Connor Haynes

I have a question which i am having trouble with, as im pretty new to Max. Basically i am doing an installation which requires me to randomly trigger samples (around 30 different samples for 12 different triggers). I have made a patch where this is the case using a umenu and chained sfplay~ using a gate, so that multiple samples can play back upon simultaneous bangs. The only problem is this potentially could be too much on the disk drive to load as i may have 12 of these triggers going off at one time, which would be alot of samples for the hard drive to search for in real time! (30samples x 12triggers) So i was wondering if there is anyway i could modify the patch, and use buffers so the samples are reading off the RAM/Max instead, but still keeping the round robin cycle aspect ect. Using a play object which i can modulate the playback speed with the velocity would be a massive bonus too which i know sfplay~ and groove~ has, but to add, im not sure that the playback speed would sound good if the 2nd trigger pitched the both samples playing as the previous sample was playing out. Il include my patch up here so you can see what im doing, thank you very much for any input on this guys! Cheers!

Round Robin Help Patch.maxpat

I have a question which i am having trouble with, as im pretty new to Max. Basically i am doing an installation which requires me to randomly trigger samples (around 30 different samples for 12 different triggers). I have made a patch where this is the case using a umenu and chained sfplay~ using a gate, so that multiple samples can play back upon simultaneous bangs. The only problem is this potentially could be too much on the disk drive to load as i may have 12 of these triggers going off at one time, which would be alot of samples for the hard drive to search for in real time! (30samples x 12triggers) So i was wondering if there is anyway i could modify the patch, and use buffers so the samples are reading off the RAM/Max instead, but still keeping the round robin cycle aspect ect. Using a play object which i can modulate the playback speed with the velocity would be a massive bonus too which i know sfplay~ and groove~ has, but to add, im not sure that the playback speed would sound good if the 2nd trigger pitched the both samples playing as the previous sample was playing out. Il include my patch up here so you can see what im doing, thank you very much for any input on this guys! Cheers! 

Hi everyone,
I have a question which i am having trouble with, as im pretty new to Max. Basically i am doing an installation which requires 

round-robin-sample-triggering-using-buffers-appreciate-any-help!

I would instead preload all samples into 12 sfplayers, and recall samples by numbers.
That way one can switch between samples fast enough.
Look into sfplay help file 



Thank you for the reply! That makes sense,  but Im looking into it now in the help file but im having trouble understanding how that would work. I sort of have an idea but I cant materialise it, please could you give me an idea of how that would chain together? Cheers!

I load everything into a polybuffer~ then use this poly~ to play individual samples.

So for example, if you create a polybuffer~ named "drums" and then name the poly~ [poly~ {filename} 8 args drums] and then just send it numbers (i.e. "1", "2", "3", etc...) it will play back the corresponding buffer inside of the named polybuffer~.

(The reason it's inside a poly is so that each sample can play it's full duration if another sample is triggered before it completes. The "8" in the poly~ name is how manyn simultaneous samples you can have.)

Advantage of polybuffer is  that You don't need to learn how to preload sfplay
with samples,  but if the samples are very long and many then it takes a time till buffers are loaded.
Then sfplay could be a better solution.
You have to decide upon that parameter and max  number of polyphony needed,
and if all triggers share same amount of voices, or not.
Here is example how to preload samples into sfplay :

Guys, i cant thank you enough, really really appreciate it. Both work a charm! THANK YOU

Hi! I've tried to create a poly sampler with a polybuffer object as described by Rodrigo, but I can only get one note at the same time ... My poly object has the patch by Rodrigo (saved as "drumpoly.maxpat") , and inside the box there is "poly~ drumpoly 6 @args cymbals". Here the name "cymbals" correspond to the name of the polybuffer~ object. The poly~ has on inlet with a number to change the sample. When I change the number a new sample plays, but as soon as the number is changed the old sample stops and the new one plays, so no real polyphony.

Silly question perhaps, but do you have `prepend note` before the `poly~`? That's what lets the `poly~` know to do the polyphonic thing.

   Thanks Rodrigo!  I managed to get the polyphony working!

Round Robin sample triggering using buffers? Appreciate any help!

jasch

I'm the process of recompiling for FAT externals, and I come across this broken functionality:

It's basically traversing a linked list of instances of another class using the mess4 macro from ext_mess.h to remotely call the paint methods from these clients.

Operand of type 'double' cannot be cast to a pointer type

Operand of type 't_pt' (aka 'struct _pt') where arithmetic or pointer type is required

t_pt origin;
double size;
void  *client;

client = linklist_getindex(x->registered_controllers, 0);
while(client)
{
    mess4((t_object *)client, ps_paint, g, origin, size, 1);
    linklist_next(x->registered_controllers, client, (void *)&client);
}

P.S. this could also be due to an update of Xcode to 12.4, which might react more strictly to types?

It's basically traversing a linked list of instances of another class using the mess4 macro from ext_mess.h to remotely call the paint methods from these  clients.

both errors are on the mess4 line:
Operand of type 'double' cannot be cast to a pointer type
Operand of type 't_pt' (aka 'struct _pt') where arithmetic or pointer type is required

t_pt origin;
double size;
void  *client;

client = linklist_getindex(x->registered_controllers, 0);
while(client)
{
	mess4((t_object *)client, ps_paint, g, origin, size, 1);
	linklist_next(x->registered_controllers, client, (void *)&client);
}

Hi, 
I'm the process of recompiling for FAT externals, and I come across this broken functionality:
It's basically traversing a linked list 

trouble-with-mess4-macro-from-ext-mess-for-8-2-api

hi Jasch, thanks much for diving into updating your externals!

I can make some suggestions, but it's tricky to diagnose this without the source. There are significant changes wrt to calling object methods due to apple silicon compatibility needs. Passing double arguments to untyped method calls is no longer supported. I believe the solution is to use object_method_direct, which requires passing in both the return type and signature of the calling method. You can check the header for the full documentation of this call.

for your example it would look something like this (assuming a return type of t_max_err, and argument g is a t_jgraphics):

object_method_direct(t_max_err, (t_object *, t_jgraphics*, t_pt, double, long),(t_object *)client, ps_paint, g, origin, size, 1);
Please let me know if that doesn't get you going.

Adapting the remote method calls from the mess4 macro to the object_method_direct seems to have done the trick. tbc.

cheers

trouble with mess4 macro from ext.mess for 8.2 API

Arne

Camera in jitter always looks to a target (usually position 0 0 0, unless you change it with 'lookat' attribute). But can you turn off this 'looking at'? So that when you move your camera it doesn't turn towards a target.

Camera in jitter always looks to a target (usually position 0 0 0, unless you change it with 'lookat' attribute). But can you turn off this 

'turn-off'-lookat-attribute

[in case you wondered: I was using jit.gl.render with its camera-attribute (as shown in Jitter Tutorial 32: camera view). Jit.gl.render doesn't have a locklook-attribute. Your suggestion made me find out there also is a jit.gl.camera, that has the locklook-attribute. Great!]

'turn off' lookat-attribute

yaniki

I am struggling with a problem related to handling dictionaries, maybe someone will be able to help me.

I have a dictionary containing an array of arrays (in other words: 2-dimensional array) containing "objects". For example something like this:

    "a" : [ [             {

And after being formatted more human readable:

I am trying to access to individual values stored in particular objects (

) with the "get" message sent to the [dict] object. Unfortunately, I'm unable to construct a syntactically correct "get" message. I guess it should be something like this:

but this method produces error messages in the console:

dict: no dictionary named a[0], cannot get value

dict: index out of range for key a[0][0]::b

I would be grateful for any tips. I read a lot of threads on the forum about working with dictionaries, but unfortunately I did not find a satisfactory solution. Of course, I can ditch the dicts and use a different approach, but for many reasons dictionaries are attracting me as they are a flexible containers for, let's say, "irregular" data. Also, I need the native objects (due to event-driven timing) - so I don't want to use scripting.

{ 
    "a":[
        [{"b":1,"c":0}, {"b":0,"c":1}],
        [{"b":1,"c":1}, {"b":0,"c":0}]
    ]
}

I am trying to access to individual values stored in particular objects ({"b":?,"c":?}) with the "get" message sent to the [dict] object. Unfortunately,  I'm unable to construct a syntactically correct "get" message. I guess it should be something like this:

dict: no dictionary named a[0], cannot get value
dict: index out of range for key a[0][0]::b

Dear Friends
I am struggling with a problem related to handling dictionaries, maybe someone will be able to help me.
I have a dictionary con

dictionaries-and-2d-arrays-of-objects

dictionaries and 2d arrays of objects

dirkdebruin

Max 8.2 beta crashes when using one of my latest projects that works fine on 8.1.11

I have trimmed down the patch to find the source and I think it is related to nearestpix in jit.gen. It only crashes when using a matrix with 5 or more dimensions. Sometimes it crashes right away when changing to 5 or more dimensions, and sometimes it crashes when closing the patcher.

I'm on an M1 mac mini, which overall shows some significant performance improvement with 8.2 compared to 8.1.11 so far. 

8.2 crash test.maxpat

I'm on an M1 mac mini, which overall shows some significant performance improvement with 8.2 compared to 8.1.11 so far. 


Max 8.2 beta crashes when using one of my latest projects that works fine on 8.1.11
I have trimmed down the patch to find the source and I t

crash-report-1

Okay. AFAICT, the code for converting smaller vectors to larger vectors in jit.gen, are not zero padding the arrays that the smaller vectors are copied into if those larger vectors are of size greater than 4. Examining the code, this doesn't look particularly new. However, it could be that under previous versions, the memory allocation conditions were zeroing memory themselves and wouldn't have required any logic to do so or things were just very lucky. I have some more experimentation to do here.

In the meantime, if you want a patch that works, you may zero pad your smaller vectors to whatever the larger vector you will be converting them to. In this case the [- 1 0] can be replaced with [- 1 0 0 0 0 0 0 0] or however many zeros you need to match the dim count.

Thanks! That seems to fix the problem and is probably good practice to be doing anyway. 
I'm probably not going to change the dim count dynamically so it is not going to be a problem for me anymore but it might be good to give the object a default behaviour for these situations anyway. 

crash report

Hideki Kozakura

Screen Shot 2021-10-04 at 16.01.27.png

I have tried to install and reinstall the app several times but no such object appear every time when I open the saved file. How can it be solved?


 
I have tried to install and reinstall the app several times but no such object appear every time when I open the saved file. How can it b

no-such-object-1

It looks like you added an object to the patcher and gave it the name "(function)", which is not recognised as a known object type.  A recognisable name would be one of Max's objects or a patcher of your own that is in the same folder as your test_generator patcher or elsewhere in Max's search path.  Also,  and I'm not sure about this, but I don't think parentheses () are valid in object names.

Thanks a lot, dear Simon. The problem is now solved! 

NO SUCH OBJECT

benjamin vyger

Inside the patcher potar ratio. Problem to trig $i2 or $i3 at the good moment

In my patch (at the top), I have the position, here 19550, of a motor that I transform into data from 0 to 127 for a knob that I use in max4live. One round is 6400 step.

https://www.youtube.com/watch?v=CTfMDb0Wu3A.

 at the end, the motors go very fast and I can, when recording in Live, put a ratio of 0.5 for example and therefore my pots go half as fast as the real position of the motors. In fact, I would like to control the knob"phase8bis" with a 1/2 ratio from "phase8" recorded with a 1/1 ratio. The difficulty is that when the knob "phase8" arrives at 127 then 0, either the "phase8bis" is at 64 and must go to 65, or "phase8bis" is at 127 and must go to 0. My patcher potar ratio works but not always.

I have a problem to trig $i2 and $i3 at the good moment. That's all, I hope to have been clear, and that you will be able to find the solution ;)

potar 1 ratio.maxpat

Inside the patcher potar ratio. Problem to trig $i2 or  $i3 at the good moment

In my patch (at the top), I have the position, here 19550,  of a motor that I transform into data from 0 to 127 for a knob that I use in max4live. One round is 6400 step.
As you can see https://www.youtube.com/watch?v=CTfMDb0Wu3A. at the end, the motors go very fast and I can, when recording in Live, put a ratio of 0.5 for example and therefore my pots go half as fast as the real position of the motors. In fact, I would like to control the knob"phase8bis" with a 1/2 ratio from "phase8" recorded with a 1/1 ratio. The difficulty is that when the knob  "phase8" arrives at 127 then 0, either the "phase8bis" is at 64 and must go to 65, or "phase8bis" is at 127 and must go to 0. My patcher potar ratio works but not always.
I have a problem to trig  $i2 and  $i3 at the good moment.  That's all, I hope to have been clear, and that you will be able to find the solution ;) 

Hi guys and girls,
 
Inside the patcher potar ratio. Problem to trig $i2 or  $i3 at the good moment
 

In my patch (at the top), I have the 

transform-speed-of-a-potar-second-request-better-explained


of course there is always interaction between layout and code in both directions... but in this case... i would suggest to clear the layout a bit in order to understand what you are doing and where the error could be.

the 9 numberboxes and 3 triggers makes the patch really hard to read and removing them would be of great help.


if you choose 7196 with this first number box, you will understand how works the patcher potar ratio.
I have changed something useless inside.
Actually, in the patcher potar ratio the number box 4 trig a  bang when the number is upper than 127. Between 7196 and 7200 you have 4 times the number upper than 127, that's why I put a counter until 4, in order to have the number 1.  And if $i1 < 2 then $i2 means, let the numbers from box 3 come.
I hope it will help. Thanks ;)

It's hard to understand what your patch is doing. The description of the problem is also a bit hard to follow, it would be better if you tried formulating it in a more abstract way.
However, I think I understand what you are trying to do, after scaling the value from the motors, you have a dial that wraps around 127 and you want a dial that moves with half the speed. When the normal-speed dial wraps from 127 to 0 the half-speed dial needs to go from 64 to 65 (or 0 to 127), this means that you need to keep track of wrap-arounds and change the half-speed dial's value accordingly.
I solved this issue a while ago, here is my solution.

Tell me if I'm understanding your problem correctly and if my patch is helping you.
(my patch doesn't work with negative numbers, I believe you need to change all modulos (%) with something else for it to work)

Thanks man, it looks like working very well. I 'll try to integrate it to my M4L controller tomorrow. Thanks again. I kiss you as much as possible



if i remove all the superflous litter from the patch it looks like that.

now we can freeze the order using triggers



...and now you have a chance to see which part might be wrong. :)


Transform speed of a potar (second request better explained)

Joe Kaplan

I'm using MIN to build a package of Max externals and I'd like to make use of Max's notification system.

I found the relevant functions in the C API docs.

I discovered that these functions are accessible in MIN using the "c74::max" namespace. But I'm having some difficulty implementing them.

automatically defining a t_object for the present class. 

But I'm not sure what the syntax is to access it in my register function?

I'm struggling to understand how to initializes the arguments to that function. Particularly the names parameter which takes type t_symbol***.

I'm sure a lot of this is just my own unfamiliarity with C and C++, but I'm struggling to get this going from the present documentation.

Are there any code examples available that show the Notifications system in action in MIN? Min-Devkit includes some buffer~ related objects that show a framework for receiving notifications. But I'm not finding any examples of how to register objects in a namespace, view a list of registered objects in a namespace, or send notifications.

I'm using MIN to build a package of Max externals and I'd like to make use of Max's notification system.  

 I found the relevant functions in the C API docs.  

 object_notify()
object_attach()
object_register()
object_subscribe()
 object_register_getnames() 

I discovered that these functions are accessible in MIN using the "c74::max" namespace.  But I'm having some difficulty implementing them.  

For example, it looks like MIN is automatically defining a t_object for the present class. 
But I'm not sure what the syntax is to access it in my register function?

Also, I'm looking at the  object_register_getnames().
I'm struggling to understand how to initializes the arguments to that function.  Particularly the names parameter which takes type t_symbol***.

 Are there any code examples available that show the Notifications system in action in MIN?  Min-Devkit includes some buffer~ related objects that show a framework for receiving notifications.  But I'm not finding any examples of how to register objects in a namespace, view a list of registered objects in a namespace, or send notifications.

Greetings,
I'm using MIN to build a package of Max externals and I'd like to make use of Max's notification system.  
 I found the relevant 

working-with-notifications-in-min-how-to-register-register_getnames-and-attach

you can use the the min function maxobj() to return the t_object pointer from your min class.
calling object_register from your class would look something like: c74::max::object_register(namespace_sym, name_sym, maxobj());

I'm not sure of any code examples using Min, but the max-sdk demonstrate this, and the basic functionality is the same in both sdks. 

basic usage of getnames is something like this:

long		namecount = 0;
t_symbol	**names = NULL;
t_max_err err = object_register_getnames(namespace_sym, &namecount, &names);
if (!err && names) {
	for (i=0; i<namecount; i++) {
		t_symbol *name = names[i];
	}
}

Thanks Rob.  This is super helpful!  I gave it a shot, but was unsuccessful.   Perhaps you can help me see where I'm going wrong?  I have two objects.

message<> maxclass_setup { this, "maxclass_setup",
        MIN_FUNCTION {
            c74::max::t_symbol* Namespace = c74::max::gensym("Namespace");
            c74::max::t_symbol* sender = c74::max::gensym("sender");
            c74::max::object_register(Namespace, sender, maxobj());
            return {};
        }
    };
    
message<> bang{ this, "bang", "Post the greeting.",
        MIN_FUNCTION {
            c74::max::t_symbol* notification = c74::max::gensym("well, hello");
            c74::max::object_notify(maxobj(), notification, NULL);
            return {};
        }
    };

message<> maxclass_setup{ this, "maxclass_setup",
        MIN_FUNCTION {
            c74::max::t_symbol * Namespace = c74::max::gensym("Namespace");
            c74::max::t_symbol* sender = c74::max::gensym("sender");
            c74::max::object_attach(Namespace, sender, maxobj());
            return {};
        }
    };
    
message<> notify { this, "notify", "Receive a Notification.",
        MIN_FUNCTION {
            cout << "notified" << endl;
            notification n {args};
            cout << "received notification from " << n.source() << endl;
            cout << "it said " << n.name() << endl;
            return {};
        }
    };

The idea is that the "Send" object gets registered to the "Namespace" namespace with the name "sender" when the object is loaded.   The receive object attached to the object named "sender" in the namespace "Namespace".

Then when "Send" gets a bang, "Recieve" gets a notification and posts a bunch of stuff to the Max window.

This code builds happily and I can create the objects in Max.  But when I give my sending object a bang, nothing happens.

can you zip up your project source and send my way? you can do this privately if you prefer, my email is linked in my profile (click on the avatar).

ok you were mostly there but there were a few fundamental issues. You should be calling these registration and attach functions in object context (e.g. an object constructor) not class context (e.g. maxclass_setup) as you were doing. you can step into the code for maxobj() and see it is returning NULL if called from maxclass_setup.

if you move the code for both classes to a constructor function you will be getting close. The other issue is that depending on your use case, you may wish to use object_subscribe rather than object_register. If you register rather than subscribe, the registration functionality is dependent on the server object being created prior to the client. if you use subscribe there is no dependence on order.

i'm assuming this example is simply for testing, and that your real implementation will provide proper unique id as the server name, and test for return values from the api calls for success.

one minor caveat related to min is that dummy objects are created that will call the constructor one time as part of the class initialization. in this case maxobj() will return NULL, and therefore no registration or subscription will take place in those dummy constructors (since passing NULL to those calls fails silently, there is no issue) but you could test maxobj() output if desired and necessary, to ensure you are not in the dummy constructor.

NotificationsTest2_Send() {
	c74::max::t_symbol* Namespace = c74::max::gensym("Namespace");
	c74::max::t_symbol* sender = c74::max::gensym("sender");
	c74::max::object_register(Namespace, sender, maxobj());
}

....

NotificationsTest2_Receive() {
	c74::max::t_symbol* Namespace = c74::max::gensym("Namespace");
	c74::max::t_symbol* sender = c74::max::gensym("sender");
	c74::max::object_subscribe(Namespace, sender, c74::max::gensym("NotificationsTest2.Send"), maxobj());
}

Rob, this works!  I can't believe it!   I think I can get on with building my project now.  I'm so relieved.  Thank you! 

A couple of housekeeping questions from the novice programmer:

 i'm assuming this example is simply for testing, and that your real implementation will provide proper unique id as the server name, 

 Yes.  I had considered this.  Though I haven't yet thought through how to achieve it.  Do you have any suggestions for how to get a count  of how many instances of an object are currently in the patch?   I could use that to generate the server names.   Or, come to think it, perhaps this is where object_register_getnames() comes in handy.

 and test for return values from the api calls for success. 

I had not considered this.  It looks like all of these functions return a pointer to the registered server object.  If I'm understanding you correctly, I should test make sure I get back a pointer to the registered server object and not "NULL" or something else?  Is that on the right track?

 one minor caveat related to min is that dummy objects are created that will call the constructor one time as part of the class initialization. 

I read about this in the Min Documentation for writing objects.   If I'm understanding correctly,  I need to watch out for this if I'm doing other logic in my constructor that might fail less silently if maxobj() is null.  Is that right?

Working with Notifications in Min:  How to register,  register_getnames, and attach?

Karthik Pai

I want audio to trigger based on certain values of X axis and Y axis corresponding to a time stamp

how-to-trigger-audio-based-on-tabular-data-of-time-stamps

How to trigger audio based on tabular data of time stamps

Dom Aversano

I just wanted to let you know, I´m very happy to be running an online course soon with Music Hackspace, that will take a look at how South Indian Carnatic rhythmic theory can challenge our way of thinking about time and rhythm – both in terms of composition, and instrument design!

If you're interested you can find out more here.

https://musichackspace.org/events/understanding-indian-rhythm-through-simple-algorithms-workshop-series-19th-october/

If you're interested you can find out more here. 

Hello all, 
I just wanted to let you know, I´m very happy to be running an online course soon with Music Hackspace, that will take a look at

understanding-indian-rhythms-simple-algorithms

Understanding Indian rhythms simple algorithms

Ben Bracken

We are excited to announce the availability of our 

 as a public beta. We’ve worked closely with 

. The Eventide H9000 is a rack-sized multi-effects unit with powerful DSP processing capabilities and numerous high quality built-in effects. We’re excited to enable RNBO export for this and open the door to unique custom studio effects and instruments without the strict limitations of other hardware platforms.

RNBO support is similar to the existing Gen support, but more streamlined, and uses the built-in 

, rather than a watcher patcher, to manage the export process. Eventually, all of these changes will be rolled into the standard release versions of RNBO and this package.

Get the preview version of the RNBO package (v1.2.7-eventide.0):

Get the preview version of the "Eventide H9000" package (v1.0.4-beta.3):

We’d love to hear any experiences and feedback from RNBO users with access to an H9000. To join the discussion, use the 

We are excited to announce the availability of our Eventide package for RNBO as a public beta. We’ve worked closely with Eventide to bring...

eventide-h9000-package-for-rnbo-public-beta

Eventide H9000 Package for RNBO - Public Beta

This update focuses on Max for Live stability, fixing regression bugs, and a few new 

The full list of improvements are below, or on the Max 8.6.2 release page:

: Updated LOM documentation to the state on 12.0.x.

We are happy to announce the release of Max 8.6.2.

This update focuses on Max for Live stability, fixing regression bugs, and a few new...

Cycling '74 Forums

We are excited to announce the availability of our Eventide package for RNBO as a public beta. We’ve worked closely with Eventide to...

We are happy to announce the release of Max 8.6.2. This update focuses on Max for Live stability, fixing regression bugs, and a few new...

Thanks Rodrigo! I managed to get the polyphony working!

Thanks Rob, Adapting the remote method calls from the mess4 macro to the object_method_direct seems to have done the trick....

Hi Rob, thanks for your solution! [in case you wondered: I was using jit.gl.render with its camera-attribute (as shown in Jitter Tutorial...

Dear Friends I am struggling with a problem related to handling dictionaries, maybe someone will be able to help me. I have a dictionary con

Thanks! That seems to fix the problem and is probably good practice to be doing anyway. I'm probably not going to change the dim count...

Thanks a lot, dear Simon. The problem is now solved!

if i remove all the superflous litter from the patch it looks like that. now we can freeze the order using triggers ...and now you...

Rob, this works! I can't believe it! I think I can get on with building my project now. I'm so relieved. Thank you! A couple of...

I want audio to trigger based on certain values of X axis and Y axis corresponding to a time stamp

Hello all, I just wanted to let you know, I´m very happy to be running an online course soon with Music Hackspace, that will take a look at