha
>
>
> Peter Brooks wrote:
> > Good to hear that there is some active development on this, any work i
> > do in this will enhance my degree anyway. I am also idling on irc with
> > the nick theneb but very little time for idle chit chat until my level
> > 2 exams are over. (end of next week)   The current plan is that for
> > the computer science project I am undertaking next year I will have
> > access to a research server with full matlab access, something i'm
> > very much looking forward to. Anyway I'll get more into all this
> > within the next few weeks.
> >
> > On 14/05/07, Juha Heljoranta <juha.heljoranta at iki.fi> wrote:
> >> Hi,
> >>
> >> Welcome! Glad to see that people are interested in this :)
> >>
> >> I'm currently working with the finger print function it self. I still
> >> have to perform few tests to see how my current implementation works.
> >> The implementation is very similar to one described in wiki.
> >>
> >> I'll let you know the test results when I get them (hopefully in few
> >> days:).
> >>
> >> I will also put the code available after I have cleaned it a bit.
>
> _______________________________________________
> MusicBrainz-devel mailing list
> MusicBrainz-devel at lists.musicbrainz.org
> http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.musicbrainz.org/pipermail/musicbrainz-devel/attachments/20070520/9ec38fdf/attachment.htm