Xpressive plugin #3259

gnudles · 2017-01-14T19:53:49Z

tresf added screenshot Feb 5 '17

Xpressive plugin allows you to write your own expressions for sound creation.

AndiEcker · 2017-01-17T09:04:23Z

+1 - WOW - looks like another interesting instrument 2017-01-14 19:53 GMT+00:00 gnudles <notifications@github.com>:

Xpressive plugin allows you to write your own expressions for sound creation. ------------------------------ You can view, comment on, or merge this pull request online at: #3259 Commit Summary - Small fixes: - First Preview of the X-Pressive Plugin - Xpressive Plug-In: - X-Pressive Plug-In: Few more fixes - Special version of ExprTk with modified license (BSL) for the LMMS project https://github.com/LMMS/lmms - Merge pull request #1 from ArashPartow/xpressive_plugin File Changes - *A* data/presets/X-Pressive/X-Pressive1.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-0> (21) - *A* data/presets/X-Pressive/X-Pressive10.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-1> (21) - *A* data/presets/X-Pressive/X-Pressive11.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-2> (21) - *A* data/presets/X-Pressive/X-Pressive12.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-3> (21) - *A* data/presets/X-Pressive/X-Pressive13.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-4> (21) - *A* data/presets/X-Pressive/X-Pressive14.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-5> (21) - *A* data/presets/X-Pressive/X-Pressive2.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-6> (21) - *A* data/presets/X-Pressive/X-Pressive3.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-7> (21) - *A* data/presets/X-Pressive/X-Pressive4.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-8> (21) - *A* data/presets/X-Pressive/X-Pressive5.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-9> (21) - *A* data/presets/X-Pressive/X-Pressive6.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-10> (21) - *A* data/presets/X-Pressive/X-Pressive7.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-11> (21) - *A* data/presets/X-Pressive/X-Pressive8.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-12> (21) - *A* data/presets/X-Pressive/X-Pressive9.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-13> (21) - *A* data/presets/X-Pressive/watsyn_like.xpf <https://github.com/LMMS/lmms/pull/3259/files#diff-14> (21) - *M* include/Graph.h <https://github.com/LMMS/lmms/pull/3259/files#diff-15> (1) - *M* plugins/CMakeLists.txt <https://github.com/LMMS/lmms/pull/3259/files#diff-16> (1) - *A* plugins/xpressive/CMakeLists.txt <https://github.com/LMMS/lmms/pull/3259/files#diff-17> (4) - *A* plugins/xpressive/artwork.png <https://github.com/LMMS/lmms/pull/3259/files#diff-18> (0) - *A* plugins/xpressive/expressive_plugin.cpp <https://github.com/LMMS/lmms/pull/3259/files#diff-19> (885) - *A* plugins/xpressive/expressive_plugin.h <https://github.com/LMMS/lmms/pull/3259/files#diff-20> (0) - *M* plugins/xpressive/exprfront.cpp <https://github.com/LMMS/lmms/pull/3259/files#diff-21> (0) - *M* plugins/xpressive/exprfront.h <https://github.com/LMMS/lmms/pull/3259/files#diff-22> (0) - *M* plugins/xpressive/exprtk.hpp <https://github.com/LMMS/lmms/pull/3259/files#diff-23> (0) - *M* plugins/xpressive/help_active.png <https://github.com/LMMS/lmms/pull/3259/files#diff-24> (0) - *M* plugins/xpressive/help_inactive.png <https://github.com/LMMS/lmms/pull/3259/files#diff-25> (0) - *M* plugins/xpressive/logo.png <https://github.com/LMMS/lmms/pull/3259/files#diff-26> (0) - *M* plugins/xpressive/o1_active.png <https://github.com/LMMS/lmms/pull/3259/files#diff-27> (0) - *M* plugins/xpressive/o1_inactive.png <https://github.com/LMMS/lmms/pull/3259/files#diff-28> (0) - *M* plugins/xpressive/o2_active.png <https://github.com/LMMS/lmms/pull/3259/files#diff-29> (0) - *M* plugins/xpressive/o2_inactive.png <https://github.com/LMMS/lmms/pull/3259/files#diff-30> (0) - *M* plugins/xpressive/w1_active.png <https://github.com/LMMS/lmms/pull/3259/files#diff-31> (0) - *M* plugins/xpressive/w1_inactive.png <https://github.com/LMMS/lmms/pull/3259/files#diff-32> (0) - *M* plugins/xpressive/w2_active.png <https://github.com/LMMS/lmms/pull/3259/files#diff-33> (0) - *M* plugins/xpressive/w2_inactive.png <https://github.com/LMMS/lmms/pull/3259/files#diff-34> (0) - *M* plugins/xpressive/w3_active.png <https://github.com/LMMS/lmms/pull/3259/files#diff-35> (0) - *M* plugins/xpressive/w3_inactive.png <https://github.com/LMMS/lmms/pull/3259/files#diff-36> (0) - *M* plugins/xpressive/wavegraph.png <https://github.com/LMMS/lmms/pull/3259/files#diff-37> (0) - *M* src/gui/PianoView.cpp <https://github.com/LMMS/lmms/pull/3259/files#diff-38> (0) - *M* src/gui/TrackContainerView.cpp <https://github.com/LMMS/lmms/pull/3259/files#diff-39> (0) - *M* src/gui/editors/AutomationEditor.cpp <https://github.com/LMMS/lmms/pull/3259/files#diff-40> (0) - *M* src/gui/widgets/Graph.cpp <https://github.com/LMMS/lmms/pull/3259/files#diff-41> (0) Patch Links: - https://github.com/LMMS/lmms/pull/3259.patch - https://github.com/LMMS/lmms/pull/3259.diff — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#3259>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADlPbl76fEXaRN1LIu-LmwDd1v8B7wR9ks5rSSfPgaJpZM4LjuVX> .

zonkmachine · 2017-01-17T11:23:42Z

😮 Totally tripping right now. It's so cool. Like if Bit Invader grew up and took classes in engineering.

gnudles · 2017-01-18T10:10:19Z

What should be done with this error? :
/tmp/cc62tQLz.s: Fatal error: can't write CMakeFiles/xpressive.dir/exprfront.cpp.obj: File too big

ArashPartow · 2017-01-18T11:06:17Z

What should be done with this error? :

The version of ld shipped with mingw64 is Literally Broken - last good version was circa GCC4.3. In short to get rid of the build error you need to do the following two things:

Step 1.
Add the following only to the mingw64 build options in CMake:

-Wa,-mbig-obj

Step 2.
In order to make the ExprTk TU section size smaller add the following defines either in the mingw64 CMake build line

-Dexprtk_disable_break_continue
-Dexprtk_disable_sc_andor
-Dexprtk_disable_enhanced_features
-Dexprtk_disable_string_capabilities
-Dexprtk_disable_rtl_io_file

Or #ifdef on mingw64 just prior to the #include "exprtk.hpp" and define

#define exprtk_disable_break_continue
#define exprtk_disable_sc_andor
#define exprtk_disable_enhanced_features
#define exprtk_disable_string_capabilities
#define exprtk_disable_rtl_io_file

I'd also review [Section 19 - Enabling & Disabling Features] of the documentation - specifically 19.2 Disabling Control-Flow Structures such as for-loop, while-loop etc. I'm guessing you don't want users invoking such functionality in their expressions.

ArashPartow

You've applied the defines to all compiler settings which is not what I suggested. It should only be applied to the mingw64 build

Furthermore you've applied the -Wa,-mbig-obj option to all mingw compilers it should only be the mingw64 compiler.

gnudles · 2017-01-18T13:46:55Z

oops, I'll replace
IF(MINGW_PREFIX)
with
IF(WIN64)
I invoke the parser with:

	parser_t::settings_store sstore;
	sstore.disable_all_logic_ops();
	sstore.disable_all_assignment_ops();
	sstore.disable_all_control_structures();
	parser_t parser(sstore);

So I don't understand why these defines should be used only in a certain build. Furthermore, I added them to reduce build time.
I do want all the users to be able to use
(t>0)&(t<1)
I saw that using the flag exprtk_disable_sc_andor changes the implementation of the "&" and "|" signs.
I hope it's OK.

tresf · 2017-01-18T17:21:22Z

@ArashPartow Thanks for such specific insight into this issue. From what I'm reading, we can set Step 1 globally, right?

https://github.com/LMMS/lmms/blob/master/cmake/modules/MinGWCrossCompile.cmake#L17-L22

, and

https://github.com/LMMS/lmms/blob/master/cmake/modules/MinGWCrossCompile.cmake#L61-L67

ArashPartow · 2017-01-18T17:23:16Z

I think it should be something like:

IF(MINGW_PREFIX and WIN64)

As WIN64 probably implies ALL windows 64 builds, perhaps including MSVC

Perhaps only assignment and loop control_structures as you might want if-statements/ternary-statements with logical operators to be made available:

parser_t::settings_store sstore;
sstore.disable_all_assignment_ops();
sstore.disable_control_structure(settings_t::e_ctrl_for_loop   );
sstore.disable_control_structure(settings_t::e_ctrl_while_loop );
sstore.disable_control_structure(settings_t::e_ctrl_repeat_loop);
parser_t parser(sstore);

So I don't understand why these defines should be used only in a certain build.

It's explained in my post above.

I do want all the users to be able to use
(t>0)&(t<1)
I saw that using the flag exprtk_disable_sc_andor changes the implementation of the "&" and "|" signs.

Yeah get rid of it from the build line. It wont save much for section size issues and it is probably better off to have it available/present.

ArashPartow · 2017-01-18T17:29:11Z

@tresf Nothing should be global. All of this is related to a broken version of mingw64 - So all the changes must only be specific to 64-bit builds using mingw64 and no other compilers or architectures.

The build changes presented are a hack/kludge to reduce the section size of the TU so as to allow the broken ld that is shipped with mingw64 to build the final executable binary.

In short the defines disable certain features such as expression optimizations so as to reduce the section size, hence it is not an option one would want to add to all builds as it drastically retards the performance of the expression engine.

tresf · 2017-01-18T17:36:33Z

@tresf Nothing should be global. All of this is related to a broken version of mingw64 - So all the changes must only be specific to 64-bit builds using mingw64 and no other compilers or architectures.

Please look at the lines that are linked. #3259 (comment)

ArashPartow · 2017-01-18T17:38:49Z

@tresf ooh yeah I missed this line:

https://github.com/LMMS/lmms/blob/master/cmake/modules/MinGWCrossCompile.cmake#L10

Should be all good if it is only within that IF(LMMS_BUILD_MSYS) scope.

tresf · 2017-01-18T17:43:11Z

Should be all good if it is only within that IF(LMMS_BUILD_MSYS) scope.

Great but to be precise, I believe it needs to be for both scopes linked or moved into a separate conditional.

ArashPartow · 2017-01-18T17:48:44Z

@tresf Now I'm confused. According to the build artifacts the 32-bit mingw build was successful however the 64-bit mingw build failed.

On line:

https://github.com/LMMS/lmms/blob/master/cmake/modules/MinGWCrossCompile.cmake#L18

The "IF(WIN64)" seems to suggest 64-bit but the comments around and in the scope and the "COMPILER32" stuff seem to suggest 32-bit.

I'll let you guys figure it out. However you end up doing it my recommendation is that it should only apply to 64-bit builds using mingw64.

tresf · 2017-01-18T18:04:01Z

Correct, we explicitly build the 32-bit host wrapper on 64-bit. Naturally, the description would have to change if this were added.

softrabbit · 2017-01-19T12:11:00Z

Actually, this shouldn't be (just) a plugin. @gnudles, could you get the expression evalution worked into the standard oscillator class(es)? That way it could be used in at least TripleOscillator or Organic as well, maybe some other plugin too.

(I haven't tried this plugin yet, but I like the idea.)

gnudles · 2017-01-20T10:57:59Z

@softrabbit , I think that the next step is to create a sound scripting language, compile it with llvm (or some other jit engine) on runtime, and make a generic plug-in infrastructure... But then you need to add llvm to the dependency list...
@tresf what's the bottom line?
@zonkmachine - it's better to use integrate(f) instead of t*f because it stays stable with frequency change (with base detuning). integrate(f) holds internal counter that counts f and return the sum divided by sample_rate.

zapashcanon · 2017-01-21T14:56:53Z

src/gui/widgets/Graph.cpp

+{
+	// store values in temporary array
+	QVector<float> temp = m_samples;
+	int graph_length = length();


const int graph_length = length();

zapashcanon · 2017-01-21T15:07:57Z

plugins/xpressive/exprfront.cpp

+		}
+		int m_rseed;
+	};
+}*/


A comment to explain why this is commented would be better I guess.

zapashcanon · 2017-01-21T15:10:20Z

plugins/xpressive/exprfront.cpp

+		const int data_size=sizeof(random_data)/sizeof(int);
+		int xi=(int)index;
+		int si=m_rseed%data_size;
+		int sa=m_rseed/data_size;


const int xi = [...]; const int si = [...]; const int sa = [...];

zapashcanon · 2017-01-21T15:12:47Z

plugins/xpressive/exprfront.cpp

+		{
+			return vec[ix];
+		}
+		int ixnext=(ix+1)%size;


const int ixnext=(ix+1)%size;

zapashcanon · 2017-01-21T15:13:01Z

plugins/xpressive/exprfront.cpp

+		//to disable interpolation uncomment:
+		//return static_cast<WaveSample*>(w)->samples[ix];
+
+		float xfrc=fraction(x);


const float xfrc=fraction(x);

zapashcanon · 2017-01-21T15:25:45Z

plugins/xpressive/expressive_plugin.cpp

+		ExprFront expr(text.constData());
+		float t,f=10,key=2,rel=0,v=0.5;
+		unsigned int i;
+		unsigned int sample_rate=m_raw_graph->length();


float t; const float f=10,key=2,rel=0,v=0.5; unsigned int i; const unsigned int sample_rate=m_raw_graph->length();

zapashcanon · 2017-01-21T15:28:21Z

plugins/xpressive/expressive_plugin.cpp

+		float t,f=10,key=2,rel=0,v=0.5;
+		unsigned int i;
+		unsigned int sample_rate=m_raw_graph->length();
+		expr.add_variable("t", t);


t not initialized ?

It works, but I'll initialize it for clarity.

I mean, the value is set before it's evaluated...

zapashcanon · 2017-01-21T15:29:19Z

plugins/xpressive/expressive_plugin.cpp

+
+static void clearGraph(graphModel * g)
+{
+	int length = g->length();


const int length = g->length();

zapashcanon · 2017-01-21T15:31:27Z

plugins/xpressive/expressive_plugin.cpp

+	m_panning2.loadSettings(_this,"PAN2");
+	m_relTransition.loadSettings(_this,"RELTRANS");
+
+	int size = 0;


Not sure for this one, as I don't know the effect of the next function calls, but:

const int size = 0;

Let's keep it that way...

zapashcanon · 2017-01-21T15:35:47Z

plugins/xpressive/expressive_plugin.cpp

+
+	float o1,o2,pn1,pn2;
+	float new_freq=nph->frequency();
+	float freq_inc=(new_freq-frequency)/_frames;


const float new_freq=nph->frequency(); const float freq_inc=(new_freq-frequency)/_frames;

tresf · 2017-02-05T17:55:37Z

Is this ready for merge?

zonkmachine · 2017-02-05T20:36:51Z

Is this ready for merge?

I think it is.

I've seen a very minor issue. The presets are not saved with unique names. It's not obvious when you save a preset that the file name and the preset name isn't necessarily the same. One could argue that this is something we could change actually.

zonkmachine

Please re-save the pre-sets with their chosen names and not the default name="X-Pressive"

zonkmachine · 2017-02-05T20:44:29Z

data/presets/X-Pressive/X-Pressive1.xpf

+<!DOCTYPE lmms-project>
+<lmms-project version="1.0" creator="LMMS" creatorversion="1.2.0-rc2.6" type="instrumenttracksettings">
+  <head/>
+  <instrumenttracksettings muted="0" type="0" name="X-Pressive" solo="0">


name="X-Pressive"

The preset name that shows up when you load an instrument isn't the name of the .xpf but this setting. You need to edit the name in the rename dialogue on the top of the plugin.

Tell you the truth, I'm not sure that those presets should be merge. And even if they do, I think they should have more meaningful naming... I'm not so good at naming sounds so... if someone volunteers...
Also if someone made nice presets which he wants them to be merged...

The user will want to have some good presets to look at for this one. In any case I don't think this should hold back a merge. You or someone else can tweak the names later and it's easier for people to try it out and make presets if it's merged anyway.

Can be fixed later

gnudles · 2017-02-08T18:30:08Z

Oh BTW, what's the best way for a plugin to deallocate memory on shut-down? Because I need to deallocate the help window.

jasp00

Although this is work in progress, you should start following the coding conventions.

It is required to include the Boost Software License.

jasp00 · 2017-02-14T17:44:19Z

plugins/xpressive/expressive_plugin.h

+{
+	Q_OBJECT
+public:
+	static expressiveHelpView *getInstance()


what's the best way for a plugin to deallocate memory on shut-down?

We should add some construction and destruction methods for plug-ins. In the meantime, you may try implementing __attribute__((destructor)) static void module_destroy() in your class.

jasp00 · 2017-02-14T17:49:31Z

include/Graph.h

@@ -145,6 +145,8 @@ class EXPORT graphModel : public Model
 		return( m_samples.data() );
 	}

+	void convolve(const float *_convolution, const int _conv_length, const int _center_offset);


Coding conventions: Parameters are not prefixed with underscore.

jasp00 · 2017-02-14T17:52:52Z

plugins/CMakeLists.txt

@@ -79,6 +79,7 @@ IF("${PLUGIN_LIST}" STREQUAL "")
 		waveshaper
 		vibed
 		zynaddsubfx
+		xpressive


You should sort by name. I know it is not sorted.

jasp00 · 2017-02-14T18:01:30Z

plugins/xpressive/CMakeLists.txt

@@ -0,0 +1,7 @@
+INCLUDE(BuildPlugin)
+
+SET(CMAKE_CXX_FLAGS "-Dexprtk_disable_sc_andor -Dexprtk_disable_enhanced_features -Dexprtk_disable_break_continue -Dexprtk_disable_comments -Dexprtk_disable_string_capabilities -Dexprtk_disable_rtl_io_file -Dexprtk_disable_rtl_vecops ${WERROR_FLAGS}")


Please include previous value of CMAKE_CXX_FLAGS. You should apply all these definitions only for the MinGW-w64 build, as recommended at #3259 (comment).

But I want to disable those features on any platform...

Oh... now it's not compiling because exprtk.hpp still throws exception in line 2733
@ArashPartow, do you have any idea?
@tresf do I have to follow that -fno-exception flag?
????

@gnudles which plaform are you building where you see that exception being thrown?

The exception is not being thrown. But I can't compile it with the flag -fno-exceptions

jasp00 · 2017-02-14T18:05:17Z

plugins/xpressive/expressive_plugin.cpp

+/*
+ * expressive_plugin.cpp - instrument which uses a mathematical formula parser
+ *
+ * Copyright (c) 2016-2017 Orr Dvori


The convention is to add an email address, but it is not strictly necessary.

jasp00 · 2017-02-14T18:28:34Z

plugins/xpressive/expressive_plugin.h

+
+	BoolModel m_exprValid;
+
+	friend class expressiveView;


Be aware this is not a clean model/view separation.

I saw it in bitInvader (I started by copying bitInvader)...

If you have some spare time, please use inline getters and setters instead of friend.

jasp00 · 2017-02-14T18:29:44Z

plugins/xpressive/expressive_plugin.h

+	static QString HelpText;
+
+};
+class SubWindow;


Please move these declarations to the top.

jasp00 · 2017-02-14T18:31:20Z

plugins/xpressive/expressive_plugin.cpp

+		delete exprO2;
+}
+float saw_wave(float);
+float sin_wave(float);


Where are these functions used?

leftovers...

jasp00 · 2017-02-14T18:48:05Z

plugins/xpressive/exprfront.cpp

+{
+	static inline float process()
+	{
+		return 1.0f - rand() * 2.0f / RAND_MAX;


rand is not thread-safe. It should be replaced later.

jasp00 · 2017-02-14T18:50:09Z

plugins/xpressive/exprfront.cpp

+	size_t last_pos=0;
+	size_t count = 0;
+	size_t len=strlen(needle);
+	while (1)


You should use the standard true constant.

gnudles · 2017-02-23T13:44:18Z

I fixed everything that @jasp00 noted,

Please include previous value of CMAKE_CXX_FLAGS

but unfortunately when I added the previous value, it couldn't be compiled, because there is still another throw statement in exprTk that slipped through all those disabling defines.
When it will be fixed, I think we are ready for merge...

simonvanderveldt · 2017-02-25T11:15:06Z

Cool stuff!
Small question: Shouldn't new plugins use the new styling? See #2831
Or will this be done after this PR is merged?

zonkmachine · 2017-02-25T19:59:56Z

I fixed everything that @jasp00 noted,

@gnudles Awesome! Push the changes.

Please include previous value of CMAKE_CXX_FLAGS

but unfortunately when I added the previous value, it couldn't be compiled, because there is still another throw statement in exprTk that slipped through all those disabling defines.
When it will be fixed, I think we are ready for merge...

Is there an upstream issue about this we can keep track of?

Small question: Shouldn't new plugins use the new styling? See #2831
Or will this be done after this PR is merged?

#2831 Is concerning the native effect plugins specifically.

simonvanderveldt · 2017-02-25T20:38:50Z

#2831 Is concerning the native effect plugins specifically.

@zonkmachine yeah, I always forget there's a distinction between instruments and effects :x

gnudles · 2017-02-28T09:42:10Z

@gnudles Awesome! Push the changes.

But if I do, it won't compile

Is there an upstream issue about this we can keep track of?

What do you mean?

zonkmachine · 2017-06-19T16:45:23Z

The second (+) line above is indeed in exprtk.hpp upstream but neither of the lines is anywhere in the versions I have.

Edit: apart from the backed up original...

gnudles · 2017-06-20T10:16:04Z

@zonkmachine, you can always get the newest exprtk.hpp from upstream. Thanks for the fixes BTW.
The exprtk.hpp in the plugin is the original file without any change, so it should be OK to take the newest.

zonkmachine · 2017-06-20T13:30:16Z

Aight, upated! 😉

zonkmachine · 2017-06-25T14:02:39Z

Not good to merge. Due to the rebasing to stable-1.2 this includes some commits from master that are not present in stable-1.2, such as de3b344 and 9f905bc.

I believe this is fixed now. Merge?

lukas-w · 2017-06-25T14:49:23Z

Merge?

👍

zonkmachine · 2017-06-26T15:23:55Z

@tresf This is a big one. Does anything get squashed or just merge it as it is?

tresf · 2017-06-26T15:59:34Z

@zonkmachine if we wanted we could isolate the Graph additions and plugin cmake cleanup, but they're actually relatively small, I vote squash 'em all.

zonkmachine · 2017-06-26T17:27:56Z

I vote squash 'em all.

I'm fine with that but how about keeping the two main contributors apart and squashing the exprtk.hpp commits by Arash (amended by my update) separately so it gets recognized on the contributors list. I'm a bit sentimental that way. I tested it in a separate branch and the rebase comes out with no bugs and there is no diff against gnudles:xpressive_plugin.

It would look something like:

commit fa59472eed44f5e40d011bf3366874949720be89
Author: Orr Dvori <dvoreader@gmail.com>
Date:   Wed May 17 22:43:31 2017 +0300

    X-Pressive plugin - added graph drawing feature

commit acee8d3c6aefdcb7eaa047ce83881e3852ee2701
Author: ArashPartow <partow@gmail.com>
Date:   Sat Jan 14 05:21:23 2017 +1100

    Update ExprTk to tip of branch

commit 9b646c5c01266ef9229a6ff3f49c4ba32fcdb356
Author: Orr Dvori <dvoreader@gmail.com>
Date:   Sun Jan 8 14:31:01 2017 +0200

    X-Pressive Plugin
    
    Synthesizer plugin based on Bitinvader and 'The C++ Mathematical
    Expression Toolkit Library (ExprTk)'
    
    Some initial functions.
    
    available keys:
    f- note's frequency. available only in the output expressions
    t- time in seconds. in the Waves (W1,W2,W3) it's in the range [0,1)
    key- the note's keyboard key. available only in the output expressions.
    v- the note's velocity (divided by 255.0 so it is in the range [0,

...

tresf · 2017-06-26T17:31:05Z

In my experience squash maintains the contributors.

zonkmachine · 2017-06-26T17:37:43Z

Ah, OK! Then just a Squash and merge.

zonkmachine · 2017-06-26T18:08:26Z

In my experience squash maintains the contributors.

It doesn't look like it.

tresf · 2017-06-26T18:14:47Z

What is the concern? It gives him credit?

(Edit: If you're concerned about @ArashPartow I understand. I would have expected him to be in there as well.)

zonkmachine · 2017-06-26T18:18:47Z

What is the concern? It gives him credit?

Yes, gnudles but not Arash. It's not a big concern though, I'm more trying to get to grips with git and github.

ArashPartow · 2017-06-28T10:30:35Z

@zonkmachine - Gnudles did all the heavy lifting, so I'm all good with that. :D

gnudles · 2017-06-30T08:57:41Z

Maybe we should credit Arash in the plugin itself in the help window.
BTW, shouldn't it be merged to stable1.2?

* First Preview of the X-Pressive Plugin (exprtk.hpp is not included, get it from my exprtk fork in the branch internal_data_functions) available keys: f- note's frequency. available only in the output expressions t- time in seconds. in the Waves (W1,W2,W3) it's in the range [0,1) key- the note's keyboard key. available only in the output expressions. v- the note's velocity (divided by 255.0 so it is in the range [0,1]). available only in the output expressions. rel- gives 0 while the key is holded, and 1 after the key release. available only in the output expressions. A1,A2,A3- general purpose knobs (you can control them with the automations). available only in the output expressions. W1,W2,W3- precalculated wave forms. can be also load from file. you can use them only in the output expressions available functions: cent(x)- gives pow(2,x/1200) rand()- random number generator. in range [-1,1). each call gives other value. randv(i)- random vector (with pseudo infinite integer cells). the values are in range [-1,1). it's stays consistent only across the note playback. so each note playback will get other vector (even on the same key). sinew- sine wave with period of 1. saww- saw wave with period of 1. squarew- square wave with period of 1. trianglew- triangle wave with period of 1. expw- exponent wave with period of 1. expnw- another exponent wave with period of 1. moogw- moog wave with period of 1. moogsaww- moog-saw wave with period of 1. you can use * % ^ / + - pow sin log pi etc. * Xpressive Plug-In: Added Release transition knob that control the "rel" variable. (the duration of transit from 0 to 1) Fixed some problems in the displays. (update display when changing A1,A2,A3, clear display with invalid expression. * X-Pressive Plug-In: Few more fixes Changed the callbacks in exprfront.cpp to be templated. Added help. Added ExprTk.hpp. some bug fixes (inf issues). Added integrate function. * Special version of ExprTk with modified license (BSL) for the LMMS project https://github.com/LMMS/lmms * Xpressive Plug-In- fixed some building errors. Added the "e" euler's constant. * Xpressive Plug-In - fix mingw64 issues * X-Pressive Plug-in: Added "trel" (time since release) variable. The integrate function can now have unlimited usage. Added selective interpolation per wave. Improved a little the random vector function. Some other improvements, code cleaning, etc... * Xpressive Plug-In: move clearGraph definition into Graph.cpp. fixed compilation errors. (oops..) * X-Pressive plug-in: updated presets names * X-Pressive plug-in added semitone function, added sample-rate variable * X-Pressive plug-in, code cleaning, changed the rendering function to achieve performace gain. * X-Pressive plug-in - fix the string counting function * X-Pressive plug-in - until somebody will find a better solution, exprtk.hpp is patched under the name exprtk.patched.hpp ... * X-Pressive plug-in - fix compiling errors. * X-Pressive plug-in - added patch file for exprtk.hpp, added last function that gives last calculated samples. moved ExprSynth to be with ExprFront for performance reasons. * X-Pressive plugin - moved the patched file back to the source tree, added .gitignore file.. * X-Pressive plugin - fix compilation error. (isnan isinf) * X-Pressive plugin - tried to fix embed.cpp problem, added new variable to the parser (tempo) * X-Pressive plugin - fixed cmake script * X-Pressive plugin - updated the license and the diff file. * Updates to ExprTk * Added return statement enable/disable via parser settings Added exprtk_disable_return_statement macro for disabling return statements and associated exceptions at the source code level. * X-Pressive plugin - updated CMakeLists.txt to use the correct flags on each platform. also added exprtk.hpp as a dependency for the patch command. Updated the exprtk diff file. * X-Pressive plugin - moved the enhanced features flag to the WIN64 installation. * X-Pressive plugin - another fix for CMakeLists.txt * Minor updates to ExprTk Updated multi-sub expression operator to return final sub-expression type. Updates to exprtk_disable_return_statement macro for disabling return statements and associated exceptions at the source code level. * X-Pressive plug-in - added try-block around exprtk calls and enabled the -fexceptions flag, so patch file is no longer needed. * X-Pressive plug-in - small fix in CMakeLists.txt * Update ExprTk to tip of branch. * X-Pressive plugin - added graph drawing feature.. * Updating exprtk.hpp to the latest upstream version

zonkmachine · 2017-10-01T19:04:32Z

BTW, shouldn't it be merged to stable1.2?

I agree. In the works here: #3845

gnudles · 2018-09-22T20:21:56Z

You know, I think it can be nice to have a JavaScript version of this plugin. It can use the built-in JS support in Qt via QtScript. The performance issue should be considered, but it seems a very nice idea.
Recently, I did a demo with Qt outside of LMMS to play a note with QtScript (JS), and it responded well.
What do you think?

Unesty · 2019-09-01T19:08:06Z

How to add external sample to W?

gnudles · 2019-09-03T06:42:54Z

you have the load button

Unesty · 2019-09-09T17:52:04Z

I tried to load sample and it became distorted. Is that how supposed to be?

gnudles · 2019-09-09T19:33:53Z

the wave is loaded to a buffer with 4096 samples or so (4096 floats or shorts, can't recall). 4096 is a huge buffer compared to other plugins in lmms. but if you load a long sample it will become distorted (resampled). it is not a bug, it's a feature :)

* First Preview of the X-Pressive Plugin (exprtk.hpp is not included, get it from my exprtk fork in the branch internal_data_functions) available keys: f- note's frequency. available only in the output expressions t- time in seconds. in the Waves (W1,W2,W3) it's in the range [0,1) key- the note's keyboard key. available only in the output expressions. v- the note's velocity (divided by 255.0 so it is in the range [0,1]). available only in the output expressions. rel- gives 0 while the key is holded, and 1 after the key release. available only in the output expressions. A1,A2,A3- general purpose knobs (you can control them with the automations). available only in the output expressions. W1,W2,W3- precalculated wave forms. can be also load from file. you can use them only in the output expressions available functions: cent(x)- gives pow(2,x/1200) rand()- random number generator. in range [-1,1). each call gives other value. randv(i)- random vector (with pseudo infinite integer cells). the values are in range [-1,1). it's stays consistent only across the note playback. so each note playback will get other vector (even on the same key). sinew- sine wave with period of 1. saww- saw wave with period of 1. squarew- square wave with period of 1. trianglew- triangle wave with period of 1. expw- exponent wave with period of 1. expnw- another exponent wave with period of 1. moogw- moog wave with period of 1. moogsaww- moog-saw wave with period of 1. you can use * % ^ / + - pow sin log pi etc. * Xpressive Plug-In: Added Release transition knob that control the "rel" variable. (the duration of transit from 0 to 1) Fixed some problems in the displays. (update display when changing A1,A2,A3, clear display with invalid expression. * X-Pressive Plug-In: Few more fixes Changed the callbacks in exprfront.cpp to be templated. Added help. Added ExprTk.hpp. some bug fixes (inf issues). Added integrate function. * Special version of ExprTk with modified license (BSL) for the LMMS project https://github.com/LMMS/lmms * Xpressive Plug-In- fixed some building errors. Added the "e" euler's constant. * Xpressive Plug-In - fix mingw64 issues * X-Pressive Plug-in: Added "trel" (time since release) variable. The integrate function can now have unlimited usage. Added selective interpolation per wave. Improved a little the random vector function. Some other improvements, code cleaning, etc... * Xpressive Plug-In: move clearGraph definition into Graph.cpp. fixed compilation errors. (oops..) * X-Pressive plug-in: updated presets names * X-Pressive plug-in added semitone function, added sample-rate variable * X-Pressive plug-in, code cleaning, changed the rendering function to achieve performace gain. * X-Pressive plug-in - fix the string counting function * X-Pressive plug-in - until somebody will find a better solution, exprtk.hpp is patched under the name exprtk.patched.hpp ... * X-Pressive plug-in - fix compiling errors. * X-Pressive plug-in - added patch file for exprtk.hpp, added last function that gives last calculated samples. moved ExprSynth to be with ExprFront for performance reasons. * X-Pressive plugin - moved the patched file back to the source tree, added .gitignore file.. * X-Pressive plugin - fix compilation error. (isnan isinf) * X-Pressive plugin - tried to fix embed.cpp problem, added new variable to the parser (tempo) * X-Pressive plugin - fixed cmake script * X-Pressive plugin - updated the license and the diff file. * Updates to ExprTk * Added return statement enable/disable via parser settings Added exprtk_disable_return_statement macro for disabling return statements and associated exceptions at the source code level. * X-Pressive plugin - updated CMakeLists.txt to use the correct flags on each platform. also added exprtk.hpp as a dependency for the patch command. Updated the exprtk diff file. * X-Pressive plugin - moved the enhanced features flag to the WIN64 installation. * X-Pressive plugin - another fix for CMakeLists.txt * Minor updates to ExprTk Updated multi-sub expression operator to return final sub-expression type. Updates to exprtk_disable_return_statement macro for disabling return statements and associated exceptions at the source code level. * X-Pressive plug-in - added try-block around exprtk calls and enabled the -fexceptions flag, so patch file is no longer needed. * X-Pressive plug-in - small fix in CMakeLists.txt * Update ExprTk to tip of branch. * X-Pressive plugin - added graph drawing feature.. * Updating exprtk.hpp to the latest upstream version

ArashPartow suggested changes Jan 18, 2017

View reviewed changes

zapashcanon suggested changes Jan 21, 2017

View reviewed changes

zonkmachine previously requested changes Feb 5, 2017

View reviewed changes

jasp00 reviewed Feb 14, 2017

View reviewed changes

PaulBatchelor mentioned this pull request Feb 18, 2017

Implement Sporth Editor as LMMS effect plugin #3368

Closed

Updating exprtk.hpp to the latest upstream version

5e66f1b

zonkmachine merged commit dff76b2 into LMMS:master Jun 26, 2017

PhysSong mentioned this pull request Oct 26, 2018

Multitap Echo sometimes crashes when filter steps above 16 are changed #4679

Closed

		@@ -0,0 +1,7 @@
		INCLUDE(BuildPlugin)

		SET(CMAKE_CXX_FLAGS "-Dexprtk_disable_sc_andor -Dexprtk_disable_enhanced_features -Dexprtk_disable_break_continue -Dexprtk_disable_comments -Dexprtk_disable_string_capabilities -Dexprtk_disable_rtl_io_file -Dexprtk_disable_rtl_vecops ${WERROR_FLAGS}")

Xpressive plugin #3259

Xpressive plugin #3259

Conversation

gnudles commented Jan 14, 2017 • edited by tresf Loading

AndiEcker commented Jan 17, 2017 via email

zonkmachine commented Jan 17, 2017

gnudles commented Jan 18, 2017

ArashPartow commented Jan 18, 2017 • edited Loading

ArashPartow left a comment • edited Loading

Choose a reason for hiding this comment

gnudles commented Jan 18, 2017 • edited Loading

tresf commented Jan 18, 2017

ArashPartow commented Jan 18, 2017 • edited Loading

ArashPartow commented Jan 18, 2017 • edited Loading

tresf commented Jan 18, 2017

ArashPartow commented Jan 18, 2017 • edited Loading

tresf commented Jan 18, 2017

ArashPartow commented Jan 18, 2017

tresf commented Jan 18, 2017

softrabbit commented Jan 19, 2017

gnudles commented Jan 20, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tresf commented Feb 5, 2017

zonkmachine commented Feb 5, 2017

zonkmachine left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gnudles commented Feb 8, 2017

jasp00 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gnudles Feb 16, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gnudles commented Feb 23, 2017 • edited Loading

simonvanderveldt commented Feb 25, 2017

zonkmachine commented Feb 25, 2017

simonvanderveldt commented Feb 25, 2017

gnudles commented Feb 28, 2017

zonkmachine commented Jun 19, 2017

gnudles commented Jun 20, 2017 • edited Loading

zonkmachine commented Jun 20, 2017 • edited Loading

zonkmachine commented Jun 25, 2017

lukas-w commented Jun 25, 2017

zonkmachine commented Jun 26, 2017

tresf commented Jun 26, 2017

zonkmachine commented Jun 26, 2017

tresf commented Jun 26, 2017

zonkmachine commented Jun 26, 2017

zonkmachine commented Jun 26, 2017

tresf commented Jun 26, 2017 • edited Loading

zonkmachine commented Jun 26, 2017

gnudles commented Jan 14, 2017 •

edited by tresf

Loading

ArashPartow commented Jan 18, 2017 •

edited

Loading

ArashPartow left a comment •

edited

Loading

gnudles commented Jan 18, 2017 •

edited

Loading

ArashPartow commented Jan 18, 2017 •

edited

Loading

ArashPartow commented Jan 18, 2017 •

edited

Loading

ArashPartow commented Jan 18, 2017 •

edited

Loading

gnudles Feb 16, 2017 •

edited

Loading

gnudles commented Feb 23, 2017 •

edited

Loading

gnudles commented Jun 20, 2017 •

edited

Loading

zonkmachine commented Jun 20, 2017 •

edited

Loading

tresf commented Jun 26, 2017 •

edited

Loading

ArashPartow commented Jun 28, 2017 •

edited

Loading

zonkmachine commented Oct 1, 2017 •

edited

Loading

gnudles commented Sep 22, 2018 •

edited

Loading