on 2013-08-10 at 19:39 Machina wrote:

> I was looking up the state-of-the-art pitch detectors online and found
> TARTINI:
> 
> http://miracle.otago.ac.nz/tartini/

looks very promising indeed. unfortunately i couldn't compile it. i guess
the packaging is not mature for a wide release. the Makefile has some
hard-coded paths that would only work in the author's system, i had to
modify it by hand to make it find some libraries. even then compilation
failed. but then again IANAP and i didn't even try very hard, someone more
knowledgeable than me will doubtlessly succeed. there are also windows and
mac binaries, perhaps those work.

adapting the algorithm to csound opcodes would be great, the code is GPL
and the algorithm is well documented.

ciao!


lj

did you try plltrack? That's a pretty recent (and decent) algorithm.
On 10 Aug 2013, at 13:34, Machina wrote:

> Hi
> 
> i know CSound has some pitch detectors. i tried all of them, with AMDF performing best but not good enough IMO. 
> I was looking up the state-of-the-art pitch detectors online and found TARTINI:

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie

On 10/08/13 14:50, luis jure wrote:
> on 2013-08-10 at 19:39 Machina wrote:
>
>> I was looking up the state-of-the-art pitch detectors online and found
>> TARTINI:
>>
>> http://miracle.otago.ac.nz/tartini/
> looks very promising indeed. unfortunately i couldn't compile it. i guess
> the packaging is not mature for a wide release. the Makefile has some
> hard-coded paths that would only work in the author's system, i had to
> modify it by hand to make it find some libraries. even then compilation
> failed. but then again IANAP and i didn't even try very hard, someone more
> knowledgeable than me will doubtlessly succeed. there are also windows and
> mac binaries, perhaps those work.
I tried compiling it, and got some way. I'm good enough at C++ to get
through most of the errors. Unfortunately I then hit a Qt/Qtw error that
I couldn't solve, namely:

.moc/moc_ledindicator.cpp:14:2: error: #error "This file was generated
using the moc from 4.2.3. It"
.moc/moc_ledindicator.cpp:15:2: error: #error "cannot be used with the
include files from this version of Qt."
.moc/moc_ledindicator.cpp:16:2: error: #error "(The moc has changed too
much.)"

That's meaningless to me :-(

I might have a go at figuring where the algorithm is in amongst the gui
code, and see if I can extract it,

regards
Alan

-- 
Alan Fitch

On 11/08/13 16:27, Andres Cabrera wrote:
> It sounds like the sources contain moc files which should be generated
> by the build system, not be included in the sources as they depend on
> the version of Qt you are using. You would need to get Qt 4.2.3 and
> make it the default to get this to work. But it might be better just
> to find the dsp and rewrap that.
>
Thanks Andres,

 I guess I could carry on asking tedious questions one-by-one such as
"what is a dsp" and "how do I rewrap it" - but to save your time, it's
probably better if you can recommend to me an idiot's guide to Qt :-)

kind regards
Alan

-- 
Alan Fitch

On 11/08/13 16:27, Andres Cabrera wrote:
> It sounds like the sources contain moc files which should be generated
> by the build system, not be included in the sources as they depend on
> the version of Qt you are using. You would need to get Qt 4.2.3 and
> make it the default to get this to work. But it might be better just
> to find the dsp and rewrap that.
>
> Cheers,
> Andrés
>
>

I figured it out and I've managed to build tartini (on Fedora 18).

I'll have to write up what I did. I've no idea if it's working as I
don't have a microphone to hand, though there may be a .wav file provided.

Watch this space...

regards
Alan

P.S. The clue to the .moc files was simply to do
make mocclean
make


-- 
Alan Fitch

On 11/08/13 22:23, Alan Peter Fitch wrote:

I've got tartini to load, however if I try and record any sound it crashes.

The backtrace says

#0  0x0000000000459598 in MyTransforms::calculateAnalysisData
(this=0x1991118, chunk=5, ch=0x1c84c60) at general/mytransforms.cpp:497
#1  0x0000000000421be3 in Channel::processNewChunk (this=0x1c84c60,
filterState=0x7fffe1d25c90) at sound/channel.cpp:334
#2  0x000000000042ada6 in SoundFile::finishRecordChunk (this=0x19910a0,
n=1024) at sound/soundfile.cpp:540
#3  0x0000000000419ac3 in AudioThread::doStuff
(this=this@entry=0x920a28) at sound/audio_thread.cpp:176
#4  0x0000000000419cb5 in AudioThread::run (this=0x920a28) at
sound/audio_thread.cpp:106
#5  0x00007ffff764697c in QThreadPrivate::start(void*) () from
/lib64/libQtCore.so.4
#6  0x000000307b607d15 in start_thread (arg=0x7fffe1d26700) at
pthread_create.c:308
#7  0x000000307b2f253d in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:114


Does that mean anything to anyone? I'm running on an Intel Core i5 64
bit processor, fedora 18.

Thanks for any clues,

Alan

-- 
Alan Fitch