I have embedded python as a midi scripting language in my audio
thread, and need to migrate the calls to libpython to a new process
per each audio thread to avoid acquiring python's the global
interpreter lock (GIL). Does anyone know what the theoretical
reliability of basic memory-based IPC mechanisms are like with respect
to realtime audio threads? Will I be causing even bigger performance
problems with just the IPC overhead to make some function calls, or is
it negligible?
Thanks
Hi all,
I'm having trouble getting a working compile of the freeverb3 plugin from the
CMT plugin set. I remember that a while back it had issues with de-normals.
However, I thought those issues got worked out. Furthermore, I was under the
impression that the whole de-normals problem was specific to Intel
processors. (Is that incorrect?)
The output sounds the same as what people described when having denormal
issues. i.e. sounds like a crappy plate reverb, with no control over tail
length.
It used to work, so I'm wondering if this might be an issue to do with using
newer versions of gcc?
-Reuben
Fons Adriaensen:
>
> On Wed, Sep 10, 2008 at 05:02:12PM +0200, Kjetil S. Matheussen wrote:
>
>> jack_capture
>> ============
>> jack_capture is a program for recording soundfiles with jack. Its default
>> operation is to capture whatever sound is going out to your speakers into
>> a file. (But it can do a number of other operations as well...)
>
> Very useful, just one remark: for more than 2 channels
> the default should be WAVEX instead of WAV (easy enough
> if you use libsndfile). WAVs with more than 2 channels
> are technically illegal and refused by some software.
>
Thanks for the input. But do you think ch>2 wavex
files are supported by more software than ch>2 wav
files? Because I haven't even heard of wavex files
before... (wavex files are supported in jack_capture
though, "-f wavex" should do the job, thanks to libsndfile)
I posted this exact message to the Vamp forum, noticed the dates of the
last posts, there, and decided to try here instead.
I'm having a most unusual problem compiling the Vamp-SDK on my machine.
This has actually been an on-going problem for some time now. Looking
over the Vamp forums, I see one other person seems to have successfully
compiled and installed Vamp on his 64-bit Ubuntu install, so I'm
wondering what keeps going wrong here.
The first thing I did was to edit the Makefile to install to /usr
instead of /usr/local. That was simple enough. Then when I run make, I
get the following output (notice the error at the end):
-----
g++ -O2 -Wall -I. -fPIC -c -o vamp-sdk/PluginAdapter.o
vamp-sdk/PluginAdapter.cpp
g++ -O2 -Wall -I. -fPIC -c -o vamp-sdk/RealTime.o vamp-sdk/RealTime.cpp
ar r vamp-sdk/libvamp-sdk.a vamp-sdk/PluginAdapter.o vamp-sdk/RealTime.o
g++ -O2 -Wall -I. -fPIC -c -o vamp-sdk/PluginHostAdapter.o
vamp-sdk/PluginHostAdapter.cpp
g++ -O2 -Wall -I. -fPIC -c -o
vamp-sdk/hostext/PluginBufferingAdapter.o
vamp-sdk/hostext/PluginBufferingAdapter.cpp
g++ -O2 -Wall -I. -fPIC -c -o vamp-sdk/hostext/PluginChannelAdapter.o
vamp-sdk/hostext/PluginChannelAdapter.cpp
g++ -O2 -Wall -I. -fPIC -c -o
vamp-sdk/hostext/PluginInputDomainAdapter.o
vamp-sdk/hostext/PluginInputDomainAdapter.cpp
g++ -O2 -Wall -I. -fPIC -c -o vamp-sdk/hostext/PluginLoader.o
vamp-sdk/hostext/PluginLoader.cpp
g++ -O2 -Wall -I. -fPIC -c -o vamp-sdk/hostext/PluginWrapper.o
vamp-sdk/hostext/PluginWrapper.cpp
ar r vamp-sdk/libvamp-hostsdk.a vamp-sdk/PluginHostAdapter.o
vamp-sdk/hostext/PluginBufferingAdapter.o
vamp-sdk/hostext/PluginChannelAdapter.o
vamp-sdk/hostext/PluginInputDomainAdapter.o
vamp-sdk/hostext/PluginLoader.o vamp-sdk/hostext/PluginWrapper.o
vamp-sdk/RealTime.o
ranlib vamp-sdk/libvamp-sdk.a
ranlib vamp-sdk/libvamp-hostsdk.a
g++ -static-libgcc -shared -Wl,-Bsymbolic -Wl,-soname=libvamp-sdk.so.1
-o vamp-sdk/libvamp-sdk.so vamp-sdk/PluginAdapter.o vamp-sdk/RealTime.o
g++ -static-libgcc -shared -Wl,-Bsymbolic
-Wl,-soname=libvamp-hostsdk.so.2 -o vamp-sdk/libvamp-hostsdk.so
vamp-sdk/PluginHostAdapter.o vamp-sdk/hostext/PluginBufferingAdapter.o
vamp-sdk/hostext/PluginChannelAdapter.o
vamp-sdk/hostext/PluginInputDomainAdapter.o
vamp-sdk/hostext/PluginLoader.o vamp-sdk/hostext/PluginWrapper.o
vamp-sdk/RealTime.o
g++ -O2 -Wall -I. -fPIC -c -o examples/SpectralCentroid.o
examples/SpectralCentroid.cpp
g++ -O2 -Wall -I. -fPIC -c -o examples/PercussionOnsetDetector.o
examples/PercussionOnsetDetector.cpp
g++ -O2 -Wall -I. -fPIC -c -o examples/AmplitudeFollower.o
examples/AmplitudeFollower.cpp
g++ -O2 -Wall -I. -fPIC -c -o examples/ZeroCrossing.o
examples/ZeroCrossing.cpp
g++ -O2 -Wall -I. -fPIC -c -o examples/plugins.o examples/plugins.cpp
g++ -static-libgcc -shared -Wl,-Bsymbolic
-Wl,--version-script=vamp-plugin.map -o examples/vamp-example-plugins.so
examples/SpectralCentroid.o examples/PercussionOnsetDetector.o
examples/AmplitudeFollower.o examples/ZeroCrossing.o examples/plugins.o
vamp-sdk/libvamp-sdk.a /usr/lib/gcc/x86_64-linux-gnu/4.2.3/libstdc++.a
/usr/bin/ld:
/usr/lib/gcc/x86_64-linux-gnu/4.2.3/libstdc++.a(functexcept.o):
relocation R_X86_64_32 against `std::bad_typeid::~bad_typeid()' can not
be used when making a shared object; recompile with -fPIC
/usr/lib/gcc/x86_64-linux-gnu/4.2.3/libstdc++.a: could not read symbols:
Bad value
collect2: ld returned 1 exit status
make: *** [examples/vamp-example-plugins.so] Error 1
-----
Is it trying to tell me that libstdc++ is compiled incorrectly on my
machine? It's the one that comes with every Ubuntu 64-bit system. Any
help would be greatly appreciated. Thank you!
My system:
AMD64 X2 with Kubuntu 64-bit, 2.4.24-19-rt kernel, 2GB DDR-800 RAM,
SATA2 hard disks.
Regards,
Darren Landrum
Hi all,
Does anyone else get clicks which occur continuously after getting a
single xrun in jack? (i.e no further xruns are reported, but the audio is
crackling).
I'm setting up an art installation which has to run for hours on end
without intervention, and this seems to be happening after some time
(running jack in realtime mode). I'm using a stock kernel in OpenSuse -
would it help to use a lowlatency one? The audio doesn't need lowlat - I'm
using a buffer size of 1024.
I don't care about the odd xrun - as long as the audio goes back to normal
afterwards, any help gratefully received. This is running in my own
software so there is more than a chance it's something I'm doing wrong.
cheers,
dave
Hi, sorry for the double post, but maybe a user
knows the answer to this...
What is the right way to call fst_init() ?
I have installed fst-1.9 (tried 1.7 - 1.9), and
it runs standalone and loads vst's OK.
However, from within a simple 'hello world' app,
when I call fst_init(), it causes a seg fault.
The seg fault happens early within fst, in the 'wine_premain'
code, at getModuleHandleA().
I have tried that trick of reserving and initializing 1,000,000
stack bytes (taken from from jfst_reserve_memory()). No luck.
I really would like to get this working. Can you help?
Thanks. Tim.
terminator356 <at> users <dot> sourceforge <dot> net.
(Using a different computer)
On Wed, Sep 10, 2008 at 05:02:12PM +0200, Kjetil S. Matheussen wrote:
> jack_capture
> ============
> jack_capture is a program for recording soundfiles with jack. Its default
> operation is to capture whatever sound is going out to your speakers into
> a file. (But it can do a number of other operations as well...)
Very useful, just one remark: for more than 2 channels
the default should be WAVEX instead of WAV (easy enough
if you use libsndfile). WAVs with more than 2 channels
are technically illegal and refused by some software.
Ciao,
--
FA
Laboratorio di Acustica ed Elettroacustica
Parma, Italia
Wie der Mond heute Nacht aussieht !
Ist es nicht ein seltsames Bild ?
i would like to have solution where i can use any headphones and where
there is no bigger latency then 10ms (at least ;).... all pro
solutions are so f* expensive....
Assuming a 7.1/192K/24bit $3:- chip onboard ... What next?
First I will say, it actually will do the planned send and return
to/from the two pieces of stereo outboard equipment I have (that is not
going anywhere,) as well as do separate headphones and monitors. So I
should be a happy camper ...
Suppose one was to face the real world of playing live again? In my
experience this can be a torturous venture into the realms of cheap
lights and the evil thyristors and diacs of this world, all trying to
make their voices heard through my equipment ...
So, how about turning a single unbalanced stereo jack into a single
balanced mono jack by converting a single mono signal to a stereo signal
with inverted phase, would that be a good idea? If so, I believe this
might work in a simalar way on microphone input as well.
If the above holds and we then have two (instead of one) DAC's working
push/pull on the line, would it be possible to take advantage of this?
What I have in mind here is that, the lower you get in the "bittiness",
the more the systenatic errors of distortion will be apparent.
For starters, how about having one the phases at a level slightly below
the other? This should ideally trigger transisitions between absolute
levels at slightly different times for the two phases, giving us an
extra 6dB of useable headroom.
mvh // Jens M Andreasen
Hello all,
I want to know if is there anybody that know anything about open source
implementation of Speaker Recognition[1]
or what kind of tool i can use to develop this kind of tool on linux?
Regards,
Ste
[1]http://en.wikipedia.org/wiki/Speaker_recognition
--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f
Sponsor:
Cerchi un hotel a Riccione, Rimini o Misano Adriatico ? Visita il sito
rivieraparkhotels.it.
Gli alberghi dei parchi divertimento
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=8262&d=20080910
