hi all -

the third release of the csound floss manual is out at
www.flossmanuals.net/csound

there is a lot of new stuff in it. see below the "what's new" in detail. 
thanks goes to all contributors, and in particular to alexandre abrioux 
for his diligent proof reading.

all the csd example files (nearly 250 now) and audio samples can be 
downloaded here:
http://csound-tutorial.net/filebrowser/download/576
they will also be included in the next release of csoundqt.

alex hofmann will push a printed version at lulu.com in a short time.

please let us know any bugs, errata and suggestions.

enjoy -

	joachim and iain



WHAT'S NEW IN THIS RELEASE

     New chapters:
         03E ARRAYS (Tarmo Johannes, Joachim Heintz)
         04H SCANNED SYNTHESIS (Christopher Saunders)
         08B CSOUND AND ARDUINO (Iain McCurdy)
         12B PYTHON INSIDE CSOUND (Andrés Cabrera, Joachim Heintz)
         12C PYTHON IN CSOUNDQT (Tarmo Johannes, Joachim Heintz)

     Revised chapters:
         02A MAKE CSOUND RUN: Updated section about Windows install (Jim 
Aikin) and new sections about Csound on Android and iOS (Jacques Laplat)
         03A INITIALIZATION AND PERFORMANCE PASS has completely been 
rewritten (Joachim Heintz)
         04A ADDITIVE SYNTHESIS has been expanded (Iain McCurdy, Bjørn 
Houdorf)
         05B PANNING AND SPATIALIZATION now contains descriptions about 
multi-channel audio in Csound in general, and VBAP and Ambisonics in 
particular (Iain McCurdy, Joachim Heintz)
         10A CSOUNDQT now contains a description of the options and 
choices in CsoundQt's Configure Panel (Peiman Khosravi, Joachim Heintz)
         10D CABBAGE has been updated and covers now some of the 
exciting new developments (Rory Walsh)
         12A THE CSOUND API has been revised and extended (Francois Pinot)
         The OPCODE GUIDE has been updated (Iain McCurdy)
         The METHODS OF WRITING CSOUND SCORES now contain a description 
of Pysco (Jacob Joaquin)

     General additions and changes:
         The code examples now also carry some (hopefully) meaningful 
names in addition to the numbers.
         Many improvements to existing examples have been made by Iain 
McCurdy.
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user

This is great news, Joachim and Iain and others, good work!
I will buy the paper version when available - something to read in the
holidays :)

Thanks a lot!




--
View this message in context: http://csound.1045644.n5.nabble.com/3rd-release-of-csound-floss-manual-released-tp5722755p5722757.html
Sent from the Csound - General mailing list archive at Nabble.com.

In chapter 12B on Python, there is a limitation that the probabilities of a table for a Markov chain must equal 1. Stack Overflow contains a post that solves this problem:

def weighted_choice(choices):
   total = sum(w for c, w in choices)
   r = random.uniform(0, total)
   upto = 0
   for c, w in choices:
      if upto + w > r:
         return c
      upto += w
   assert False, "Shouldn't get here"

On Wed, May 1, 2013 at 3:31 AM, joachim heintz <jh@joachimheintz.de> wrote:

hi all -

the third release of the csound floss manual is out at
www.flossmanuals.net/csound

there is a lot of new stuff in it. see below the "what's new" in detail. thanks goes to all contributors, and in particular to alexandre abrioux for his diligent proof reading.

all the csd example files (nearly 250 now) and audio samples can be downloaded here:
http://csound-tutorial.net/filebrowser/download/576
they will also be included in the next release of csoundqt.

alex hofmann will push a printed version at lulu.com in a short time.

please let us know any bugs, errata and suggestions.

enjoy -

        joachim and iain



WHAT'S NEW IN THIS RELEASE

    New chapters:
        03E ARRAYS (Tarmo Johannes, Joachim Heintz)
        04H SCANNED SYNTHESIS (Christopher Saunders)
        08B CSOUND AND ARDUINO (Iain McCurdy)
        12B PYTHON INSIDE CSOUND (Andrés Cabrera, Joachim Heintz)
        12C PYTHON IN CSOUNDQT (Tarmo Johannes, Joachim Heintz)

    Revised chapters:
        02A MAKE CSOUND RUN: Updated section about Windows install (Jim Aikin) and new sections about Csound on Android and iOS (Jacques Laplat)
        03A INITIALIZATION AND PERFORMANCE PASS has completely been rewritten (Joachim Heintz)
        04A ADDITIVE SYNTHESIS has been expanded (Iain McCurdy, Bjørn Houdorf)
        05B PANNING AND SPATIALIZATION now contains descriptions about multi-channel audio in Csound in general, and VBAP and Ambisonics in particular (Iain McCurdy, Joachim Heintz)
        10A CSOUNDQT now contains a description of the options and choices in CsoundQt's Configure Panel (Peiman Khosravi, Joachim Heintz)
        10D CABBAGE has been updated and covers now some of the exciting new developments (Rory Walsh)
        12A THE CSOUND API has been revised and extended (Francois Pinot)
        The OPCODE GUIDE has been updated (Iain McCurdy)
        The METHODS OF WRITING CSOUND SCORES now contain a description of Pysco (Jacob Joaquin)

    General additions and changes:
        The code examples now also carry some (hopefully) meaningful names in addition to the numbers.
        Many improvements to existing examples have been made by Iain McCurdy.


Send bugs reports to the Sourceforge bug tracker
           https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

I am a bit puzzled by this, as I always understood that the requirement 
that all the probabilities add up to 1 was, well, a requirement, part of 
the definition of a Markov chain. So not a problem to be solved as such. 
It basically says that given any note in the set, it always be followed 
by another note from that set. A note can have a probability of zero, 
saying that given the current note, it can never be followed by that 
note - which means that the probabilities for the other available notes 
will be that much higher.

Richard Dobson

On 01/05/2013 16:26, Mark Brophy wrote:
> In chapter 12B on Python, there is a limitation that the probabilities
> of a table for a Markov chain must equal 1. Stack Overflow contains a
> post that solves this problem
> :
>
>
> |def  weighted_choice(choices):
>     total=  sum(wfor  c,  win  choices)
>     r=  random.uniform(0,  total)
>     upto=  0
>     for  c,  win  choices:
>        if  upto+  w>  r:
>           return  c
>        upto+=  w
>     assert  False,  "Shouldn't get here"|
>

Still not really a "problem" - you can preprocess things however you 
like (no different in principle from any other more or less routine 
normalisation procedure - you can even make a user-defined opcode for it 
parametrised by the value you are changing), so long as the final ratios 
sum to unity.  The further advantage is that the resulting code is 
self-documenting.


Richard Dobson


On 03/05/2013 20:31, Mark Brophy wrote:
> The problem with restricted it to 1 is that when you want to make one
> item more probable relative to the others, you must not only increase
> the item but decrease the others, too. This wastes time; from a math
> standpoint, the software can divide all items until the sum equals one.
> A Markov chain needs only relative weightings.
>

Interestingly enough, the idea of "problem solving" is one of the two 
core themes (I am tempted to call them idées fixes) in the UK education 
drive to move from ICT to Computer Science. So just about everything is 
presented as a form of problem solving.  If the Markov requirement for 
probabilites to sum to unity is "a problem", it is not alone. Much the 
same "problem" arises in, for example, simple music notation. By 
definition, the notes and rests in a bar of 4/4 must sum to 4/4. If you 
want to change a crotchet to a minim, again other notes or rests have to 
be reduced to compensate, to ensure the durations still add up 
correctly. So we can either call that a "problem" to be solved, or 
simply "the way things are, because this is a 4/4 bar". You either 
modify the other notes by hand, or make up a bit of code to do it 
automatically.

So the question becomes, however you do it: is that "solving a problem", 
or simply "composing"?


It seems that Markov processes are now going right down the age groups:

http://scratch.mit.edu/projects/alexruthmann/943184

(NB, timing is bad in a browser, better to install Scratch, download the 
project, run and enjoy!)

Richard Dobson






On 03/05/2013 21:29, Mark Brophy wrote:
> That's essentially my point. I contributed the preprocessing code.
>
>
> On Fri, May 3, 2013 at 3:17 PM, Richard Dobson
> >
> wrote:
>
>     Still not really a "problem" - you can preprocess things however you
>     like (no different in principle from any other more or less routine
>     normalisation procedure - you can even make a user-defined opcode
>     for it parametrised by the value you are changing), so long as the
>     final ratios sum to unity.  The further advantage is that the
>     resulting code is self-documenting.
>
>

hi francesco -

thanks for the report. yes obviously the math formulas are not working 
anymore, although nothing has changed for this page since the previous 
release (and it worked).

i will report this to fm ...

best -

	joachim


Am 03.05.2013 14:19, schrieb Francesco Scagliola:
> Hi All,
> Thanks for the very good work. There is some little problems.
> 1) The pdf download doesn't work;
> 2) At the Intesities page doesn't work the image that stand for SIL
>
> Best,
> Francesco scagliola
>
>
> 2013/5/1 John Clements  >
>
>     Nice, Mark.  Thank you much for sharing this solution.
>
>     John Clements
>
>
>
>
>     On May 1, 2013, at 8:26 AM, Mark Brophy      > wrote:
>
>>     In chapter 12B on Python, there is a limitation that the
>>     probabilities of a table for a Markov chain must equal 1. Stack
>>     Overflow contains a post that solves this problem
>>     :
>>
>>
>>     |def  weighted_choice(choices):
>>         total=  sum(wfor  c,  win  choices)
>>         r=  random.uniform(0,  total)
>>         upto=  0
>>         for  c,  win  choices:
>>            if  upto+  w>  r:
>>               return  c
>>            upto+=  w
>>         assert  False,  "Shouldn't get here"|
>>
>>
>>
>>
>>
>>     On Wed, May 1, 2013 at 3:31 AM, joachim heintz
>>     > wrote:
>>
>>         hi all -
>>
>>         the third release of the csound floss manual is out at
>>         www.flossmanuals.net/csound 
>>
>>         there is a lot of new stuff in it. see below the "what's new"
>>         in detail. thanks goes to all contributors, and in particular
>>         to alexandre abrioux for his diligent proof reading.
>>
>>         all the csd example files (nearly 250 now) and audio samples
>>         can be downloaded here:
>>         http://csound-tutorial.net/__filebrowser/download/576
>>         
>>         they will also be included in the next release of csoundqt.
>>
>>         alex hofmann will push a printed version at lulu.com
>>          in a short time.
>>
>>         please let us know any bugs, errata and suggestions.
>>
>>         enjoy -
>>
>>                 joachim and iain
>>
>>
>>
>>         WHAT'S NEW IN THIS RELEASE
>>
>>             New chapters:
>>                 03E ARRAYS (Tarmo Johannes, Joachim Heintz)
>>                 04H SCANNED SYNTHESIS (Christopher Saunders)
>>                 08B CSOUND AND ARDUINO (Iain McCurdy)
>>                 12B PYTHON INSIDE CSOUND (Andrés Cabrera, Joachim Heintz)
>>                 12C PYTHON IN CSOUNDQT (Tarmo Johannes, Joachim Heintz)
>>
>>             Revised chapters:
>>                 02A MAKE CSOUND RUN: Updated section about Windows
>>         install (Jim Aikin) and new sections about Csound on Android
>>         and iOS (Jacques Laplat)
>>                 03A INITIALIZATION AND PERFORMANCE PASS has completely
>>         been rewritten (Joachim Heintz)
>>                 04A ADDITIVE SYNTHESIS has been expanded (Iain
>>         McCurdy, Bjørn Houdorf)
>>                 05B PANNING AND SPATIALIZATION now contains
>>         descriptions about multi-channel audio in Csound in general,
>>         and VBAP and Ambisonics in particular (Iain McCurdy, Joachim
>>         Heintz)
>>                 10A CSOUNDQT now contains a description of the options
>>         and choices in CsoundQt's Configure Panel (Peiman Khosravi,
>>         Joachim Heintz)
>>                 10D CABBAGE has been updated and covers now some of
>>         the exciting new developments (Rory Walsh)
>>                 12A THE CSOUND API has been revised and extended
>>         (Francois Pinot)
>>                 The OPCODE GUIDE has been updated (Iain McCurdy)
>>                 The METHODS OF WRITING CSOUND SCORES now contain a
>>         description of Pysco (Jacob Joaquin)
>>
>>             General additions and changes:
>>                 The code examples now also carry some (hopefully)
>>         meaningful names in addition to the numbers.
>>                 Many improvements to existing examples have been made
>>         by Iain McCurdy.
>>
>>
>>         Send bugs reports to the Sourceforge bug tracker
>>         https://sourceforge.net/__tracker/?group_id=81968&atid=__564599 
>>         Discussions of bugs and features can be posted here
>>         To unsubscribe, send email sympa@lists.bath.ac.uk
>>          with body "unsubscribe csound"
>>
>>
>