CNMAT, SDIF and SPEAR
I am doing some analysis based synthesis and I am studying CNMAT's SDIF
objects. Michael Klingbeil's SPEAR saves analysis data in several SDIF
formats. I am wondering which to use. Does anyone have a patch that
illustrates the basics of this?
Cheers
Gary Lee Nelson
Oberlin College
www.timara.oberlin.edu/GaryLeeNelson
The CNMAT spectral tutorials (highly recommended) have answered most of my
questions. I have been able to move analysis files from SPEAR to MSP with a
few problems that I think I can work out.
On 11/10/06 6:49 PM, "Gary Lee Nelson" wrote:
> I am doing some analysis based synthesis and I am studying CNMAT's SDIF
> objects. Michael Klingbeil's SPEAR saves analysis data in several SDIF
> formats. I am wondering which to use. Does anyone have a patch that
> illustrates the basics of this?
>
> Cheers
> Gary Lee Nelson
> Oberlin College
> www.timara.oberlin.edu/GaryLeeNelson
>
>
Hi Gary-
Thanks for the recommendation.
I don't have SPEAR in front of me, but I'm looking at the ICMC paper about it. It outputs SDIF files with either 1TRC or RBEP frames, right? So, 1TRC is what I've used in the tutorials. RBEP is more Loris style, but should work with sinusoids~ in bandwidth enhanced mode (in the help file).
I'll put together a tutorial using RBEP frames and sinusoids~ bwe. Basically, there's one more column of data to be handled.
It would be possible, pretty cool, for SPEAR to write both analysis types in different streams of the *same* SDIF file. I wonder if I can convince the developer to offer that.
mz
Quote: Gary Lee Nelson wrote on Sat, 11 November 2006 12:29
----------------------------------------------------
> The CNMAT spectral tutorials (highly recommended) have answered most of my
> questions. I have been able to move analysis files from SPEAR to MSP with a
> few problems that I think I can work out.
>
>
> On 11/10/06 6:49 PM, "Gary Lee Nelson" wrote:
>
> > I am doing some analysis based synthesis and I am studying CNMAT's SDIF
> > objects. Michael Klingbeil's SPEAR saves analysis data in several SDIF
> > formats. I am wondering which to use. Does anyone have a patch that
> > illustrates the basics of this?
> >
> > Cheers
> > Gary Lee Nelson
> > Oberlin College
> > www.timara.oberlin.edu/GaryLeeNelson
> >
> >
>
>
>
>
----------------------------------------------------
Yes, 1TRC is what I am using now. I have been trying to find a description
of the various types of SDF files I download from CNMAT. Some say simply
.SDIF but some say res.sdif and some loud.sdif. Are these distinctions
described somewhere?
On 11/11/06 5:59 PM, "mzed" wrote:
>
> Hi Gary-
>
> Thanks for the recommendation.
>
> I don't have SPEAR in front of me, but I'm looking at the ICMC paper about it.
> It outputs SDIF files with either 1TRC or RBEP frames, right? So, 1TRC is
> what I've used in the tutorials. RBEP is more Loris style, but should work
> with sinusoids~ in bandwidth enhanced mode (in the help file).
>
> I'll put together a tutorial using RBEP frames and sinusoids~ bwe. Basically,
> there's one more column of data to be handled.
>
> It would be possible, pretty cool, for SPEAR to write both analysis types in
> different streams of the *same* SDIF file. I wonder if I can convince the
> developer to offer that.
>
> mz
>
>
>
> Quote: Gary Lee Nelson wrote on Sat, 11 November 2006 12:29
> ----------------------------------------------------
>> The CNMAT spectral tutorials (highly recommended) have answered most of my
>> questions. I have been able to move analysis files from SPEAR to MSP with a
>> few problems that I think I can work out.
>>
>>
>> On 11/10/06 6:49 PM, "Gary Lee Nelson" wrote:
>>
>>> I am doing some analysis based synthesis and I am studying CNMAT's SDIF
>>> objects. Michael Klingbeil's SPEAR saves analysis data in several SDIF
>>> formats. I am wondering which to use. Does anyone have a patch that
>>> illustrates the basics of this?
>>>
>>> Cheers
>>> Gary Lee Nelson
>>> Oberlin College
>>> www.timara.oberlin.edu/GaryLeeNelson
>>>
>>>
>>
>>
>>
>>
> ----------------------------------------------------
>
>
> --
> || michael f. zbyszynski -- molecular gastronimist
> || mzed@cnmat.berkeley.edu -- +1.510.643.9990x314
> || http://www.cnmat.berkeley.edu/
> || http://www.piratelemur.com/
>
>
>
Cheers
Gary Lee Nelson
Oberlin College
www.timara.oberlin.edu/GaryLeeNelson
Quote: Gary Lee Nelson wrote on Sun, 12 November 2006 17:42
----------------------------------------------------
> Yes, 1TRC is what I am using now. I have been trying to find a description
> of the various types of SDF files I download from CNMAT. Some say simply
> .SDIF but some say res.sdif and some loud.sdif. Are these distinctions
> described somewhere?
>
I admit, it's a little disorganized.
Anything that says xxxxx.res.sdif should be a resonant model, ie send the data to resonators~. I should make a tutorial around that.
Anything that is xxxx.f0.sdif is a fundamental estimate. I forget what frame type, but it's not going to be very interesting for resynthesis, anyway.
The others are... probably 1TRC's. That is, additive partial tracks. Here's a short reference:
Unfortunately, IRCAM and CNMAT have trouble seeing eye-to-eye on the future of SDIF. RBEP is another frametype that had a short life because IRCAM didn't buy into it. My recent workarounds to accommodate AudioSculpt SDIF files is also a symptom of our non-cooperation on this venture. Maybe Xavier Rodet can help explain it to us.
mz
> Anything that says xxxxx.res.sdif should be a resonant model, ie send the data to resonators~. I should make a tutorial around that.
>
> Anything that is xxxx.f0.sdif is a fundamental estimate. I forget what frame type, but it's not going to be very interesting for resynthesis, anyway.
>
> The others are... probably 1TRC's. That is, additive partial tracks. Here's a short reference:
what i find most difficult is to make resonance models
fitting into resonators~ (7-bit limit!) with spear,
because spear lacks a "high and low limit" for bands;
so that you could choose to analyse only between 500-2500 Hz, but still have 128 frequency bands.
>
> what i find most difficult is to make resonance models
> fitting into resonators~ (7-bit limit!) with spear,
> because spear lacks a "high and low limit" for bands;
> so that you could choose to analyse only between 500-2500 Hz, but still have 128 frequency bands.
I'm synthesizing a model with almost 700 frequency bands here... are you sure you're using the current version (1.997) of resonators? The limit was increased to 1024 in version 1.995.
goto: http://cnmat.berkeley.edu/MAX/downloads/ for the current version.
mz
>> I'm synthesizing a model with almost 700 frequency bands here... are you sure you're using the current version (1.997) of resonators? The limit was increased to 1024 in version 1.995.
nice to hear - but i am using it on OS 9 - and in
conjunction with coll.
i am ok with 128 bands - its just difficult to create
such in SPEAR! (when the source file has complex
harmonics but you dont need the stuff under 500 Hz)
Hi Michael,
I tried downloading all the CNMAT externals for PPC, but the sit archive
seems to be corrupt. I've updated to latest version of Stuffit expander
and tried downloading several times, but I'm not able to expand. Would
it be possible to save the PPC externals as a .dmg instead of .sit?
Thanks,
Trond
mzed wrote:
>
>> what i find most difficult is to make resonance models
>> fitting into resonators~ (7-bit limit!) with spear,
>> because spear lacks a "high and low limit" for bands;
>> so that you could choose to analyse only between 500-2500 Hz, but still have 128 frequency bands.
>>
>
>
> I'm synthesizing a model with almost 700 frequency bands here... are you sure you're using the current version (1.997) of resonators? The limit was increased to 1024 in version 1.995.
>
> goto: http://cnmat.berkeley.edu/MAX/downloads/ for the current version.
>
> mz
>
> --
> || michael f. zbyszynski -- molecular gastronimist
> || mzed@cnmat.berkeley.edu -- +1.510.643.9990x314
> || http://www.cnmat.berkeley.edu/
> || http://www.piratelemur.com/
>
>
>
>
>
Try version 9 of Stuffit... and avoid version 10 and 11 !
Nicolas
Hi,
it's corrupt for me too.
All the best
Alessandro Fogar
2006/11/16, Trond Lossius :
> Hi Michael,
>
> I tried downloading all the CNMAT externals for PPC, but the sit archive
> seems to be corrupt. I've updated to latest version of Stuffit expander
> and tried downloading several times, but I'm not able to expand. Would
> it be possible to save the PPC externals as a .dmg instead of .sit?
>
> Thanks,
> Trond
>
>
> mzed wrote:
> >
> >> what i find most difficult is to make resonance models
> >> fitting into resonators~ (7-bit limit!) with spear,
> >> because spear lacks a "high and low limit" for bands;
> >> so that you could choose to analyse only between 500-2500 Hz, but still have 128 frequency bands.
> >>
> >
> >
> > I'm synthesizing a model with almost 700 frequency bands here... are you sure you're using the current version (1.997) of resonators? The limit was increased to 1024 in version 1.995.
> >
> > goto: http://cnmat.berkeley.edu/MAX/downloads/ for the current version.
> >
> > mz
> >
> > --
> > || michael f. zbyszynski -- molecular gastronimist
> > || mzed@cnmat.berkeley.edu -- +1.510.643.9990x314
> > || http://www.cnmat.berkeley.edu/
> > || http://www.piratelemur.com/
> >
> >
> >
> >
> >
>
>
Hi Trond,
Have you tried opening up Terminal, cd'ing to the directory you
downloaded the archive to and then doing a 'tar xvzf
CNMAT-Everything-OSX-MachO.tar.gz'? IIRC this worked for me. For me
Stuffit often seems to complain about certain archives, usually it
seems Stuffit is the problem because almost always using unzip or tar
from the terminal correctly decompresses the archive.
On Nov 16, 2006, at 1:13 AM, Trond Lossius wrote:
> Hi Michael,
>
> I tried downloading all the CNMAT externals for PPC, but the sit
> archive seems to be corrupt. I've updated to latest version of Stuffit
> expander and tried downloading several times, but I'm not able to
> expand. Would it be possible to save the PPC externals as a .dmg
> instead of .sit?
>
> Thanks,
> Trond
>
>
> mzed wrote:
>>
>>> what i find most difficult is to make resonance models
>>> fitting into resonators~ (7-bit limit!) with spear, because spear
>>> lacks a "high and low limit" for bands;
>>> so that you could choose to analyse only between 500-2500 Hz, but
>>> still have 128 frequency bands.
>>>
>>
>>
>> I'm synthesizing a model with almost 700 frequency bands here... are
>> you sure you're using the current version (1.997) of resonators? The
>> limit was increased to 1024 in version 1.995.
>>
>> goto: http://cnmat.berkeley.edu/MAX/downloads/ for the current
>> version.
>>
>> mz
>>
>> --
>> || michael f. zbyszynski -- molecular gastronimist
>> || mzed@cnmat.berkeley.edu -- +1.510.643.9990x314
>> || http://www.cnmat.berkeley.edu/
>> || http://www.piratelemur.com/
>>
>>
>>
>>
>>
>
>
Quote: nimar wrote on Thu, 16 November 2006 03:00
----------------------------------------------------
> Try version 9 of Stuffit... and avoid version 10 and 11 !
>
> Nicolas
----------------------------------------------------
yea this file opened without a problem on OS 9, it should
work in classic enviroment too.
(funny that you even have the spear app in it too now!)
man, i hate this stuff it sh*t, really.
just last night i had a dozen of files corrupted in 10.3
which i zipped with aladdin dropzip on OS 9 a few years ago.