[Clam-devel] Re: TODOs
parumi at iua.upf.edu
Wed Jul 23 10:35:19 PDT 2008
On dt, 2008-07-22 at 14:45 -0400, Han, Yushen wrote:
> To make sure that the synthesizer read the right score, here is the
> syntax to use the synthesizer from the command lines
> ./SimpleOboeSynthesizer <where the SDIF files are> <the text score
> file> testOutput.wav synthesis.xml
> The text score files could be found in testScore.tar, C#D.txt is an
> easy start where OboeDMajorAt5.txt is a challenge.
We need to make testing really easy. So commit (via sending a patch) the
testing scores into the plugin dir. Also make sure that it works well
with the clam-test-data SDIF DB, else send a new one to commit.
> Do you think I should keep his structure or this design when I merge
> our plugins?
> 4. To make it plugin, I think I should be able to get the score from a
> MIDI controller and be able to load the samples in another thread. (Do
> you think the latter is a must?)
> Now I tried to compile the osc plugin but it still did not work.
> g++ -o LibloSource.os -c -fPIC -DCLAM_FLOAT -D_DEBUG -DUSE_PTHREADS=1
> -DUSE_XERCES=1 -DCLAM_USE_XML -DUSE_SNDFILE=1 -DUSE_OGGVORBIS=1
> -DWITH_VORBIS=1 -DUSE_MAD=1 -DWITH_MAD=1 -DUSE_ID3=1 -DUSE_PORTAUDIO=1
> -D__MACOSX_CORE__ -DUSE_RTAUDIO=1 -DUSE_FFTW3=1 -I/usr/local/include
> -I/opt/local/include -I/usr/local/include -I/usr/local/include/CLAM
> In file included from LibloSource.cxx:1:
> LibloSource.hxx:8:19: error: lo/lo.h: No such file or directory
> LibloSource.hxx:95: error: 'lo_arg' has not been declared
> I search my hard drive but there is no lo.h.
> Do I need more dependencies? Thank in advance for your suggestion.
Yes, you need the "liblo" library
> 5. I think a blog instead of the wikipage that I am writing will be very good.
> Do you have any spot for me? If not, I will use my google page to post
> the information for potential testers.
blogger.com (aka blogspot - bought by google) and wordpress.com are
popular services for personal blogs
About the threads question: I'd load all the samples during
configuration phase (in ConcreteConfigure). Maybe in the future we could
allow for on-the-fly sample loading, but definitely not for a first
So I don't see threads necessary here. Or maybe I am missing something?
Please comment if you have a different idea on this.
More information about the clam-devel