best pitch tracker?

[James Harkins-2]

Not sure if things have changed lately... but... what is the best patch tracker ugen currently available for sc?

I know of Qitch and Tartini by Nick Collins, which I will have to figure out how to compile myself since apparently no one else has done it for Intel mac? (Anyone...?)

Does anyone else have something that is precompiled?

I'm just finding Pitch to be a bit unstable sometimes... most of the time it's OK but sometimes it oscillates between a pair of close frequencies and if that crosses the threshold to quantize it to some temperament, the results are quite bad. I may have no alternative but to filter out the data based on some heuristics but I haven't figured out what the rules should be yet.

hjh

: H. James Harkins

: jamshark70@...

: http://www.dewdrop-world.net

.::!:.:.......:.::........:..!.::.::...:..:...:.:.:.:..:

"Come said the Muse,

Sing me a song no poet has yet chanted,

Sing me the universal."  -- Whitman

_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[jimmy johnson-7]

There is a paper by David Gerhard "pitch extraction and fundamental frequency: history and current techniques" that was very helpful to me in finding the right combination/s recently.  

http://www2.cs.uregina.ca/~gerhard/publications/TRdbg-Pitch.pdf

cheers,

=]immy
________________________________

> To: sc-users@...
> Date: Thu, 28 Feb 2008 09:00:53 -0500
> From: jamshark70@...
> Subject: [sc-users] best pitch tracker?
>
> Not sure if things have changed lately... but... what is the best patch tracker ugen currently available for sc?
>
> I know of Qitch and Tartini by Nick Collins, which I will have to figure out how to compile myself since apparently no one else has done it for Intel mac? (Anyone...?)
>
> Does anyone else have something that is precompiled?
>
> I'm just finding Pitch to be a bit unstable sometimes... most of the time it's OK but sometimes it oscillates between a pair of close frequencies and if that crosses the threshold to quantize it to some temperament, the results are quite bad. I may have no alternative but to filter out the data based on some heuristics but I haven't figured out what the rules should be yet.
>
> hjh
>
>
>
> : H. James Harkins
>
> : jamshark70@...
>
> : http://www.dewdrop-world.net
>
> .::!:.:.......:.::........:..!.::.::...:..:...:.:.:.:..:
>
>
> "Come said the Muse,
>
> Sing me a song no poet has yet chanted,
>
> Sing me the universal."  -- Whitman

_________________________________________________________________
Connect and share in new ways with Windows Live.
http://www.windowslive.com/share.html?ocid=TXT_TAGHM_Wave2_sharelife_012008
_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[Josh Parmenter]

James,

I can compile them tonight -

Nick - how would you feel about adding these to sc3-plugins? If not, I  
can send you UB builds that you can distribute. Let me know.

Best,

Josh

On Feb 28, 2008, at 10:19 AM, jimmy johnson wrote:

>
> There is a paper by David Gerhard "pitch extraction and fundamental  
> frequency: history and current techniques" that was very helpful to  
> me in finding the right combination/s recently.
>
> http://www2.cs.uregina.ca/~gerhard/publications/TRdbg-Pitch.pdf
>
> cheers,
>
> =]immy
> ________________________________
>> To: sc-users@...
>> Date: Thu, 28 Feb 2008 09:00:53 -0500
>> From: jamshark70@...
>> Subject: [sc-users] best pitch tracker?
>>
>> Not sure if things have changed lately... but... what is the best  
>> patch tracker ugen currently available for sc?
>>
>> I know of Qitch and Tartini by Nick Collins, which I will have to  
>> figure out how to compile myself since apparently no one else has  
>> done it for Intel mac? (Anyone...?)
>>
>> Does anyone else have something that is precompiled?
>>
>> I'm just finding Pitch to be a bit unstable sometimes... most of  
>> the time it's OK but sometimes it oscillates between a pair of  
>> close frequencies and if that crosses the threshold to quantize it  
>> to some temperament, the results are quite bad. I may have no  
>> alternative but to filter out the data based on some heuristics but  
>> I haven't figured out what the rules should be yet.
>>
>> hjh
>>
>>
>>
>> : H. James Harkins
>>
>> : jamshark70@...
>>
>> : http://www.dewdrop-world.net
>>
>> .::!:.:.......:.::........:..!.::.::...:..:...:.:.:.:..:
>>
>>
>> "Come said the Muse,
>>
>> Sing me a song no poet has yet chanted,
>>
>> Sing me the universal."  -- Whitman
> _________________________________________________________________
> Connect and share in new ways with Windows Live.
> http://www.windowslive.com/share.html?ocid=TXT_TAGHM_Wave2_sharelife_012008
> _______________________________________________
> sc-users mailing list
> sc-users@...
> http://lists.create.ucsb.edu/mailman/listinfo/sc-users

******************************************
/* Joshua D. Parmenter
http://www.realizedsound.net/josh/

“Every composer – at all times and in all cases – gives his own  
interpretation of how modern society is structured: whether actively  
or passively, consciously or unconsciously, he makes choices in this  
regard. He may be conservative or he may subject himself to continual  
renewal; or he may strive for a revolutionary, historical or social  
palingenesis." - Luigi Nono
*/

_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[James Harkins-2]

Cool, thanks. I should figure out how to do it but time is at a premium...

What I know now is that Pitch.kr isn't clever enough for my needs in
this piece.... I'll check out Qitch and Tartini later, and maybe some
other FFTs.

KeyTrack looks intriguing but I don't need it to identify a
pre-programmed mode... I want the pitch classes to make another kind
of analysis in the language.
hjh

On Thu, Feb 28, 2008 at 12:01 PM, Josh Parmenter <josh@...> wrote:
> James,
>
>  I can compile them tonight -

--
James Harkins /// dewdrop world
jamshark70@...
http://www.dewdrop-world.net

"Come said the Muse,
Sing me a song no poet has yet chanted,
Sing me the universal."  -- Whitman
_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[Click Nilson]

Hi James,

The MachineListening plugins set:

http://www.informatics.sussex.ac.uk/users/nc81/code/MachineListening.zip

comes with an Intel binary?

Contains Concat, Qitch and Tartini (as well as AutoTrack and  
AnalyseEvents2 for bbcut2)

With the recent machine listening upgrade in SC core, I didn't add  
Qitch because of the extra data files needed, and Tartini can't use  
FFT in a straight laced way, non-trivial to port. So these will  
remain third party for now; they rely on fftw3f, but should come  
precompiled with static lib pre-linked. Please tell me if trouble and  
I'll double check on my office MacIntel next week.

best,
Nick

btw, thanks for the offer Josh; but I hope the above should work?


On 28 Feb 2008, at 18:13, James Harkins wrote:

> Cool, thanks. I should figure out how to do it but time is at a  
> premium...
>
> What I know now is that Pitch.kr isn't clever enough for my needs in
> this piece.... I'll check out Qitch and Tartini later, and maybe some
> other FFTs.
>
> KeyTrack looks intriguing but I don't need it to identify a
> pre-programmed mode... I want the pitch classes to make another kind
> of analysis in the language.
> hjh
>
> On Thu, Feb 28, 2008 at 12:01 PM, Josh Parmenter  
> <josh@...> wrote:
>> James,
>>
>>  I can compile them tonight -
>
> --
> James Harkins /// dewdrop world
> jamshark70@...
> http://www.dewdrop-world.net
>
> "Come said the Muse,
> Sing me a song no poet has yet chanted,
> Sing me the universal."  -- Whitman
> _______________________________________________
> sc-users mailing list
> sc-users@...
> http://lists.create.ucsb.edu/mailman/listinfo/sc-users

_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[Oswald Berthold-2]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


hi click,

maybe you want to add this SConstruct to your folder for eased linux building?

bst, opt

[ATTACHMENT /usr/src/supercollider/MachineListening/source/SConstruct, text/plain]


Click Nilson writes:
 > Hi James,
 >
 > The MachineListening plugins set:
 >
 > http://www.informatics.sussex.ac.uk/users/nc81/code/MachineListening.zip
 >
 > comes with an Intel binary?
 >
 > Contains Concat, Qitch and Tartini (as well as AutoTrack and  
 > AnalyseEvents2 for bbcut2)
 >
 > With the recent machine listening upgrade in SC core, I didn't add  
 > Qitch because of the extra data files needed, and Tartini can't use  
 > FFT in a straight laced way, non-trivial to port. So these will  
 > remain third party for now; they rely on fftw3f, but should come  
 > precompiled with static lib pre-linked. Please tell me if trouble and  
 > I'll double check on my office MacIntel next week.
 >
 > best,
 > Nick
 >
 > btw, thanks for the offer Josh; but I hope the above should work?
 >
 >
 > On 28 Feb 2008, at 18:13, James Harkins wrote:
 >
 > > Cool, thanks. I should figure out how to do it but time is at a  
 > > premium...
 > >
 > > What I know now is that Pitch.kr isn't clever enough for my needs in
 > > this piece.... I'll check out Qitch and Tartini later, and maybe some
 > > other FFTs.
 > >
 > > KeyTrack looks intriguing but I don't need it to identify a
 > > pre-programmed mode... I want the pitch classes to make another kind
 > > of analysis in the language.
 > > hjh
 > >
 > > On Thu, Feb 28, 2008 at 12:01 PM, Josh Parmenter  
 > > <josh@...> wrote:
 > >> James,
 > >>
 > >>  I can compile them tonight -
 > >
 > > --
 > > James Harkins /// dewdrop world
 > > jamshark70@...
 > > http://www.dewdrop-world.net
 > >
 > > "Come said the Muse,
 > > Sing me a song no poet has yet chanted,
 > > Sing me the universal."  -- Whitman
 > > _______________________________________________
 > > sc-users mailing list
 > > sc-users@...
 > > http://lists.create.ucsb.edu/mailman/listinfo/sc-users
 >
 > _______________________________________________
 > sc-users mailing list
 > sc-users@...
 > http://lists.create.ucsb.edu/mailman/listinfo/sc-users
 >
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iD8DBQFHx/M+7u+hhWCy2kMRAsz/AKDwuKXKIoEE+WhiBfek25WCjh66+QCfR3li
JAPLv4JlT/iAs5CeI/ihIbA=
=G9Jx
-----END PGP SIGNATURE-----
_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[Oswald Berthold-2]

that didnt work did it?

alora, ancora una volta ..


- --IRRy+U3Vvw
Content-Type: text/plain; name="SConstruct"
Content-Description: sconstruct
Content-Disposition: inline;
        filename="SConstruct"
Content-Transfer-Encoding: 7bit

# scons build script.
# blackrain at realizedsound dot net - 11 2006
# Additions by Andrzej Kopec - akopec at chopin dot edu dot pl - Oct 07 2007
# Adapted for ML by Oswald Berthold - opt -x- xdv org - Feb 2008

# edit this to point to your SuperCollider3 source directory

sc3_source = '/usr/src/supercollider/SuperCollider3/'

##############################################
# simple ugens
headers = sc3_source + 'Headers'

plugs = [
        'AnalyseEvents2fftw',
        'AutoTrackfftw',
        'Concatfftw',
        'MachineListening',
        'Qitchfftw',
        'Tartini'
]

env = Environment(
    CPPPATH = [headers + '/common', headers + '/plugin_interface', headers + '/server'],
    CPPDEFINES = ['SC_LINUX', '_REENTRANT', 'NDEBUG', ('SC_MEMORY_ALIGNMENT', 1)],
    CCFLAGS = ['-Wno-unknown-pragmas'],
    CXXFLAGS =  ['-Wno-deprecated', '-O3'],
    SHLIBPREFIX = '',
    SHLIBSUFFIX = '.so'
)

plugsc = []
for plug in plugs :
        plugsc.append(plug + '.cpp')

env.SharedLibrary('MachineListening', plugsc, LIBS=['fftw3f']);


- --IRRy+U3Vvw
Content-Type: text/plain; charset=us-ascii
Content-Description: message body text
Content-Transfer-Encoding: 7bit



Oswald Berthold writes:
 > -----BEGIN PGP SIGNED MESSAGE-----
 > Hash: SHA1
 >
 >
 > hi click,
 >
 > maybe you want to add this SConstruct to your folder for eased linux building?
 >
 > bst, opt
 >
 > [ATTACHMENT /usr/src/supercollider/MachineListening/source/SConstruct, text/plain]
 >
 >
 > Click Nilson writes:
 >  > Hi James,
 >  >
 >  > The MachineListening plugins set:
 >  >
 >  > http://www.informatics.sussex.ac.uk/users/nc81/code/MachineListening.zip
 >  >
 >  > comes with an Intel binary?
 >  >
 >  > Contains Concat, Qitch and Tartini (as well as AutoTrack and  
 >  > AnalyseEvents2 for bbcut2)
 >  >
 >  > With the recent machine listening upgrade in SC core, I didn't add  
 >  > Qitch because of the extra data files needed, and Tartini can't use  
 >  > FFT in a straight laced way, non-trivial to port. So these will  
 >  > remain third party for now; they rely on fftw3f, but should come  
 >  > precompiled with static lib pre-linked. Please tell me if trouble and  
 >  > I'll double check on my office MacIntel next week.
 >  >
 >  > best,
 >  > Nick
 >  >

- --IRRy+U3Vvw--
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>

iD8DBQFHx/kR7u+hhWCy2kMRAvOUAJ9RbPQVhg7AuaegqknDKhfKSz4tOQCfRYGL
CeNbNRgQtlv32r283oqPErQ=
=ZNaq
-----END PGP SIGNATURE-----
_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[James Harkins-2]

I'm enjoying both Qitch and Tartini. With Qitch I thought it could be made easier to load the kernel buffer, so I moved the existing kernel files into a new folder QitchData (in the same location as Qitch.sc). Then:

+ Qitch {

*loadDataBuffer { |kernelSize = 2048, server|

var path = PathName(this.filenameSymbol.asString).pathOnly;

server ?? { server = Server.default };

if(server.serverRunning.not) {

Error("Server must be booted to load data.").throw;

};

path = "%QitchData/QspeckernN%SR%.wav".format(path, kernelSize, server.sampleRate);

if(File.exists(path).not) {

Error("Kernel file % does not exist.".format(path)).throw;

};

^Buffer.read(server, path)

}

}

d = Qitch.loadDataBuffer;

{ ... Qitch.kr(signal, d, ...) } ...

hjh

: H. James Harkins

: jamshark70@...

: http://www.dewdrop-world.net

.::!:.:.......:.::........:..!.::.::...:..:...:.:.:.:..:

"Come said the Muse,

Sing me a song no poet has yet chanted,

Sing me the universal."  -- Whitman

_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[Tim Walters]

Click Nilson wrote:

> http://www.informatics.sussex.ac.uk/users/nc81/code/MachineListening.zip

Am I missing something, or are the data files for Qitch not in this bundle?

Tartini works great, though!

--

Tim Walters | The Doubtful Palace | http://doubtfulpalace.com
_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users

[Richard Taylor-12]

They're not but you can grab them from the older qitch release. Wish
i'd noticed this thread before, spent a happy afternoon learning how
to write scons build scripts in order to get this lot working in
linux.. oh well guess it was educational :)

On 02/03/2008, Tim Walters <walters@...> wrote:

> Click Nilson wrote:
>
>  > http://www.informatics.sussex.ac.uk/users/nc81/code/MachineListening.zip
>
>
> Am I missing something, or are the data files for Qitch not in this bundle?
>
>  Tartini works great, though!
>
>
>  --
>
>  Tim Walters | The Doubtful Palace | http://doubtfulpalace.com
>
> _______________________________________________
>  sc-users mailing list
>  sc-users@...
>  http://lists.create.ucsb.edu/mailman/listinfo/sc-users
>

_______________________________________________
sc-users mailing list
sc-users@...
http://lists.create.ucsb.edu/mailman/listinfo/sc-users