What happens if you don't use MIDI input with the
everything else the same (but without the MIDI opcodes)?

Is this with the GUI app or with command-line Csound? I must
confess I have not really tested MIDI extensively on OSX
apart from checking it works. But others, I presume
have (David Akbari?).

I'll check.

Victor
>
>
> First let me say thanks to the all those who work so hard
> to put out   Csound5.
>
> I am trying to use Csound5 with real time MIDI input.  My
> problem is   that there are many glitches in the sound.
> Not a distortion from   samples out of range, rather, the
> sound seems to cut on and off (the   same type of cut out
> that occurs when I change from one window to   another).
> Csound5 reports that I am using PortMIDI.  Is that my only
>   choice on OSX.  If not, what flag would allow me to
> change to a   different realtime MIDI module (like
> CoreAudio).  I've tried -  +rtmidi=coremidi but that does
> not work.
>
> This instrument is simple oscil playing the current MIDI
> note (in   cps) at current volume (using current MIDI
> velocity).  This file is   perfect when run on MacCsound
> 1.2.a5.  The file is below.
>
> I am using the following flags with Csound5:
> -odac -iadc -+rtaudio=coreaudio --midi-device=1 -B128 -b64
>
> This same set of flags works perfectly (with no glitches)
> when   sending an audio signal through a Csound5, so I do
> not believe that I   have a problem with -B128 -b64.
>
> I am using:
> Csound5.00 for OS 10.4
> OS 10.4.4
>
> Thanks in advance,
> Mitch
>
>
>
>
> 
> 
> sr = 44100
> kr = 4410
> ksmps = 10
> nchnls = 2
>
> instr 1
>     ;----------Get a note from the MIDI instrument
>     icps    cpsmidi                ; the note's frequency
>     iamp1    ampmidi    127             ; the note's
> amplitude scaled to a value   between 0 and 127
>     iamp = (iamp1*iamp1)/16129        ; vel 0-127 is
> squared then divided by   16129 to give more realistic vel
> to amp relationship makes kenv range   from 0-1
>
>     ;----------E.G. (MIDI realtime)
>     irise = .1                    ; length of rise portion
> in seconds
>     idec    = .1                    ; length of decay
> portion on seconds
>         ;----test to see if in release time (this block is
> modified version   of that found under opcode "xtratim" in
> Csound Manual 4.23)
>         xtratim idec                    ;extra-time, i.e.
> release dur
>            krel init 0
>            krel release                     ;outputs
> release-stage flag (0 or 1 values)
>            if (krel > .5) kgoto rel             ;if in
> release-stage goto release   section
>         ;----attack decay portion
>         kenvatk linseg 0, irise, iamp, 50, iamp      ; ADD
> IN POINTS FOR   COMPLEX ATTACK PORTION
>         kenv = kenvatk
>         kgoto done
>         ;----release section
>             rel:
>            kenvrel linseg 1, .3*idec, .2, .7*idec, 0    ;
> ADD IN POINTS FOR   COMPLEX DECAY PORTION
>            kenv = kenvatk*kenvrel
>            done:
>
>     ;----------Main Oscilator
>     a1        oscil        5000, icps, 1
>
>     ;----------Output
>     ascl        =    a1*kenv                ; BE SURE TO
> ADD IN THE *kenv SO THAT THE E.G. IS   USED
>     aout    clip    ascl, 0, 32000                ; clip
> anything over 32000 to be sure   nothing is out of range
>     outs aout, aout
> endin
>
>
>
>
>
> 
> 
> f1 0 4096 10 1 ;0 .5 0 .3
>
> ;i#    str    dur
> i1     0         8888        ; make the duration long this
> is 8888 seconds 
>
> 
>
>
>
>

Running the attached file out of the box "as is" results in no errors; 
I'm able to play 5-6 voice polyphony without breakups. On OS 10.3.9 
still, using the Csound commandline built from CVS this morning. (10 
Feb 2006)

Flags used were:
csound -odac:1 -idevaudio:1 -+noninterleaved=1 -+rtaudio=CoreAudio 
-b1024 -B1024 -M0 -d ${FILENAME}

* noninterleaved and explicit buffer-size are because of using digi 
mbox hardware.


-David

On Feb 10, 2006, at 4:12 PM, Victor Lazzarini wrote:

> What happens if you don't use MIDI input with the
> everything else the same (but without the MIDI opcodes)?
>
> Is this with the GUI app or with command-line Csound? I must
> confess I have not really tested MIDI extensively on OSX
> apart from checking it works. But others, I presume
> have (David Akbari?).
>
> I'll check.
>
> Victor


> 
>> 
>> sr = 44100
>> kr = 4410
>> ksmps = 10
>> nchnls = 2
>>
>> instr 1
>>     ;----------Get a note from the MIDI instrument
>>     icps    cpsmidi                ; the note's frequency
>>     iamp1    ampmidi    127             ; the note's
>> amplitude scaled to a value   between 0 and 127
>>     iamp = (iamp1*iamp1)/16129        ; vel 0-127 is
>> squared then divided by   16129 to give more realistic vel
>> to amp relationship makes kenv range   from 0-1
>>
>>     ;----------E.G. (MIDI realtime)
>>     irise = .1                    ; length of rise portion
>> in seconds
>>     idec    = .1                    ; length of decay
>> portion on seconds
>>         ;----test to see if in release time (this block is
>> modified version   of that found under opcode "xtratim" in
>> Csound Manual 4.23)
>>         xtratim idec                    ;extra-time, i.e.
>> release dur
>>            krel init 0
>>            krel release                     ;outputs
>> release-stage flag (0 or 1 values)
>>            if (krel > .5) kgoto rel             ;if in
>> release-stage goto release   section
>>         ;----attack decay portion
>>         kenvatk linseg 0, irise, iamp, 50, iamp      ; ADD
>> IN POINTS FOR   COMPLEX ATTACK PORTION
>>         kenv = kenvatk
>>         kgoto done
>>         ;----release section
>>             rel:
>>            kenvrel linseg 1, .3*idec, .2, .7*idec, 0    ;
>> ADD IN POINTS FOR   COMPLEX DECAY PORTION
>>            kenv = kenvatk*kenvrel
>>            done:
>>
>>     ;----------Main Oscilator
>>     a1        oscil        5000, icps, 1
>>
>>     ;----------Output
>>     ascl        =    a1*kenv                ; BE SURE TO
>> ADD IN THE *kenv SO THAT THE E.G. IS   USED
>>     aout    clip    ascl, 0, 32000                ; clip
>> anything over 32000 to be sure   nothing is out of range
>>     outs aout, aout
>> endin
>>
>>
>>
>>
>>
>> 
>> 
>> f1 0 4096 10 1 ;0 .5 0 .3
>>
>> ;i#    str    dur
>> i1     0         8888        ; make the duration long this
>> is 8888 seconds 
>>
>>