We are pleased to announce the release of version 5.15. The sources
are on the standard Sourceforge location
as both zip and tar.gz
Platform packages will follow shortly, and the manual on Friday.
Notes for 5.15
New parser has been subjected to a great deal of work. It now has
better checking of argument types and use, better diagnostics and
increased functionality. We have only reached this stage in the last
few days so we judge it prudent to leave the old parser as standard.
We would be pleased if more users tried the new and gave the
A major reorganisation means that there are many fewer plugins and
most opcodes are in the base (about 1250 of them). A side effect of
that is that leaving old plugins from an earlier release is a
disaster, and so 5.15 will not load earlier plugins.
The multicore system is now safe (ie maintains semantics) when zak,
channels or table modification are made.
ftab2tab transfers between ftables and t-variables
tab2pvs tsig - pvs conversion
pvs2tab pvs - tsig conversion
cpumeter-- not really new but now available in OSX
(EXPERIMENTAL) ftresize and ftresizei allow resizing of
existing tables. These will be permanent if the
community feel they are useful.
hrtfearly, hrtfreverb opcodes
New Gen and Macros:
Code to allow GEN49 to be deferred [NB does not seem to work]
Modified Opcodes and Gens:
socksend and sockrecv no longer uses MTFU check and work on Windows
mpulse changed so if next event is at negative time use the absolute value
serial opcode now runs on Windows as will as Un*x
out, out2, outq, outh, outo outx and out32 are now identical
opcodes and will take up to as many arguments as nchnls.
This replaces the current remapping of opcodes
turnoff2 now polymorphic wrt S and k types (ie accepts instrumnet names)
jacko: fixed a segfault removing the unused JackSessionID option
doppler memory leak fixed
transegr fixed in release mode when skipping most of envelope
FLPack now agrees with manual
max_k now agrees with manual
atsa code now works on Windows in more cases
tabmorph bug fixed
fixed problem with user-defined opcodes having no outputs
Various fixes to * ... */ comments
Various licence issues sorted
Loris is no longer part of the Csound tree
Memory leaks fixed
If no score is given a dummy that runs for over 100 years is
All score processing takes place in memory without temporary
String memory now expandable and no size limitation
#if #else #end now in new parser
Adjustments to MIDI file precision in output
On OSX move from Coreaudio to AuHAL
Multicore now safe for ZAK, Channels and modifying tables
New coremidi module
Virtual Keyboard improved:
1) Dropdown for choosing base octave (the one that
starts with the virtual key mapped to physical key
Z). Default value is 5 which is backwards compatible.
2) Shift-X mappings which add two octaves to X
mappings for a total of 4 octaves playable from the
physical keyboard (starting from selected base octave).
3) Control-N / Control-Shift-N mappings to increment
/ decrement slider for control N.
4) Mouse wheel now controls sliders.
tsig type for vectors
tsigs and fsigs allowed as arguments in UDOs
Minor version upped
Very, very, very many!