Nonrealtime sfrecord, metro and metro~

Oct 22, 2006 at 9:03pm

Nonrealtime sfrecord, metro and metro~

I have a VST synth receiving midievent messages generated from an algorithm
that is triggered by bangs from a metro 250. It plays fine in real time
although, as you can guess, it drives up the cpu % substantially. When I
put it into nonrealtime in DSPstatus the metro seems to tick very slowly. I
also tried metro~ with same results.

Here is a short patch that illustrates the problem.

max v2;
#N vpatcher 175 238 976 790;
#P window setfont “Sans Serif” 9.;
#P window linecount 1;
#P newex 144 244 32 196617 print;
#P button 146 213 15 0;
#P toggle 246 87 15 0;
#P newex 251 131 58 196617 metro 250;
#P newex 253 231 31 196617 mtof;
#P newex 254 195 34 196617 + 36.;
#P newex 251 166 58 196617 random 49;
#P message 454 342 30 196617 open;
#P toggle 407 286 15 0;
#P newex 404 394 64 196617 sfrecord~ 2;
#P user ezdac~ 229 376 273 409 0;
#P newex 319 275 40 196617 cycle~;
#P connect 10 0 11 0;
#P connect 8 0 10 0;
#P connect 0 0 1 0;
#P connect 9 0 8 0;
#P connect 8 0 5 0;
#P connect 6 0 7 0;
#P connect 5 0 6 0;
#P connect 0 0 1 1;
#P connect 7 0 0 0;
#P connect 0 0 2 0;
#P connect 4 0 2 0;
#P connect 3 0 2 0;
#P connect 0 0 2 1;
#P pop;

Obviously I am missing something.

Cheers
Gary Lee Nelson
Oberlin College
http://www.timara.oberlin.edu/GaryLeeNelson

#28307
Oct 23, 2006 at 6:07am

Gary Lee Nelson wrote:
> I have a VST synth receiving midievent messages generated from an algorithm
> that is triggered by bangs from a metro 250. It plays fine in real time
> although, as you can guess, it drives up the cpu % substantially. When I
> put it into nonrealtime in DSPstatus the metro seems to tick very slowly. I
> also tried metro~ with same results.

To have scheduler events correctly timed in nonrealtime mode you have to
put “scheduler in audointerrupt” on. Elsewise they are not synced…

Stefan


Stefan Tiedje————x——-
–_____———–|————–
–(_|_ —-|—–|—–()——-
– _|_)—-|—–()————–
———-()——–www.ccmix.com

#86763
Oct 23, 2006 at 9:01am

Hi,
You have to know “scheduler in audo interrupt” does not work fine in
nonrealtime mode (with sfplay~Z?). After a while, you’ll notice a
delay. I did not really had the time to check properly where the
problem comes from, but there is a problem.

#86764
Oct 23, 2006 at 2:08pm

Thanks Stefan. That did the trick with adstatus~ takeover = 1.

On 10/23/06 2:07 AM, “Stefan Tiedje” wrote:

> Gary Lee Nelson wrote:
>> I have a VST synth receiving midievent messages generated from an algorithm
>> that is triggered by bangs from a metro 250. It plays fine in real time
>> although, as you can guess, it drives up the cpu % substantially. When I
>> put it into nonrealtime in DSPstatus the metro seems to tick very slowly. I
>> also tried metro~ with same results.
>
> To have scheduler events correctly timed in nonrealtime mode you have to
> put “scheduler in audointerrupt” on. Elsewise they are not synced…
>
> Stefan

Cheers
Gary Lee Nelson
Oberlin College
http://www.timara.oberlin.edu/GaryLeeNelson

#86765
Oct 23, 2006 at 4:00pm

I should have recalled this from my experience with render_node. I use
record~ into a buffer. That works fine and is very fast.

On 10/23/06 5:01 AM, “olivier pasquet” wrote:

> Hi,
> You have to know “scheduler in audo interrupt” does not work fine in
> nonrealtime mode (with sfplay~Z?). After a while, you’ll notice a
> delay. I did not really had the time to check properly where the
> problem comes from, but there is a problem.
> _______________________________________________
> maxmsp mailing list
> maxmsp@cycling74.com
> http://www.cycling74.com/mailman/listinfo/maxmsp

Cheers
Gary Lee Nelson
Oberlin College
http://www.timara.oberlin.edu/GaryLeeNelson

#86766

You must be logged in to reply to this topic.