| Message written at 3 Aug 1999 09:10:05 +0100
I realise that it is only two weeks since 3.56 but there are a number
of changes which I wanted to get out to ensure that they work. There
are a number bug fixes to clock, divz, wgbrass and marimba, as well as
new opcodes.
The new opcodes are varied. From matt ingalls we have a set of
opcodes to interrogate sound files from within the orchestra (filelen,
filesr, filechnls and filepeak). From Gabriel Maldonado there are
expsega (and more accurate expseg), logbtwo and powoftwo for fast
arithmetic. In addition, from a collaboration between Gabriel,
Richard Boulanger and myself there are active, cpuprc, prealloc and
maxalloc which control and investigate the number of active
instruments, mainly for realtime activity
In addition Gen23 has been extended to allow deferred length (from
Paris Smaragdis)
By teh time this message is sent the system for Windows, PC-console
mode, Linux and SGI shoudl be available; and sources of course.
Solaris and Mac68k systems to follow. Others at some time I guess
==John ffitch
========================================================================
Release Notes for 3.57
======================
These are the release notes for version 3.57
Bug Fixes
---------
clock opcodes all mapped to clock 0 -- fixed
divz was decoded incorrectly in parsing
The triple strike in marimba never happened due to programming error.
The percentage of doubles and singles are variable as two optional
arguments, both defaulting to 20%.
Some error and warning strings were wrong. Extensively reviewed and
fixed
Language Changes
----------------
In GEN23 (read ASCII file of numbers it is now possible to have a
length of 0 and have the generator calculated from the number of
numbers in the file.
Opcode Fixes
------------
in buzz and gbuzz at least 1 harmonic is always used, and the absolute
value of the number is used rather than giving a warning
bug in wgbrass fixed which could lead to crashes
New Opcodes
-----------
active -- tell how many active instances there are of an instrument
cpuprc -- limit number of allocations of an instrument by load
maxalloc -- limit number of allocations of an instrument count
prealloc -- create a pool of unactive instances
expsega -- a-rate exponential segments
logbtwo
powoftwo -- fast versions of pow and log in both i and k position
ilen filelen ifilcod ; returns length of "ifilcod" in seconds
isr filesr ifilcod ; returns the sample rate of "ifilcod"
inchnls filenchnls ifilcod ; returns the number of chnls of "ifilcod"
ipeak filepeak ifilcod, [ichnl] ; returns peak absolute value of
; "ifilcod"
; if ichnl=0, returns peak out of all channels
; if ichnl>0, returns the peak of ichnl
; if ichnl is not specified, returns peak of
; the entire file.
; currently only supports AIFF_C float files
Other Changes:
-------------
The pvtool utility has been included in utils2, and as a -U option
Windows GUI Changes
-------------------
New dialog for pvlook utility
------------------------------------------------------------------------
==John ff
1999 August 3
========================================================================
active
inum active ins
DESCRIPTION
Returns the number of active instances of instrument number ins
------------------------------------------------------------------------
expsega
asig expsega ia, idur1, ib[, idur2, ic[...]]
DESCRIPTION
An exponential segment generator operating at a-rate. This unit is
almost identical to expseg, but very more precise when defining
segments with very short duration (i.e. in percussive attack phase) at
audio-rate.
Note that expseg opcode does not operate correctly at audio rate when
segments are shorter than a k-period.
INITIALISATION
ia - starting value. Zero is illegal.
ib, ic, etc. - value after idur1 seconds, etc. must be non-zero and
must agree in sign with ia.
idur1 - duration in seconds of first segment.
A zero or negative value will cause all initialization to be skipped.
idur2, idur3, etc. - duration in seconds of subsequent segments.
A zero or negative value will terminate the initialization process
with the preceding point, permitting the last-defined line or curve to
be continued indefinitely in performance. The default is zero.
PERFORMANCE
These units generate control or audio signals whose values can pass
through 2 or more specified points. The sum of dur values may or may
not equal the instrument's performance time: a shorter performance
will truncate the specified pattern, while a longer one will cause the
last-defined segment to continue on in the same direction.
------------------------------------------------------------------------
powoftwo(x)
logbtwo(x)
powoftwo() function returns 2 ^ x and allows positive and negatives
numbers as argument.
logbtwo() returns the logarithm base two of x.
If the argument is in the range [-5,+5] for powoftwo() or [0.25,4]
for logbatwo() then an internal table is used, allowing a precision
more fine than one cent in a range of ten octaves. Outside those
ranges the value is calculated afresh and will be as slow as use of
pow or log.
logbtwo() returns the logarithm base two of x.
Also they are very useful when working with tuning ratios. They work
at i and k-rate.
------------------------------------------------------------------------
cpuprc
maxalloc
cpuprc instrnum, ipercent
maxalloc instrnum, icount
prealloc instrnum, icount
DESCRIPTION
cpuprc sets the cpu processing-time percent usage of an instrument in
order to avoid buffer underrun in realtime performances
maxalloc limits the number of allocations of an instrument.
prealloc creates space for instruments but does not run them
INITIALIZATION
instrnum - instrument number
ipercent - percent of cpu processing-time to assign
icount -- number of instruments instances that can be allocated
cpuprc is an opcode that enables a sort of polyphony theshold. The
user must set ipercent value for each instrument he want to activate
in realtime. It is assumed that the total theorical processing time
of the cpu of the computer is 100%, but note that this percent value
can only be defined empirically.
For example if ipercent is set to 5% for instrument 1, the maximum
number of voices that can be allocated at any time will be 20 (as 5% X
20 = 100%). If the user attempts to play a further note while the 20
previous notes are still playing, Csound inhibits the allocation of
that note and will display a warning message,
In order to avoid audio buffer underruns, it is suggested to set
the maximum number of voices a bit below the real processing power
of the computer, because sometimes an instrument can require more
processing time than normal (for example, if the instrument contains
an oscillator which reads a table that doesn't fit in cache memory,
it will be slower than normal; also, any concurrent program which run
in multitasking, can subtract more processing power in some cases,
less power in other cases etc.)
Initially all instruments are set to a default value of ipercent =
0.0% (i.e. zero processing time or rather infinite cpu
processing-speed). Note that this opcode can be used either at
instrument 0 time or dynamically, when it only affects later
instruments. Any active instuenmt whose load is changed may lead to
incorrect or anomolous results.
In maxalloc setting the number of maximum allocation to 0 means
unlimited allocations are allowed. A negative allocation disallows
any allocation.
example:
sr=44100
kr=441
ksmps=100
nchnls=2
cpuprc 1, 2.5 ;** set instr 1 to 2.5% of cpu, max 40 voices
cpuprc 2, 33.333 ;** set instr 2 to 33.333% of cpu, max 3 voices
instr 1
...body...
endin
instr 2
....body...
endin
|