Hello,

A spectre is haunting Europe. That spectre is Native Instruments. ; )

When i used to play around with their "Reaktor" modular environment, as part
of it's Granular "suite" was a module called "Pitch Former".

"Pitch Former" imposes a fundamental on any sample you load into it, &
allows playback of it at any arbitrary pitch, with the capacity to "tweak"
the Formant value of the output sound. (Basically creating an effect not to
dissimilar to Vocoding - except you don't need a modulator signal). as well
as apply Time Stretch etc etc - all the "usual" granular "side effects"...

So there is no "transposition" of the sample playback - just the application
of a given "fundamental value" to the samples periodic content, & away you
go. 

Input pitch == 60: 
    get out ANYTHING played back AS Middle C.
    (so a sample singing "DOH RE MI FA" outputs "DOH DOH DOH DOH" & sounds a
bit ill...)

Sometime shortly after all this, i went out & bought myself Computer Music
Tutorial. It says (on page 174...)

"Pitch Syncronous Granular Synthesis (PSGS) is a technique designed for the
generation of tones with one or more formant regions in their spectra" & a
brief multi-stage process is described. "This must be how Pitch Former
works" I thought to myself. 

Anyway, it was a cool module but like much of Reaktor (IMHO) it lacked the
capacity to provide any meaningful precision interfacing with it's sonic
capabilities - (that & the fact there was no way to impose any order of
events with any certainty ..... indeterminacy is fine, but only when i, the
user, say so please!)

But now i have Csound & Python, so if i could do this effect using these 2
environments instead, well, i could pump out top 40 in my spare time & make
millions of dollars to give to Dr Boulanger to spread Linux throughout the
third world ; )

So... has anyone done this in csound? I thought FOG might be a good starting
point - but i didn't have much sucess (including armed with the Csound
book...)

I possibly need to to a pitch track analysis on a sample, & derive some
complex re-windowing based upon it & execute it via some some FOF / FOG type
process? (anyone with Computer Music Tutorial please go & open it to page
174 right about now...)

But, err, shouldn't this be available through one or 2 opcodes maybe? Part
of Partikkel Perhaps? (a "parttikeltrack" perhaps?)

Now, i am yet to study partikkel in detail, but as it's entry in the manual
specifically referes to "time domain" processing, i'm wondering if PSGS is
in fact still "a bridge too far" for the average Csound user - & if so, is
it time to do something about it?

I think i can say with some certainty that if you want to attract more
"bedroom producer types" to csound - this effect may single handedly be the
most potent recruitment tool in your arsenal. 

So, who amongst you knows something about this, & can anyone report having
achieved this process already at all? (by means simple or elaborately
complex..)

& if there is an out the box opcode solution already, i'm sorry I missed it,
but the opportunity to start this thread was just too gooder opportunity to
miss (& i'll blame David W for bringing up postmodernism in the first place
! ; ) ...)

& yes, the reason i started thinking about this was because the new Britney
song totally kicks ass!! (but THAT vocal sound on this occasion sounds a lot
like the reaktor module to me, & not so much like any Vocoder / FFT type
effect - but i've only heard it once, on the radio like....)

How did they achieve the effect the analog way anyway? (there's some old EMS
looking suitcase type synth that does it i'm sure - plus the old "talkbox"
guitar effect...)

(& yes, i got bored with programming Python today...)






-- 
View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14707962.html
Sent from the Csound - General mailing list archive at Nabble.com.

Tim,

what a coincidence! I've just read Microsound by Curtis Roads - it's a fantastic
book, you should take a look at it. Anyway, back to your question: I'm not
sure if there's an opcode that does exactly what you want. But I often implement
granular synthesis techniques with two instruments: one is the "slave" and
generates a single grain, the other is the master instrument that calls the
slave through schedkwhen. It's obviously extremely slow, but also very versatile,
because you can implement all the parameters you want.

Alex


>-- Originalnachricht --
>Date: Wed, 9 Jan 2008 02:00:17 -0800 (PST)
>From: Tim Mortimer 
>To: csound@lists.bath.ac.uk
>Subject: [Csnd]  Pitch Syncronous Granular Synthesis (or does Curtis Roads
> dream of electric Britneys)
>Reply-To: csound@lists.bath.ac.uk
>
>
>
>Hello,
>
>A spectre is haunting Europe. That spectre is Native Instruments. ; )
>
>When i used to play around with their "Reaktor" modular environment, as
part
>of it's Granular "suite" was a module called "Pitch Former".
>
>"Pitch Former" imposes a fundamental on any sample you load into it, &
>allows playback of it at any arbitrary pitch, with the capacity to "tweak"
>the Formant value of the output sound. (Basically creating an effect not
>to
>dissimilar to Vocoding - except you don't need a modulator signal). as well
>as apply Time Stretch etc etc - all the "usual" granular "side effects"...
>
>So there is no "transposition" of the sample playback - just the application
>of a given "fundamental value" to the samples periodic content, & away you
>go.
>
>Input pitch == 60:
>    get out ANYTHING played back AS Middle C.
>    (so a sample singing "DOH RE MI FA" outputs "DOH DOH DOH DOH" & sounds
>a
>bit ill...)
>
>Sometime shortly after all this, i went out & bought myself Computer Music
>Tutorial. It says (on page 174...)
>
>"Pitch Syncronous Granular Synthesis (PSGS) is a technique designed for
the
>generation of tones with one or more formant regions in their spectra" &
>a
>brief multi-stage process is described. "This must be how Pitch Former
>works" I thought to myself.
>
>Anyway, it was a cool module but like much of Reaktor (IMHO) it lacked the
>capacity to provide any meaningful precision interfacing with it's sonic
>capabilities - (that & the fact there was no way to impose any order of
>events with any certainty ..... indeterminacy is fine, but only when i,
the
>user, say so please!)
>
>But now i have Csound & Python, so if i could do this effect using these
>2
>environments instead, well, i could pump out top 40 in my spare time & make
>millions of dollars to give to Dr Boulanger to spread Linux throughout the
>third world ; )
>
>So... has anyone done this in csound? I thought FOG might be a good starting
>point - but i didn't have much sucess (including armed with the Csound
>book...)
>
>I possibly need to to a pitch track analysis on a sample, & derive some
>complex re-windowing based upon it & execute it via some some FOF / FOG
type
>process? (anyone with Computer Music Tutorial please go & open it to page
>174 right about now...)
>
>But, err, shouldn't this be available through one or 2 opcodes maybe? Part
>of Partikkel Perhaps? (a "parttikeltrack" perhaps?)
>
>Now, i am yet to study partikkel in detail, but as it's entry in the manual
>specifically referes to "time domain" processing, i'm wondering if PSGS
is
>in fact still "a bridge too far" for the average Csound user - & if so,
is
>it time to do something about it?
>
>I think i can say with some certainty that if you want to attract more
>"bedroom producer types" to csound - this effect may single handedly be
the
>most potent recruitment tool in your arsenal.
>
>So, who amongst you knows something about this, & can anyone report having
>achieved this process already at all? (by means simple or elaborately
>complex..)
>
>& if there is an out the box opcode solution already, i'm sorry I missed
>it,
>but the opportunity to start this thread was just too gooder opportunity
>to
>miss (& i'll blame David W for bringing up postmodernism in the first place
>! ; ) ...)
>
>& yes, the reason i started thinking about this was because the new Britney
>song totally kicks ass!! (but THAT vocal sound on this occasion sounds a
>lot
>like the reaktor module to me, & not so much like any Vocoder / FFT type
>effect - but i've only heard it once, on the radio like....)
>
>How did they achieve the effect the analog way anyway? (there's some old
>EMS
>looking suitcase type synth that does it i'm sure - plus the old "talkbox"
>guitar effect...)
>
>(& yes, i got bored with programming Python today...)
>
>
>
>
>
>
>--
>View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14707962.html
>Sent from the Csound - General mailing list archive at Nabble.com.
>
>
>
>Send bugs reports to this list.
>To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>csound"


Sunrise ADSL mit automatischer Installation - die einfachste Art, ins schnelle
Internet einzusteigen. Ab CHF 9.90 pro Monat. http://www.sunrise.ch/adsl

& as it happens, that 2 inst style of granular was pretty much what my very
first Csound experiment ever set out to achieve... 

i actually think it MIGHT be possible using "standard existing repetoire" of
opcodes, 

but i honestly think if "pvsnotch" or whatever it is gets its own opcode,
then so should this...... 

"spectrum analysis is performed on each grain?" (in a window that's == grain
size?) (With grains == successive wavecycle lengths?)

"an impulse response is then derived which sets the params on a resynth
filter..."

& then, well, it goes on.... 

"pulse trains delivered into FIR filters at the detected pitch period.."

"The output signal results from the excitation of the pulse train on the
weighted sum of the impulse responses of all the filters" what does that
mean???

& then there's an overlap mechanism....

it can even be used to create a "quasi harmonic part" & a residual noise
compnent....

(De Poli & Piccialli 1991 - Piccialli et al 1992)

it's not a simple process....


Alex Weiss-3 wrote:
> 
> Tim,
> 
> But I often implement
> granular synthesis techniques with two instruments: one is the "slave" and
> generates a single grain, the other is the master instrument that calls
> the
> slave through schedkwhen. It's obviously extremely slow, but also very
> versatile,
> because you can implement all the parameters you want.
> 
> Alex
> 
> 

-- 
View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14710166.html
Sent from the Csound - General mailing list archive at Nabble.com.

Hi Tim,

That sounds awesome, I hope it 'sounds' as good as I imagine it!!

Peiman

On 9 Jan 2008, at 11:36, Tim Mortimer wrote:

>
> & as it happens, that 2 inst style of granular was pretty much what  
> my very
> first Csound experiment ever set out to achieve...
>
> i actually think it MIGHT be possible using "standard existing  
> repetoire" of
> opcodes,
>
> but i honestly think if "pvsnotch" or whatever it is gets its own  
> opcode,
> then so should this......
>
> "spectrum analysis is performed on each grain?" (in a window that's  
> == grain
> size?) (With grains == successive wavecycle lengths?)
>
> "an impulse response is then derived which sets the params on a  
> resynth
> filter..."
>
> & then, well, it goes on....
>
> "pulse trains delivered into FIR filters at the detected pitch  
> period.."
>
> "The output signal results from the excitation of the pulse train  
> on the
> weighted sum of the impulse responses of all the filters" what does  
> that
> mean???
>
> & then there's an overlap mechanism....
>
> it can even be used to create a "quasi harmonic part" & a residual  
> noise
> compnent....
>
> (De Poli & Piccialli 1991 - Piccialli et al 1992)
>
> it's not a simple process....
>
>
> Alex Weiss-3 wrote:
>>
>> Tim,
>>
>> But I often implement
>> granular synthesis techniques with two instruments: one is the  
>> "slave" and
>> generates a single grain, the other is the master instrument that  
>> calls
>> the
>> slave through schedkwhen. It's obviously extremely slow, but also  
>> very
>> versatile,
>> because you can implement all the parameters you want.
>>
>> Alex
>>
>>
>
> -- 
> View this message in context: http://www.nabble.com/Pitch- 
> Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of- 
> electric-Britneys%29-tp14707962p14710166.html
> Sent from the Csound - General mailing list archive at Nabble.com.
>
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body  
> "unsubscribe csound"

Yep piece of me - that's the schizz! ; )

i just dloaded it. I'm not sure it's Reaktor - but there's some pretty
interesting stuff going on in there definately. & it's that backing vocal
i'm talking about. But the Reaktor effect is good to (& i also heard it make
"phase synced grains" & they sounded good too....)

I'm still on 5.06, so haven't tried partikkel yet (doesn't work?). But glad
to hear you think your on top of the PSGS process...

& I'll be happy to help you investigate once i up to 5.08..... but maybe you
have the inspiration to implement an example?

There was some Britney worship on microsound list also i glanced earlier
today... 

-- 
View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14710184.html
Sent from the Csound - General mailing list archive at Nabble.com.

well i think my point that "the kids want it" is QED...

Oeyvind, try this...

http://www.nabble.com/lists-tt14642461.html
-- 
View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14710491.html
Sent from the Csound - General mailing list archive at Nabble.com.

Hi,

I just had a listen to the song on youtube (http://www.youtube.com/ 
watch?v=89oS4SN4mNg)! OK the exaggerated glottal sounds are pretty  
interesting. I think all you would need for that is a top quality  
gate, compressor or limiter (possibly multi-band), to exaggerated the  
grainy micro attacks in the sound (this is then mixed with the  
original signal at varying rations: I can hear a fade in and out at  
least once). I think that particular sound has also been enhanced  
with a harmonic exciter so it really jumps out.

That's the boring option, more interestingly they might have attached  
a contact mic to her throat and recorded that on a parallel track.  
This could very well be the case. I wonder if it has been done before  
(?)

Peiman

On 9 Jan 2008, at 11:35, Oeyvind Brandtsegg wrote:

> Yes, Hah, coincidentally, I've been thinking of modeling some of the
> sounds on the latest Britney Spears record with granular techniques
> too... I'm not sure exactly what you mean by "the new song", but I'm
> totally fascinated with the granular sounding sweep on "Piece of Me"
> (the very first sound on the track, and appearing before every chorus)
> on the Blackout album.
>
> I've also contemplated doing a version of hat I *think* they must have
> done to the vocals in certain places, when they exaggerate the
> characteristic Britney "rasping" (click click click at the start of a
> phrase, seemingly from singular vocal cord exitations, don't know how
> to describe this in precise words actually).
> They exaggerate this effect to the degree of it appearing as
> percussion sometimes. This could be done by some sort of granular
> gating, based upon an analysis of the natural occurrence of such
> rasping sounds in the input audio (original vocal track).
>
> But, back to your question about PSGS:
> I don't have the book (Microsound) in front of me right now, and I
> don't remember all the details, but PSGS is described in Microsound
> along the same lines as you describe it in your post. You could
> definately do this with partikkel (it was designed to be able to do
> every form of time domain granular synthesis, and if it can't do PSGS,
> I've done something wrong). I haven't implemented an example myself
> yet, though. If I remember (the details of PSGS) correctly, I guess
> you would use a pitch tracker opcode of your choice to find the
> fundamental pitch in your input sample and then control the grain rate
> based on the pitch track signal. There's probably a million little
> things to consider, but I *think* this is the basic setup to enable
> PSGS.
>
> Oeyvind
>
> 2008/1/9, Tim Mortimer :
>>
>> Hello,
>>
>> A spectre is haunting Europe. That spectre is Native Instruments. ; )
>>
>> When i used to play around with their "Reaktor" modular  
>> environment, as part
>> of it's Granular "suite" was a module called "Pitch Former".
>>
>> "Pitch Former" imposes a fundamental on any sample you load into  
>> it, &
>> allows playback of it at any arbitrary pitch, with the capacity to  
>> "tweak"
>> the Formant value of the output sound. (Basically creating an  
>> effect not to
>> dissimilar to Vocoding - except you don't need a modulator  
>> signal). as well
>> as apply Time Stretch etc etc - all the "usual" granular "side  
>> effects"...
>>
>> So there is no "transposition" of the sample playback - just the  
>> application
>> of a given "fundamental value" to the samples periodic content, &  
>> away you
>> go.
>>
>> Input pitch == 60:
>>    get out ANYTHING played back AS Middle C.
>>    (so a sample singing "DOH RE MI FA" outputs "DOH DOH DOH DOH" &  
>> sounds a
>> bit ill...)
>>
>> Sometime shortly after all this, i went out & bought myself  
>> Computer Music
>> Tutorial. It says (on page 174...)
>>
>> "Pitch Syncronous Granular Synthesis (PSGS) is a technique  
>> designed for the
>> generation of tones with one or more formant regions in their  
>> spectra" & a
>> brief multi-stage process is described. "This must be how Pitch  
>> Former
>> works" I thought to myself.
>>
>> Anyway, it was a cool module but like much of Reaktor (IMHO) it  
>> lacked the
>> capacity to provide any meaningful precision interfacing with it's  
>> sonic
>> capabilities - (that & the fact there was no way to impose any  
>> order of
>> events with any certainty ..... indeterminacy is fine, but only  
>> when i, the
>> user, say so please!)
>>
>> But now i have Csound & Python, so if i could do this effect using  
>> these 2
>> environments instead, well, i could pump out top 40 in my spare  
>> time & make
>> millions of dollars to give to Dr Boulanger to spread Linux  
>> throughout the
>> third world ; )
>>
>> So... has anyone done this in csound? I thought FOG might be a  
>> good starting
>> point - but i didn't have much sucess (including armed with the  
>> Csound
>> book...)
>>
>> I possibly need to to a pitch track analysis on a sample, & derive  
>> some
>> complex re-windowing based upon it & execute it via some some  
>> FOF / FOG type
>> process? (anyone with Computer Music Tutorial please go & open it  
>> to page
>> 174 right about now...)
>>
>> But, err, shouldn't this be available through one or 2 opcodes  
>> maybe? Part
>> of Partikkel Perhaps? (a "parttikeltrack" perhaps?)
>>
>> Now, i am yet to study partikkel in detail, but as it's entry in  
>> the manual
>> specifically referes to "time domain" processing, i'm wondering if  
>> PSGS is
>> in fact still "a bridge too far" for the average Csound user - &  
>> if so, is
>> it time to do something about it?
>>
>> I think i can say with some certainty that if you want to attract  
>> more
>> "bedroom producer types" to csound - this effect may single  
>> handedly be the
>> most potent recruitment tool in your arsenal.
>>
>> So, who amongst you knows something about this, & can anyone  
>> report having
>> achieved this process already at all? (by means simple or elaborately
>> complex..)
>>
>> & if there is an out the box opcode solution already, i'm sorry I  
>> missed it,
>> but the opportunity to start this thread was just too gooder  
>> opportunity to
>> miss (& i'll blame David W for bringing up postmodernism in the  
>> first place
>> ! ; ) ...)
>>
>> & yes, the reason i started thinking about this was because the  
>> new Britney
>> song totally kicks ass!! (but THAT vocal sound on this occasion  
>> sounds a lot
>> like the reaktor module to me, & not so much like any Vocoder /  
>> FFT type
>> effect - but i've only heard it once, on the radio like....)
>>
>> How did they achieve the effect the analog way anyway? (there's  
>> some old EMS
>> looking suitcase type synth that does it i'm sure - plus the old  
>> "talkbox"
>> guitar effect...)
>>
>> (& yes, i got bored with programming Python today...)
>>
>>
>>
>>
>>
>>
>> --
>> View this message in context: http://www.nabble.com/Pitch- 
>> Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of- 
>> electric-Britneys%29-tp14707962p14707962.html
>> Sent from the Csound - General mailing list archive at Nabble.com.
>>
>>
>>
>> Send bugs reports to this list.
>> To unsubscribe, send email sympa@lists.bath.ac.uk with body  
>> "unsubscribe csound"
>>
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body  
> "unsubscribe csound"

Yes I'm just hearing lots of pulse synth like sweeps, & i think my backing
vocal could well be talk box (i.e. possibly the real thing..)

http://en.wikipedia.org/wiki/Talk_box

how to model such a beast digitally i'm not sure. It has just a distinctive
& "transparent" tone...

rest assured however the "formant granular" approaches i have heard are
synonymous & flexible enough to be interesting & worthy in their own right..

Then there are formant shifting plugin effects too (like the company Waves
make...) - how do they work?

If i could produce all 3 of these efffects with the accuracy of having
csound krate control of parameters i (& many others) i'm sure would be very
happy..


-- 
View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14711322.html
Sent from the Csound - General mailing list archive at Nabble.com.

Now that's taking Britney analysis to new extremes! 

But alas, we are no closer to an implementation....

but this one won't lie down i assure you - will start scouring the manual &
start looking for ANY useful inroads...

But basically i think we're looking at doing PSGS, with a subplot of any
other pitch tracking / repitching / formant shifting options that might
yield similar results...

The "evaporative" side of Granular is of less immediate interest to me, but
i look forward to listening to ANY implementations (or ideas) anyone comes
up with as a result of this thread.....

The culmination of all this will of cousre be my devastating "urban
diaspora" remix of "hello, i'm doctor richard bo-o-o-o-ulanger..."


giucant wrote:
> 
> Hi,
> hearing and watching the spectrogram i think you are
> right, it's a granular tecnique. Picture attached.
> 
> j
> --- Oeyvind Brandtsegg  ha
> scritto:
> 
>> I agree that a lot of the synth textures must be
>> layeres & phased
>> pulse synths, but:
>> Are you sure that very first sound is a pulse synth
>> too ? It sound
>> distinctly granular to me, with the noisy part of it
>> coming from a
>> slightly uneven (nonperiodic) distribution of
>> grains.
>> Oeyvind
>> 
>> 2008/1/9, Tim Mortimer :
>> >
>> > Yes I'm just hearing lots of pulse synth like
>> sweeps, & i think my backing
>> > vocal could well be talk box (i.e. possibly the
>> real thing..)
>> >
>> 
>> 
>> Send bugs reports to this list.
>> To unsubscribe, send email sympa@lists.bath.ac.uk
>> with body "unsubscribe csound"
>> 
> 
> 
> 
>       ___________________________________ 
> L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail:
> http://it.docs.yahoo.com/nowyoucan.html
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>  
> 

-- 
View this message in context: http://www.nabble.com/Pitch-Syncronous-Granular-Synthesis-%28or-does-Curtis-Roads-dream-of-electric-Britneys%29-tp14707962p14726211.html
Sent from the Csound - General mailing list archive at Nabble.com.

Your email is well timed, Oeyvind.  I was just thinking yesterday about
using trees of syncphasor opcodes to set up complex triggering patterns for
grains.  But, I did not know whether a Csound opcode existed that could
spawn individual grains from a trigger input -- I thought I might have to
write one.  So, thanks!  This is great.

Also, another note about the "old approach of generating grains with one of
the sched* opcodes.  The start and end times of Csound note events are
quantized to the k-rate so you will not get sample-accurate timing unless
ksmps = 1.

Anthony
(ashamed to be part of a thread mentioning B.S. :)

Anthony Kozar
mailing-lists-1001 AT anthonykozar DOT net
http://anthonykozar.net/

Oeyvind Brandtsegg wrote on 1/9/08 6:51 AM:

> Just as a side note:
> The "old" approach of a 2-instrument approach for full control of each
> single grain can be replicated with partikkel (in one single
> instrument if you wish), just set grain rate to zero and use the async
> input to trigger creation of indivudual grains.
> Oeyvind