Roberto Federici

How can I extract, in a message, the entire numeric list of a column inside the jit.cellblock object. Thanks so much. 

extract-column-list-jitcellblock

did you bother to read jit.cellblock help file ?

Thanks. How to multiply a list of numbers with a variable number which always gives me the multiplied list at the output

vexpr or scale, set float or int, depending on input or mutiplication factor.

Extract column list jit.cellblock

hi together,
I realized today that I cannot use the --- for my buffer object I am using for gen.
Does anybody know a solution to create an unique name space in max for live within a gen~ ?

hi together,
I realized today that I cannot use the --- for my buffer object I am using for gen.
Does anybody know a solution to create an ... 

buffer-name-for-gen

I believe you are mistaken. As long as you type the buffer name as an argument to the max [gen] object (as opposed to only inside the gen code) this should work.

thanks for the suggestion, but it is not working for me. also I would need the buffer length ...
but it seems already the @bufname is invalid.

Don't use "---bumi" in the gen~ patcher, use "bufname".

E.g. within gen~, use [buffer bufname],  [dim bufname], etc.   

The attribute "@bufname ---bumi" on the gen~ object tells the gen~ that where gen~ uses "bufname", it will map to whatever ---bumi is in Max. 

Super! Now I understood and it works just fine.
Thanks to both of you.

Is this somewhere described or maybe this falls more into the category of things which are just basic knowledge which I might not have :)


---buffer name for gen

TobyAM

I've been using the rnbo.example.juce project for my plugin for months.
In an earlier version of Max the GUI would update when parameters in RNBO would have their values change programmatically without user input. RNBO would generate a new parameter value, and it would be reflected in the JUCE GUI. Great.
In the current Max version I noticed that's not happening. Parameter changes from within RNBO are not reflected in the JUCE GUI until the plugin instance window is closed and opened again. The parameter change is not updated in the DAW host either.

I spent a whole day trying to debug C++ code for the JUCE example, only to find that rolling back the Max version and exporting and rebuilding fixed it. Perhaps it's a problem with updateSliderForParam

So this seems like an unintended change if not a bug! I'll be using version 8.5.4 in the meantime!

I've been using the rnbo.example.juce project for my plugin for months.
In an earlier version of Max the GUI would update when parameters ... 

parameter-changes-in-rnbo-not-reflected-in-juce-gui

The latest version of RNBO attempts to fix some issues with automation that previous plugin exports have. We've changed the default for a parameter notifications and added metadata to allow for altering those defaults: https://github.com/Cycling74/rnbo.adapter.juce/blob/dd233609adc495828fa9efcf756ab2a155a3bd1d/RNBO_JuceAudioProcessor.cpp#L144C5-L148 

I believe you should get your old behavior if you define RNBO_JUCE_PARAM_DEFAULT_NOTIFY in your build process.

Here is some more information about the metadata: https://rnbo.cycling74.com/learn/using-the-vst-audiounit-target 

Well, that's pretty cool! I missed the memo :) 
I just added the metadata to the parameter I needed and I'm good to go, thank you!

edit: But now presets don't recall until the window is reopened... well, more to do.

As the plugin is no longer recalling save states or presets properly, I would like to follow your advice and "define RNBO_JUCE_PARAM_DEFAULT_NOTIFY in your build process."  I haven't figured out how to make that work yet. I tried putting "set(RNBO_JUCE_PARAM_DEFAULT_NOTIFY TRUE)" in the CMakeLists.txt file, but that's not working... Any tips?

In the meantime I've added the metadata to each object individually, but with this global change, was there any other change I should know about that would allow saves and presets to be recalled?

But now presets don't recall until the window is reopened...
well, more to do.

As the plugin is no longer recalling save states or presets properly, I would like to follow your advice and "define RNBO_JUCE_PARAM_DEFAULT_NOTIFY in your build process."

I haven't figured out how to make that work yet. I tried putting "set(RNBO_JUCE_PARAM_DEFAULT_NOTIFY TRUE)" in the CMakeLists.txt file, but that's not working... Any tips?

Here is an example compile definition:
https://github.com/Cycling74/rnbo.example.juce/blob/cab65fd46cc1398152f088c82f0ac4cc61388c95/CMakeLists.txt#L54 

What you did only sets a definition within the cmake build setup, it doesn't actually become a `#define`

add_compile_definitions(JUCE_VST3_EMULATE_MIDI_CC_WITH_PARAMETERS=0 RNBO_JUCE_PARAM_DEFAULT_NOTIFY)

I'm curious to know how presets behave after you do this. There is something going on with presets in the latest rnbo code that I haven't quite figured out yet, I'd like to know if you're still having issues after this or not.

Thanks! I've implemented that as above, which works as expected which is great!

EDIT: I had to delete and rebuild builds and exports, but it's all working as expected now.

Yeah, re-exporting the source code is expected, some of the codegen/interface likely changed between your previous rnbo version and the latest.

I'm happy to hear that this new setup is working for you. So, presets work okay too?


BTW, this thread was a good reminder that I hadn't actually merged the juce template develop back into main for this release.

I have a cmake cache variable already set up to do all this there now: https://github.com/Cycling74/rnbo.example.juce/blob/299823760bdd28f2f73d525479844b2b306b4bd7/CMakeLists.txt#L26

So, for future readers, you can simply change the value of that instead of augmenting the build as suggested above.


Yes, presets seem to be working as expected, thanks again Alex!

So there are no other changes in the latest juce template I should look out for? Any straightforward way to get a buffer from RNBO to JUCE for analysis GUI? I know there's a workaround that's still slightly over my head. That's the next step for me.

Happy to hear that this is all working for you!

I don't think there are significant differences in the template.. you're probably good with what you have.

RE buffer output, I'd start a new thread with that topic and see if anyone has something to say about it?

rolling back Max to 8.5.4 resolved an issue with JUCE not seeing RNBO parameter updates.

ShelLuser

Hi gang!

First if all: huge compliments for RNBO~... I'm still very much a newbie at the time of writing but I already managed to build, and use, a very crude VST3 plugin and... it's just amazing. See: I use both Live (suite) & FL Studio side by side and while I have M4l in Live I always fantasized about what might happen if you could use Max with DAW's like 'FLS'. Now I can!

Question... everything works as expected but I'm wondering if it would be possible to 'overrule' the cloud compiler and instead use a locally set up compiler & framework? Or is this approach fully hardcoded into RNBO?
 
Note: I'm not necessarily asking how it can be done, merely if...


(edit): Thanks for the feedback!  I overlooked that option and I can easily imagine that this will make it much easier for people to filter stuff. Not gonna respond because we don't need a "ping", so... figured I'd just edit.

Hi gang!

First if all: huge compliments for RNBO~... 

would-it-be-possible-to-use-a-local-compiler-for-vst3-targets

So, yah, in hindsight this was a bit of an obvious question I suppose... 

First things first: the export sidebar features 2 options for "Source code" where the 1st option is C++ source code export. That's an obvious approach.

Second... having already experimented with VST exports (and being very impressed with it) you'll quickly learn that VST builds are powered by Juce; a "framework for audio applications and plug-in development". Best part is obvious: it can be downloaded and used for free, which is something that Cycling '74 is doing.

And there's my answer. Of course I still need a compiler for all this, but that's something which Visual Studio can handle for me.

Thanks for answering you own questions so that others with the same question can have a quick hint on what to to.
Instead of editing the title als ‚(solved)‘ there is a resolved tag you can give to the topic instead. ;)

Here is the JUCE adapter code that we bundle with the RNBO source code: https://github.com/Cycling74/rnbo.adapter.juce 
and here is a template project that we provide for creating your own customized JUCE/RNBO based projects: https://github.com/cycling74/rnbo.example.juce


Would it be possible to use a local compiler for VST3 targets?

hal bay

From documentation for get_all_notes_extended: 

"It is possible to optionally provide a single [dict] argument to this function, containing a single key-value pair: the key is "return" and the associated value is a list of the note properties as listed above in the discussion of the returned note dictionaries, e.g. ["note_id", "pitch", "velocity"]. The effect of this will be that the returned note dictionaries will only contain the key-value pairs for the specified properties, which can be useful to improve patch performance when processing large notes dictionaries."

I am trying to make use of new "get_all_notes_extended" function on a clip with notes in it. I want to make use of  filtering functionality through dict argument as explained in the docs.

I tried code below:

var filtered_notes = clip.call("get_all_notes_extended", { return: ["note_id", "pitch", "velocity"] });

post("filtered_notes", filtered_notes, "\n");

var notes_raw = clip.call("get_all_notes_extended");

As seen in the screenshot, "notes_raw" has expected values, however "filtered_notes" is 0.00 for the same clip object. So filtering with dict object with return key, as explained in the docs, is not working for me.

What am I missing here?
Thanks!


Edit: I also tried code below after learning about Dict objects of Max. However, it is still returning all of the keys like notes_raw above, and not filtering anything:

var note_filter = new Dict("note_filter").set("return", ["note_id", "pitch", "velocity"]);
var filtered_notes = clip.call("get_all_notes_extended", note_filter);

Screenshot 2023-09-16 at 14.18.04.png

From documentation for get_all_notes_extended: 

"It is possible to optionally provide a single [dict] argument to this function, ... 

filtering-properties-for-getallnotesextended-return-values-is-not-working

Solution found: Apparently, I should have use "set" method on a seperate function call, like below:

var note_filter = new Dict("note_filter");
note_filter.set("return", ["note_id", "pitch", "velocity"]);
clip.call("get_all_notes_extended", note_filter);

Filtering properties for get_all_notes_extended return values is not working

MartIn 

Hi,
i´m having problems observing a Button on Push and getting it´s current value.
Pretty sure this used to work  with "call get_control ButtonName" and sending the ID into an observer. 
I don´t want to grab the button and relieve it from its native functions, that´s why the "grab" call isn´t connected in the attached patch.

I wonder, am i doing something wrong here or is it broken?

PVObserver.amxd

Hi,
i´m having problems observing a Button on Push and getting it´s current value.
Pretty sure this used to work  with "call get_control ... 

observing-a-button-on-push-not-working-anymore

haha 😄 i am 🙋
(while i don't have a push... ) ...i observe that this thread was tagged with 'resolved', is it not?


👋🏻 
hey Raja, it’s not resolved, to my knowledge I didn’t put the Tag there,  weird :)

I´m happy to report that this was fixed in the latest release,  yay!

Observing a Button on Push not working anymore

bongo808

I am using autopattr, what does "patcher-dirty flag" mean ? Can I find anything about this in the max documentation ? I can't find anything.  Thank you !

I am using autopattr, what does "patcher-dirty flag" mean ? Can I find anything about this in the max documentation ? I can't find ... 

what-is-the-"patcher-dirty-flag"

A "dirty" patcher is a patcher that has been modified but not saved yet.

For reference for other readers, this is the documentation on setting the autopattr dirty attribute:

The word dirty, followed by a 1 or 0, enables or disables the patcher-dirty flag. The default is 0 (disabled). When enabled, the autopattr object will dirty the patch whenever its state changes.

"A "dirty" patcher is a patcher that has been modified but not saved yet." Ok, thank s a lot, seems obvious now...

What is the "patcher-dirty flag" ?

Marc Assenmacher

I am running into a problem that seemed so easy at first but drives me nuts now.

I need to manually sort a list during a live session either with drag&drop (preferred) or choosing an index and sorting with up/down-arrows (second choice).

1 2 3 5 4 67 8  >> move the 4 to the right 1 time

1 2 3 5 6 4 7 8 >> move the 4 to the right 1 time

1 2 3 5 6 7 4 8 >> move the 4 to the right 1 time

1 4 2 3 5 6 7 8 >> move the 4 to the left 5 tims

My Approach was using zl.swap and the incdec object, but my patch makes a mess. Everytime I switch directions, the old target makes another swap, making my patch useless.

Move-List-Item.maxpat

Hello,
I am running into a problem that seemed so easy at first but drives me nuts now.
I need to manually sort a list during a live ... 

manually-sort-list-with-draganddrop-or-updown-arrows

Somehow I missed this forum thread until now:

https://cycling74.com/forums/drag-and-drop-reordering-of-a-list 

For everyone interested: You can find drag and drop reordering of lists there. Amazing work!

while i am a fan of zl, this kind of job can be done best using coll´s insert, delete, merge, sort and swap messages, and combinations of them.

Manually sort list with drag&drop or up/down-arrows

soundyi

Hi there,
I stumbled upon a mighty issue (aka bug) where the output dimension of a texture is reduced to 512 x 512, using the @adapt 0 attribute in the patcher box of a JIT.GL.PIX.

The reason why I want to use adapt 0 : I use a second texture as a parameter for a JIT.GL.PIX patch, in this case a color - hence the dimensions of the second texture is 1x1  ... and an adaption would be a waste of memory.

The attached patcher demonstrates this - drag and drop the Shapes.png file on the dropfile object "target area" and you will notice that jit.fpsgui states "dim = 512x512", but it should be 1920x1080 (if you use the attached shapes.png file, which has this resolution - its shapes are white and the background is transparent, so don' let the "all white" preview fool you ;-) ... where 512x512 seems to be the default size for an empty texture.

Using @apdapt 0 with JIT.GEN I experienced no reduction of the matrices dimensions ... does the adapt attribute work differently in the matrix (CPU) & texture (GPU) domain?

Pix adapt issue.maxpat

Shapes.png

Hi there,
I stumbled upon a mighty issue (aka bug) where the output dimension of a texture is reduced to 512 x 512, using the @adapt 0 ... 

jitglpix-with-adapt-0-reduced-the-output-texture-dimension-to-512-x-512

well... it's not a bug. The @adapt 0 attribute lets jit.gl.pix ignore the input dimensions, no matter what inlet you use. So if you don't specify the @dim attribute directly in the jit.gl.pix object, it will resize every texture to 512x512.
Otherwise, if you use @adapt 1, it will adapt to the leftmost inlet resolution, but it will still resize all other textures. 
If you want to set a color without using a texture, you could set a param (see patch below)


Hi LSKA,

thanks for your info and the elegant solution for the "memory saving" ... using a parameter for a color instead of 1x1 texture is for sure the better solution.

And with your input I found out what fooled me ... names are just names and even if they are identical they may refer to different concepts ;-)

And for others who may stumble upon this :

In the case of jit.gl.pix, adapt is "native object" attribute : https://docs.cycling74.com/max8/refpages/jit.gl.pix#adapt.

In the case of jit.gen, adpat is (common) "MOP" attribute : https://docs.cycling74.com/max8/refpages/jit.gen#Matrix_Operator 

So, techniques that worked out for matrices really have to be revisted for textures ... got it ;-).

And by the way there is also a hint in the Gen docs "In jit.gl.pix , inputs can have different sizes." - so there is no need to use a "MOP adapt alike attribute" for jit.gl.pix in the first place, if one really needs to use another texture with different dimensions than the "primary / left most" texture.

See https://docs.cycling74.com/max8/vignettes/gen_overview at the bottom of the page - in the last paragraph.

JIT.GL.PIX with @adapt 0 reduced the output texture dimension to 512 x 512

michael watson

I've been experimenting with creating a visual show and I am trying to make a patch that can take a live input from a band or instrument and effect macros on vizzie devices.

For example, If I'm playing guitar the low pitches will create a lower value effect and if I play higher notes then the effect becomes more drastic.

Thanks in advance,
Fairly new to beautiful word of Max

Hey!
I've been experimenting with creating a visual show and I am trying to make a patch that can take a live input from a band or ... 

audio-frequency-to-video-data

[fzero~] might be a good starting point, at least for a single instrument. It reports pitch in Hz and peak amplitude (see help & reference).
The second outlet of the [retune~] object also reports pitch, as a signal.

Legend! Thanks for the prompt reply. 
That's done the trick. Is there a way to limit the range from the outputted data of [fzero~)?

There are many ways ....
You could try this one :

Perfect really appreciate that! As a learner coming from Live it is quite overwhelming but loving it.

I played around with the split values but nothing was  happening.
I am attempted to make the freq control any of those inputs on the vizzie feedr. 
This is where I am at so far. Having any suggestions?

Thanks again

open fzero help file and read how to configure it.
then use output values to scale them and make them usable.
why do you use this premade blocks ?

cycle~ and a float are easier to test fzero, than this modules,
which probably are totally unknown to you.




I am not familiar at all with Vizzie nor BEAP, but, as an example, here is a patch that makes the FEEDR's gain parameter move from 0. to 1. when the oscillator's frequency varies from 500 to 2000 Hz, and allows this frequency range to be modified dynamically.
Important note : as always, audio must be turned "on" to have audio signal generated.

Audio Frequency to Video Data

Ben Bracken

We are happy to announce the release of Max 8.5.6 and RNBO 1.2.1. 

On the Max side, we have a number of bug fixes, with improvements to fullscreen jit.window / thispatcher behavior, among other things. 

On the RNBO side, this release features many new improvements and bug fixes, with a focus on the Raspberry Pi target (can upload and manage the active patch), new ease objects, messages to the rnbo~ object for managing manual target exports, new integer utilities for codebox, among many other things.

As always, you can download the update here: 
https://cycling74.com/downloads

For a list of all Max 8.5.6 and RNBO 1.2.0 improvements, check out the change logs below, or here:
Max:
https://cycling74.com/releases/max/8.5.6 

RNBO:
https://cycling74.com/releases/rnbo/1.2.1 

udpreceive: 'defer' attribute to set message output to the low-priority queue

curve~ / line~: pending ramps are appropriately cancelled if interrupted

jit.dx.grab: support for outputting at device max framerate

jit.gl.handle: autohandle works with jit.gl.mesh

jit.gl.pix: fixed dim variable breakage with codebox

jit.gl.sketch: fixed usage of vertex colors

jit.gl.text: fixed draw errors on visiblity change

jit.window/jit.pwindow: fixed visibility in Max for Live

jit.world: fixed crash when enabling/disabling fullscreen or floating

live.menu: svg files are included as dependencies in Projects

Max for Live Editor: fixed issues when multiple Max instances are running

Max for Live: fixed crash when setting 'is_mpe' attribute

MIDI: improved support for multiple devices of the same type (Windows)

scala: fixed default scale "B" from 1200 to 1100 cents

thispatcher / jit.window: improved manual fullscreen behavior / fixed regressions

accum: added gen~ style features (resetmode, min, max)

C++ source code export: option to disable export of RNBO Engine

codebox: imod/iadd/imult integer operators

ease: new objects a la Max Package (ease, ease.list, ease.listinterp, ease~, ease.xfade~)

Max External Target: set default buffername for buffer attributes

number~: displaymode option (first / average)

oscquery runner: report if/when a preset was loaded

Raspberry Pi Target: add a device to the sidebar by address

Raspberry Pi Target: support for cloud compiling

Raspberry Pi Target: upload and manage multiple patches (hot swap)

Raspberry Pi Target: upload patcher without loading it

rnbo.remote: 'launchwebinterface' message to launch web interface from rnbo.remote

rnbo.remote: @patchername attribute to change current patch on device

rnbo~: dumptargetconfig message for getting the configuration dictionary for an export target (defaults c++)

rnbo~: export message for triggering a target export (defaults to c++)

codebox: allow const values in param declarations

codebox: fixed Int cast of bitwise operators

codebox: fixed parameter visibility in user defined functions

codebox: improved let/var declaration parsing

codebox: simplified Fixed Array initialization

counter/counter~: fixed crash when adding or removing ~ in object box

env2~: fixed an issue with release in hold mode

Export Sidebar: improved device add and delete UI

Export Sidebar: improved error message when Max is unauthed / authed via Live

Gen in RNBO: parameter with the name "t60" works

lockedphase~: removed (can still be used in codebox)

Max External Target: data can use Max buffers

Max External Target: overwrite warning references ‘external’

Max External Target: refpages in search path are not overwritten by new version of object

message: left inlet input always triggers output

param~: fixed issues when used in polyphonic top-level patchers

RNBO Object Inspector: shows enum attrs in a dropdown menu

RNBO patcher title bar: right-click title bar shows patcher filepath

samplerate: updates in the context of Max for Live

silencedetector~: fixed issues with cut off notes

silencedetector~: threshold and length options have some constraints

Themes: improved RNBO patcher legibility in non-default themes

VST Target: parameter automation improvements

We are happy to announce the release of Max 8.5.6 and RNBO 1.2.1. 

Cycling '74 Forums

We are happy to announce the release of Max 8.5.6 and RNBO 1.2.1.

vexpr or scale, set float or int, depending on input or mutiplication factor.

Super! Now I understood and it works just fine. Thanks to both of you. Is this somewhere described or maybe this falls more into the ...

Happy to hear that this is all working for you! I don't think there are significant differences in the template..

Here is the JUCE adapter code that we bundle with the RNBO source code: https://github.com/Cycling74/rnbo.adapter.juce and here is a ...

Solution found: Apparently, I should have use "set" method on a seperate function call, like below: var note_filter = new ...

I´m happy to report that this was fixed in the latest release, yay!

"A "dirty" patcher is a patcher that has been modified but not saved yet." Ok, thank s a lot, seems obvious now...

while i am a fan of zl, this kind of job can be done best using coll´s insert, delete, merge, sort and swap messages, and combinations of ...

Hi LSKA, thanks for your info and the elegant solution for the "memory saving" ...

Brilliant thank you!