Hello,
Yes I know C++/Python works very well together but the problem is that
after we start using Python Mixxx have to start supporting that on
Windows, Mac OS X and Linux (probably *BSD family also). Unix is easy
but Windows is pain. So what I'm trying to say yes tempo estimation
needs work but if we like to get it working in scope of GSOC Python
ain't gonna make it. In future it would be neat to have Python scripting
support along javascript. Sorry I have to be little bit too realistic
here. You can make very good proposal that we can trust it as I don't
know how good you know Mixxx code base. Can madmon's better algorithm
transported to C++?
Audio analysis time improving seems more efficient for GSOC and probably
can be done in 3 months. Please provide more information how you are
going to improve those and if you still are interested you need to come
up with official proposal.
Tuukka
21.03.2016, 17:59, 梁哲源 kirjoitti:
Hi,
The MIREX link is showing that the performance of madmom in beat
tracing task QM, so it might be worth integrating to MIXXX. In fact
there are several audio analysis algorithm in madmom performs very
well in MIREX contest, such as tempo estimation
(http://nema.lis.illinois.edu/nema_out/mirex2015/results/ate/summary.html).
Given the MIREX result I tend to think the madmom library is robust
against QM and possible to write some wrapper for the MIXXX.
I know Python and C++ but I don’t have experience in this kind of
wrapper, but I am very willing to take this chance learning. So can
shape my proposal as the madmom / MIXXX integration? I appreciate for
any advice.
Also, the running time for audio analysis seems to be another thing to
be improved. What is the bottleneck in current MIXXX?
CY
Hello,
I didn't really get what MIREX would provide to Mixxx and madmom look
great but it's written in Python and Mixxx in C++. So if you want to
use that you should write some kind of wrapper/gateway to Mixxx and
thing that would be one GSOC task by itself but there is again lot of
space to improve beat detecting.
What do you find bad about QM vamp-plugin and what other alternatives?
Sincerely,
Tuukka
21.03.2016, 03:19, 梁哲源 kirjoitti:
Hi developers,
I am MS student in Sound and Music Computing in Carnegie Mellon
University.
I wonder if developing or integrating existing beat tracking
algorithm would be a possible proposal?
From the MIREX contest
http://nema.lis.illinois.edu/nema_out/mirex2015/results/abt/smc/summary.html#top I
find there are better
algorithms instead of the CD (QM vamp-plugin). For example, the the
SB algorithms (https://github.com/CPJKU/madmom) outperforms the CD a
lot. Does this idea works for GSoC?
CY
------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
_______________________________________________
Get Mixxx, the #1 Free MP3 DJ Mixing software Today
http://mixxx.org
Mixxx-devel mailing list
Mixxx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mixxx-devel
------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
_______________________________________________
Get Mixxx, the #1 Free MP3 DJ Mixing software Today
http://mixxx.org
Mixxx-devel mailing list
Mixxx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mixxx-devel
