Go Back   Cockos Incorporated Forums > REAPER Forums > ReaScript, JSFX, REAPER Plug-in Extensions, Developer Forum

Reply
 
Thread Tools Display Modes
Old 01-22-2019, 10:49 AM   #121
JamesPeters
Human being with feelings
 
Join Date: Aug 2011
Location: Near a big lake
Posts: 3,943
Default

That gui glitch looks like the same thing I had noticed and could not repeat. (Just a bit more complex of a glitch because it's a more complex curve that's being displayed in this example.) Same graph.
JamesPeters is offline   Reply With Quote
Old 01-22-2019, 11:50 AM   #122
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
That UI glitch worries me as it smells of some memory corruption (which is notoriously tricky to debug). Bri1 has had some similar ones too in the past. Especially the fact that the zero is not at zero strikes me as odd. But it's also a clue as to where to look.

Even remembering the filter/waveshaper/mode you came from and the filter/waveshaper/mode you went to would be helpful information. Even if not fully reproducible, it reduces the number of things I have to play with.

What do you mean by delay jumps? pdc_delay? Do you remember which filter it was on? For debugging pdc_delay issues relevant info is: filter type, FIR on or off, oversampling ratio and RMS time.
If it helps, the preset you see where the graphic glitch happened is exactly the preset that caused it. I did not change anything. I was just about to edit the input waveshaper as it happened. And it showed up several times when I adjusted the resonance. BUT it was the very first time that it happened. And I did a lot with Filther. So ...

And yes, I meant the pdc-delay. This happens after switching several presets. I will take closer look if I can recognize a pattern. If yes, feedback will show up here.

EDIT: As far as I remember it did not change the sound. Though I'm not sure about this it could be a hint that its only graphical.
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆

Last edited by Eliseat; 01-22-2019 at 11:59 AM.
Eliseat is offline   Reply With Quote
Old 01-22-2019, 04:26 PM   #123
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
If it helps, the preset you see where the graphic glitch happened is exactly the preset that caused it.

hey- can you please try capturing some sessions with video software=usefull for retracing steps.. =)
keep the love flowing. filther is a winner already..honing is key now.
the more heads that focus on single issues=better for fixings.
Bri1 is offline   Reply With Quote
Old 01-23-2019, 05:15 AM   #124
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

- Added gain linking. Thing is though, I think this feature is only useful for the linear filters (when you have no waveshaper available), because the other ones naturally saturate themselves. Still, it's there if someone wants to use it. The state is not stored, since I didn't feel it was worth spending a slider slot on (I'm severely limited in sliderslots atm).

- Fixed bug Eliseat mentioned with the modulation not being synced properly. The issue was that the modulation was not taking into account the lookahead properly. This has been fixed now.

Just a word about the pdc_delay, it may be that you had the RMS value set to a really high value. Note that one of the x4's changed to x8. Larger RMS integration time will lead to a larger pdc_delay. If you're in MIDI mode, I would advise to set it as low as possible to get the least amount of pdc_delay.

The only thing I can think of so far for the graphical glitch is that for some reason init is not called, and the memory address the spline shaper is looking at is wrong. If it happens again, try doubleclicking the logo and see if that fixes it. If that does or doesn't will be another valuable clue to me.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]

Last edited by sai'ke; 01-23-2019 at 05:28 AM.
sai'ke is offline   Reply With Quote
Old 01-23-2019, 07:44 AM   #125
Sju
Human being with feelings
 
Join Date: Jun 2015
Posts: 685
Default

Quote:
Originally Posted by sai'ke View Post
- Added gain linking.
Thank you, this is really useful! I like that it doesn't just snap the sliders to the opposite value, but instead preserves the relative positions.
Sju is offline   Reply With Quote
Old 01-23-2019, 09:47 AM   #126
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
...
Just a word about the pdc_delay, it may be that you had the RMS value set to a really high value. Note that one of the x4's changed to x8. Larger RMS integration time will lead to a larger pdc_delay. If you're in MIDI mode, I would advise to set it as low as possible to get the least amount of pdc_delay...
Ah, this could be it. I never took my focus on rms values and sometimes I just altered an existing preset without noticing if rms was set or not. I will take a look at this the next time.
BUT the problem is, those huge pdc delays don't go away if I choose another preset. Even if I switch from a preset with about 100 spls to a preset which burts it to 26000 spls and then back, the 100 spls turns into 26000 or - in other words - the 26000 stays the whole time. ¯\_(ツ)_/¯
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-23-2019, 03:02 PM   #127
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Hmmm. Could you attach the two presets in question? I'd like to study them.

Changelog:
+ Added ability to activate dynamics and lfo for specific filters separately, making sure that legacy presets are still preserved (the first bit being both filters and LFO's active).
+ Added a few afterimage lines to the plot.
+ Optimized cem filter.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]

Last edited by sai'ke; 01-24-2019 at 03:10 AM.
sai'ke is offline   Reply With Quote
Old 01-24-2019, 10:22 AM   #128
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
Hmmm. Could you attach the two presets in question? I'd like to study them.

Changelog:
+ Added ability to activate dynamics and lfo for specific filters separately, making sure that legacy presets are still preserved (the first bit being both filters and LFO's active).
+ Added a few afterimage lines to the plot.
+ Optimized cem filter.
Oh, how did you do that so fast? I will test it later for sure.

But for now I have news about the graphic glitch. I noticed that sometimes changing the preset name or switching thru the presets suddenly increases cpu usage. It seems to absolutely happen by chance. I have two videos which show how this takes place on my system.

https://mega.nz/#!DjY1gQrK!jHFndlXaD...RR6yXbbEPzQ7GU

https://mega.nz/#!2jQTxYJZ!m9dnmQp-u...pXSUFhKmoYBYXg

I did exactly the same a few minutes later after restarting Reaper but this time I was able to do all those things without getting the cpu peak.

Relating to the pdc problem: I can't send you presets because I already had them resaved without rms x4 and x8. Anyway, I can try to recreate this.

Holy cow! If you really did the DYN/LFO separation ... I can't wait to check this. Many thanks! This multiplies the possibilities by factor 1000! (◕‿◕)♡
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-25-2019, 06:17 AM   #129
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Magic!

Thanks!

That CPU spike is interesting. I see three things that may have caused it.

- Denormals.
Something in the code was spitting out denormals, slowing everything down. This is possible. I had ext_no_denorm set to 1, but I set it back to zero now. This should at least make sure that filther doesn't spit out denormals afaik. The weird thing is, if this was the case, switching presets would likely get rid of it.
- Some loop was going on too long.
MIDI and cutoff/resonance automation is handled with a small buffer. This buffer is terminated with a -1337. If for whatever reason (it never should, but who knows) it managed to get beyond that marker, the loop could go on indefinitely. I kind of doubt that this is the case though since these variables are only ever touched in @block and @sample and afaik, these are on the same thread and won't have any weird race conditions. Anyways, to be extra sure, I now initialize the whole block with -1337.
- Huge pdc value causing havoc. I've had some performance issues in reaper before with huge pdc values.

If it happens again. Please try the following in this order:
1a. Toggle lookahead, see if it resolves the issue.
1b. Toggle FIR, see if this resolves it.
1c. Toggle the waveshaper off, see if this resolves it.
1d. Toggle all the dynamics processing off (modulation / dynamic), see if this resolves it.
1e. Toggle the filters off, see if this resolves it.
2. Seek to a certain time point in the arrange view (Every time you seek, @init is called).
3. Disable Filther, see if good performance comes back.
4. (Re)enable Filther, see if bad performance comes back.
5. Save preset, unload filther, load preset, see if performance is restored.

Also, note the pdc_delay value for me.

Thanks!
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]

Last edited by sai'ke; 01-25-2019 at 06:32 AM.
sai'ke is offline   Reply With Quote
Old 01-26-2019, 07:47 AM   #130
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
...

If it happens again. Please try the following in this order:
1a. Toggle lookahead, see if it resolves the issue.
1b. Toggle FIR, see if this resolves it.
1c. Toggle the waveshaper off, see if this resolves it.
1d. Toggle all the dynamics processing off (modulation / dynamic), see if this resolves it.
1e. Toggle the filters off, see if this resolves it.
2. Seek to a certain time point in the arrange view (Every time you seek, @init is called).
3. Disable Filther, see if good performance comes back.
4. (Re)enable Filther, see if bad performance comes back.
5. Save preset, unload filther, load preset, see if performance is restored.

Also, note the pdc_delay value for me.

Thanks!
I will check the list next time. The weird thing is, it happens sometimes 3 or 4 times after the first peak. And sometimes it doesn't happen even once. But I will have a look at it.

- I noticed, that the hold time slider for MIDI2 is not working

- the graphical issue shows up rarely when working in the WS windows with extreme settings, but it doesn't end up with a freeze rather then just flashing the WS curve for a 1/10 second out of the boundaries and back. This also happens sometimes when changing other parameters (dyn,lfo)

- the random LFO doesn't do random if linked with midi trigger, it starts always with the same value. Could this be made to also be random as it else isn't possible to get it in sync?

- At the end I have to say that Filther is way more stable since the last releases. It seems that filters don't break out to easy. But this also could be because I mostly use bass like square waves as sound sources.

- The MIDI2 Mode is lovely. This adds so much motion and movement if you use velocity for expression. Great. Also the LFO-DYN separations is very clever made and is smooth like a Baby-Popo! I just realized how cool it is to use one filter with LFO and one with DYN.

- A final suggestion: Would it be possible to also make the output gain slider DYN sensitive? - Why? Sometimes you have filters which don't separate enough between notes as their lowest lowcut lets to much level thru. With a (amp envelope like) triggered output level this could be compensated a bit. I know the waveshaper is dynamic but this doesn't help if you use it for really hard carving.

Hm... I see, this is not easy to explain.

Anyway. Filther rocks! And next week I will upload some videos with examples and presets. There are some lovely ones present in those.
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-27-2019, 10:02 AM   #131
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Quote:
Originally Posted by Eliseat View Post
- I noticed, that the hold time slider for MIDI2 is not working
Fixed.

Quote:
Originally Posted by Eliseat View Post
- the graphical issue shows up rarely when working in the WS windows with extreme settings, but it doesn't end up with a freeze rather then just flashing the WS curve for a 1/10 second out of the boundaries and back. This also happens sometimes when changing other parameters (dyn,lfo)
Still?

Quote:
- the random LFO doesn't do random if linked with midi trigger, it starts always with the same value. Could this be made to also be random as it else isn't possible to get it in sync?
Fixed.

Quote:
- At the end I have to say that Filther is way more stable since the last releases. It seems that filters don't break out to easy. But this also could be because I mostly use bass like square waves as sound sources.
Yes, low frequency content tends to cause fewer problems.

Quote:
- The MIDI2 Mode is lovely. This adds so much motion and movement if you use velocity for expression. Great. Also the LFO-DYN separations is very clever made and is smooth like a Baby-Popo! I just realized how cool it is to use one filter with LFO and one with DYN.
Wahey!

Quote:
- A final suggestion: Would it be possible to also make the output gain slider DYN sensitive? - Why? Sometimes you have filters which don't separate enough between notes as their lowest lowcut lets to much level thru. With a (amp envelope like) triggered output level this could be compensated a bit. I know the waveshaper is dynamic but this doesn't help if you use it for really hard carving.
Done. This was harder than I thought though, since I did not have any sliders left. I managed to recycle pos8x though (since this one is always clamped to 1 anyway). Issue is though, that it may have been set differently in other presets. I had to jump to a few hoops to make sure that older presets didn't suddenly get crazy high gain surprises. It should be fine though *crosses fingers*. Let me know if there are issues.

Also made the gain section amenable to splitting DYN and LFO now.

I have noticed some serious performance with the "experimental" filter. I would not use this one until I investigate closer.

So... to summarize.
Changelog:
Code:
  + Feature: Add ability to automate post gain. Add ability to toggle DYN / LFO for them.
  + Bugfix: Bugfix LFO random.
  + Bugfix: Bugfix hold mechanism MIDI2.
  + Bugfix: Move spline memory to location where it is less likely to get clobbered if there is an initialization issue.
  + Bugfix: Properly initialize reso, midi and cutoff memory.
  + Improvement: Increase size tooltips
  + Improvement: Display cutoff (when available for that filter) on second filter as well.
  + Bugfix: Bugfix combobox
  + Improvement: Removed no_denorm to hopefully prevent CPU spikes.
  + Bugfix: Make sure that nodes actually have to be clicked to be selected.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 01-28-2019, 03:17 AM   #132
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post

Done. This was harder than I thought though, since I did not have any sliders left. I managed to recycle pos8x though (since this one is always clamped to 1 anyway). Issue is though, that it may have been set differently in other presets. I had to jump to a few hoops to make sure that older presets didn't suddenly get crazy high gain surprises. It should be fine though *crosses fingers*. Let me know if there are issues.

Also made the gain section amenable to splitting DYN and LFO now.
Sai'ke, this output manipulation is amazing now. And I have to say it was really worth your efforts as its now possible to use this as a counterpart: If the filter creates to much spikes, just use the opposite DYN at the output to even it out. That's just perfect now! It also allows now really weird chopped or faded amp values which make the range even more incredible! Great stuff! ٩(◕‿◕。)۶

The hint with the experimental filter is a good one. I will see if its involved in any kind.

To be more specific with my experiences: when I start Filther, everything seems fine for a while. The CPU is around 1-3 and it looks fluidly. Then I start switching the presets. Sometimes it works for 20-30 presets without any changes, sometimes it only needs five switches til the CPU usage rises continuously. The sound begins to stutter at 10-15% CPU and the GUI gets choppy and course.
I deactivated all other plugins. Did not help. I deactivated WS, clip, MIDI, FIR, Oversampling, 4x, 8x but nothing changed. Even deactivating and reactivating Filther does not help. I will make a video of that.

And for now I think this could be a problem on my system. Or is any one else experiencing this? I noticed some graphical issues on my system which occur on heavy usage. Maybe that leads into the problems with Filther.
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-28-2019, 04:33 AM   #133
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

I have no idea! ¯\(°_o)/¯

It seems that Filther is overflowing something which even doesn't recover after deactivating it. I noticed this after reloading Filther while playing the song - it does not change the overload. So I repeated it and this time it solved it for a short time but run crazy shortly after.

Though I'm pretty sure Filther starts something, I'm also skeptical if its not my system which makes the trouble afterwards. Maybe I also went a bit to far by creating weird presets. I don't know.

Last thing I could do would be a list of how much CPU the presets use with a fresh start and how much they use while Filther/Reaper goes crazy. This could show if maybe only the CPU hungry presets make the trouble.
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-28-2019, 05:11 AM   #134
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

@Eliseat- try recording your whole sessions with either licecap lcf or video (with,or without audio)-- this will record all steps,and you can edit vid to show 'worst moments' if needed..an edited video would be harder to retrace steps,but still usefull.
also,try both asio+wave drivers--- there may be noticable differences using different buffer settings..
having the performance meter open+pre fx spectrogram on rendered waveforms is also quite helpfull-

i currently have 3 js on monitoring duties:

gfxanalyzer
gfxspectrograph
gfxscope

+ if you actually find something you consider a 'bug' - remember your last moves..hit edit+full recompile reset > and try to duplicate it right there again so it's 'reproducable'.

Last edited by Bri1; 01-28-2019 at 05:20 AM. Reason: 1xtra
Bri1 is offline   Reply With Quote
Old 01-28-2019, 05:42 AM   #135
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Has it happened on the latest version too? I did change something in the midi handling on the latest versions (1.69 and up), which could've caused issues in very rare cases in the way it was before.

Reloading it does not fix it? By reloading you really mean deleting the old filther and inserting a new one? Recompiling it? Or muting and unmuting?

Are there any plugs that have a memory (delays, filters, phasers, reverbs) downstream of Filther? I'm curious whether it spits out a denormal at some point which then grinds down whatever is south of Filther. Denormals are floating point values that are either extremely close to 0 or some other invalid or not representable value. Unless it's caused by a memory bug, I'd expect such a situation to be reproducible however. Next time, try muting the effects after filther as well (if any).

Again, if you can get me any sort of procedure that can trigger this (with presets or not), it would be incredibly helpful. I've been trying to cycle through my own presets on my end, but I can't seem to trigger this.

The only thing that seems to choke up the CPU at certain settings is this "Experimental" filter which I am thinking of removing.

Hmmm, actually, come to think of it... I should probably initialize the filter memory to very small values rather than zero to avoid de-normals there. I wouldn't expect that to make much of a difference in this case though.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 01-28-2019, 09:19 AM   #136
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
Has it happened on the latest version too? I did change something in the midi handling on the latest versions (1.69 and up), which could've caused issues in very rare cases in the way it was before.

Reloading it does not fix it? By reloading you really mean deleting the old filther and inserting a new one? Recompiling it? Or muting and unmuting?

Are there any plugs that have a memory (delays, filters, phasers, reverbs) downstream of Filther? I'm curious whether it spits out a denormal at some point which then grinds down whatever is south of Filther. Denormals are floating point values that are either extremely close to 0 or some other invalid or not representable value. Unless it's caused by a memory bug, I'd expect such a situation to be reproducible however. Next time, try muting the effects after filther as well (if any).

Again, if you can get me any sort of procedure that can trigger this (with presets or not), it would be incredibly helpful. I've been trying to cycle through my own presets on my end, but I can't seem to trigger this.

The only thing that seems to choke up the CPU at certain settings is this "Experimental" filter which I am thinking of removing.

Hmmm, actually, come to think of it... I should probably initialize the filter memory to very small values rather than zero to avoid de-normals there. I wouldn't expect that to make much of a difference in this case though.
I wouldn't delete the experimental. Its really good for sick stuff.

But finally I have good news for you:

I had a short session and realized that while switching the presets sometimes Filther begins to overload the CPU. But if I turn off and on some of the parameters like FIR, WS or something else it gets back to the normal state even if I have everything set back to the same as the saved preset.
For example: I choose the next saved preset and it starts with 6-7% CPU and quickly rises to about 15% CPU. This is the range where it begins to crackle etc. If I let it play and turn some parameters off, the CPU usage decreases to about 4%. And if that happens I can turn on all the parameters back again and this time it stays at 4%.

So that means that probably only one of the parameters (FIR?) misbehaves while switching from one preset to the next (does it fill something up?) while it else works without problems. This is the reason why I can create those presets without any sign of overloading. I wouldn't save a preset if it consumes more than 5% CPU. And this made me wonder why it later showed up with overload while switching thru the presets.

Do you have an idea?

Hm, I just realized that my video recording has no sound. What a bummer. (。╯︵╰。)
But at least you can see whats going on.
https://mega.nz/#!OrRF0KDR!AW1C9E5I1...2r2Np6bG1aGvzU
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-28-2019, 12:10 PM   #137
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Another important hint: It also happens if no audio computing takes place. I let the MIDI sequence play while disabling Synth1 and it also went thru the sealing. So the fact that I only created MIDI presets could have an effect.

Wouldn't it be the best to post my presets so you could try it by yourself? Maybe you get the same results and are able to figure out whats happening?
But I have to confess I have no idea how to export a bunch of presets without exporting also presets which should not be exported. (Like your default presets.) ¯\_(ツ)_/¯
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-28-2019, 03:02 PM   #138
foxAsteria
Human being with feelings
 
foxAsteria's Avatar
 
Join Date: Dec 2009
Location: Oblivion
Posts: 10,255
Default

Bit more praise for Jedi Coder Sai'ke... I haven't had this much fun playing with filters since Buzz!

I notice that now there is much less volume change when switching between types and I seem to no longer be getting any big spikes (time will tell); how did you achieve that? Some kind of limiting? Sry, I haven't read the whole thread...

The gain linking is awesome! I second the request for a randomization button (just for certain parameters, probably, but esp filter combo).

When using RMS mode, it doesn't show the time in seconds anymore. Is this intended?

Requests: 1. Could the Master and Dynamics sections be next to each other? I'm going between threshold and gain a lot, but the filters are in between them. 2. The filter descriptions take up a lot of space. Couldn't that all be done with tooltips? The whole UI could use about 2x magnification on all the controls and text, but the window is already large. 3. I think another really handy thing would be if you fixed the threshold line to the center of the the RMS / Level view and instead zoomed that window when moving the thresh slider. Otherwise, we have to keep zooming that window manually after changing the thresh, whereas this would do it automatically. In that case it would also make a lot of sense to have the thresh control physically in line with the thresh line in the window. And maybe it should change it's name to "maximum" also when in RMS mode?
4. Bigger request would be to allow the two filters to be used for L/R/Mid/Side channels in parallel. I'll just start holding my breath now...

P.S. I'd be willing to apply some time and design skills toward a more user friendly and attractive GUI if you're interested... I don't know how that works, but a nice, scale-able, vector UI would really do Filther some justice. It's able to do so much of the stuff I could never find other filters to do.
__________________
foxyyymusic

Last edited by foxAsteria; 01-28-2019 at 04:21 PM.
foxAsteria is online now   Reply With Quote
Old 01-29-2019, 03:15 AM   #139
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

@Eliseat;
For me, it's okay if you export too many. I'll just cycle until it hits something. I'll make a backup of whatever I have at the moment. You can send 'em over PM if you're afraid of someone stealing them. I won't distribute them.

Alternatively, you could backup your presets (afaik they're in %appdata%\REAPER\Presets) first. It's the ini file with filther's name. Then remove the ones you don't want to share. Then export the presets you have left to an RPL file and send me only those.

@FoxAsteria;
Yeah, random button is on the to-do list.
Parallel filters is also on the to-do list.

RMS mode thing is a bug; added to to-do list.

I could flip dynamics and the filters, yes. Assuming everyone else is fine with this. Would it make sense to move LFO too then? They're kind of similar.

I'm not a fan of centering the thresh line by default. I very often look at a larger or smaller scale than the threshold to see what's happening to the other curves.

I appreciate the offer, but a big UI redesign isn't in the books, sorry (unless it can be done with the existing widgets). UI stuff in JSFX is kind of painful. Much more painful than in Lua. I'm currently swamped at work, and planning a move and a new job soon, where I'll be doing a lot of learning new things (not a lot of time for personal projects), so I'm trying to get Filther as stable as possible in the remaining time.

If you know someone else who feels like modding the codebase for a new UI; that's absolutely fine btw.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 01-29-2019, 11:56 AM   #140
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Sai'ke, do you really think I would have a problem if someone uses my presets? Of course not. You share such a great tool - so its the least I can do. I already planned to post a video with sound examples and to share the corresponding presets here in this thread. Problem is, my kids made a lot of trouble in the last days. I had no time to adjust the levels or to hear the presets once over loud speakers before posting. But soon!

So here is the attachment Eli_Filther_data.zip. You will find my rpl preset library of Filther. Of course you also will find a lot of your own presets too. (I hope you have an idea of how to clean this up a bit.)
All of my presets are MIDI presets for bass range. In the zip file is also a picture of a synth1 preset which I used to create my Filther presets. This could be important to have, especially if it comes to sound level and frequency range. (By the way, it was not my intention to only make a screenshot, but Reaper did not export the presets for synth1.)

If you have set up synth1 and Filther you just should choose the last preset in my list, feed a MIDI bass sequence and start switching the presets upwards. The first troublemaker is called "DengDeng MIDI bass" which tends to run hot and to start the instability. A few presets later I get the known trouble with the over boarding CPU usage.

To be honest: I'm pretty curious if this also happens to you or if it is something specific on my system. And it would be cool if this could help to find the leak.

Okay. So lets see what it does.

Greetings
Eli

By the way: I have a question related MIDI triggering. I used Synth1 and ReaSynth and they worked pretty well without any problems. But today I wanted to use a Full Bucket synth. And it doesn't route/send the MIDI "note on" commands to Filther. How can this be solved? Is there a way to route the "note on" commands behind the synth so I could use it with Filther too? Many thanks.
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-29-2019, 01:50 PM   #141
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Bri1 is offline   Reply With Quote
Old 01-29-2019, 03:58 PM   #142
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

I have not had time to check it yet. I was still in the middle of a relatively big new feature that I was working on for the past few days. Filters can now optionally follow the pitch of the last MIDI key you've played.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 01-29-2019, 04:14 PM   #143
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
I have not had time to check it yet. I was still in the middle of a relatively big new feature that I was working on for the past few days. Filters can now optionally follow the pitch of the last MIDI key you've played.
Oh, that sounds promising. ٩(◕‿◕)۶

But of course take your time as it fits in your planning. Don't feel under pressure because of my posting. There is no deadline. (At least I don't know. )

__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-30-2019, 05:59 AM   #144
TonE
Human being with feelings
 
Join Date: Feb 2009
Location: Reaper HAS send control via midi !!!
Posts: 4,031
Default

Hackey filther is cool..

Last edited by TonE; 01-30-2019 at 09:05 PM.
TonE is offline   Reply With Quote
Old 01-30-2019, 07:03 AM   #145
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

TonE; Could you please post your unrelated plugin requests somewhere else?

The plugin looks cool, but is not similar to the aim of Filther at all. The thread already moves quite quickly and I'd prefer to not have this thread derail into discussion about an unrelated plugin.

ReaEq is probably closest, but not a dynamics but a frequency processor. But as I said, it's probably best to open a separate thread for a request such as this.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 01-30-2019, 03:29 PM   #146
X-Raym
Human being with feelings
 
X-Raym's Avatar
 
Join Date: Apr 2013
Location: France
Posts: 9,875
Default

Hi,


This script is really pushing the limit of JSFX ^^



Just wanted to warn you that your changelog has a wrong syntax for reapack,


You should add prefix for version and date else it will just write everything for each releases. CUrrently, changelog for this JSFX is about 564 lines long in Reapack sync update notice :S



For eg, rather than





Code:
  + Optional keyfollow                         + Minor bugfix for very fast attacks in the dynamics                         + Add ability to automate post gain. Add ability to toggle DYN / LFO for them.                         + Bugfix LFO random.                         + Bugfix hold mechanism MIDI2.                         + Move spline memory to location where it is less likely to get clobbered if there is an initialization issue.                         + Properly initialize reso, midi and cutoff memory.                         + Increase size tooltips                         + Display cutoff (when available for that filter) on second filter as well.                         + Bugfix combobox                         + Removed no_denorm to hopefully prevent CPU spikes.                         + Make sure that nodes actually have to be clicked to be selected.                         + Added ability to activate dynamics and lfo for specific filters separately.changelog:

Try


Code:
changelog:
v1.72(2018-09-12)

  + Optional keyfollow                         + Minor bugfix for very fast attacks in the dynamics                         + Add ability to automate post gain. Add ability to toggle DYN / LFO for them.                         + Bugfix LFO random.
v1.72(2018-09-12)                         + Bugfix hold mechanism MIDI2.                         + Move spline memory to location where it is less likely to get clobbered if there is an initialization issue.                         + Properly initialize reso, midi and cutoff memory.                         + Increase size tooltips                         + Display cutoff (when available for that filter) on second filter as well.                         + Bugfix combobox                         + Removed no_denorm to hopefully prevent CPU spikes.                         + Make sure that nodes actually have to be clicked to be selected.                         + Added ability to activate dynamics and lfo for specific filters separately.

Thanks for taking a look :P

EDIT/ and of course, the post editor delete my code formatting.
X-Raym is offline   Reply With Quote
Old 01-30-2019, 03:37 PM   #147
cfillion
Human being with feelings
 
cfillion's Avatar
 
Join Date: May 2015
Location: Québec, Canada
Posts: 4,937
Default

The changelog tag should only contain the changelog of the current release.
cfillion is offline   Reply With Quote
Old 01-30-2019, 04:30 PM   #148
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Ouch!

Ah okay, thanks for the heads up. I will only log the most recent changes starting now then. Any way to reset the old ones to not have duplicates? Should I just kill the index.xml and reindex it once with a changelog of everything, and then continue only adding new ones under the tag from thereon out? Or is there an easier way to do it that preserves the history?

Eliseat, those were some great presets, thank you!

I didn't notice any stuttering, but I did notice that the Deng Deng one was the most expensive one on the list. On my computer:

Deng Deng MIDI Bass takes ~11.4%
DeepDust ~10.7%

Those were the most expensive. So that's quite a lot really. Could really be that it's just too heavy. It probably comes from running the MS-20 so often which is pretty expensive (M1/S2 runs the filter twice for the side channel). I could make a slightly cheaper version of the MS-20 using a cheaper approximation to the tanh (the most expensive part).
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 01-30-2019, 04:36 PM   #149
cfillion
Human being with feelings
 
cfillion's Avatar
 
Join Date: May 2015
Location: Québec, Canada
Posts: 4,937
Default

Quote:
Originally Posted by sai'ke View Post
Any way to reset the old ones to not have duplicates? Should I just kill the index.xml and reindex it once with a changelog of everything, and then continue only adding new ones under the tag from thereon out? Or is there an easier way to do it that preserves the history?
Easiest way to fix the previous changelogs while keeping the update history would be to edit them in the index.xml manually.
cfillion is offline   Reply With Quote
Old 01-31-2019, 01:15 AM   #150
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
Ouch!

Ah okay, thanks for the heads up. I will only log the most recent changes starting now then. Any way to reset the old ones to not have duplicates? Should I just kill the index.xml and reindex it once with a changelog of everything, and then continue only adding new ones under the tag from thereon out? Or is there an easier way to do it that preserves the history?

Eliseat, those were some great presets, thank you!

I didn't notice any stuttering, but I did notice that the Deng Deng one was the most expensive one on the list. On my computer:

Deng Deng MIDI Bass takes ~11.4%
DeepDust ~10.7%

Those were the most expensive. So that's quite a lot really. Could really be that it's just too heavy. It probably comes from running the MS-20 so often which is pretty expensive (M1/S2 runs the filter twice for the side channel). I could make a slightly cheaper version of the MS-20 using a cheaper approximation to the tanh (the most expensive part).
Yes, I know some of them are expensive but that wasn't the problem in the first place. What occurs is that after going up in CPU consumption once it changes the behavior of the less expensive presets. I could use a preset with 1 or 2 %, changing it to "DengDeng", and after going back to the previous preset its not 1-2% anymore but 5-6%.

I recorded a whole session which showed that. But for some reason it did not record the audio. (I still don't know why.) I will repeat this as soon as I get free time for that. Everyone around me is ill at the moment. So its a bit (way) more stress right now.

And I also have to test the new key dependency. That is really great stuff. By the way. Wouldn't it be more consistent to change the MIDI2 button to a fourth mini button next to the filter so that there is VEL, KEY, DYN and LFO? Have no idea how much work this would take or if its really complicated but I see no big difference between VEL and KEY cutoff dependency.

Greetings
Eli
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-31-2019, 02:30 AM   #151
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Yeah, that occurred to me too and that was how I initially wanted to implement it, but the reason for the way it is implemented now is that vel overrides the dynamics variable with velocity and is incompatible with using the RMS option. Key doesn't override the dynamics variable (it would be nice to have both key and dynamics), but is additive to whatever pitch is being played last. Hence it is compatible with using the RMS following option.

That's interesting. I will look at it again tonight, see if I can see this behavior happen. I'll also have a close look at the video. I mean, usually it does take a while for the CPU meter to adjust to the new preset. This is normal I think, and is because it is probably some moving average, but if it really stays, then that might indicate an issue in filther.

I will report back!

Hope everyone gets well soon.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]

Last edited by sai'ke; 01-31-2019 at 03:00 AM.
sai'ke is offline   Reply With Quote
Old 01-31-2019, 04:17 AM   #152
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
And I also have to test the new key dependency.

lol-- you do not have to
as much i try to love all things reaper and what all other people come up with using it- i still do not think paying customers or users should be 'testing' anything at all- just using what they paid for!

in the case of loving creators like sai'ke- a single user giving freely to the public at much expense to themselves- this is fine+really helpfull user feedbacks to this dev.


really though-all devs (imo) should be rectifying their own mistakings- not relying on users for that...i think this 'looks' bad for devs.. releasing 'products' that are not quite- ready!.
it does not mean any are incompetent or inadequate-quite the reverse in most cases--->>> some are just needing help: ! =)

1love.great stuff sai'ke!! champ.
Bri1 is offline   Reply With Quote
Old 01-31-2019, 05:36 AM   #153
Eliseat
Human being with feelings
 
Eliseat's Avatar
 
Join Date: Mar 2018
Location: Cologne
Posts: 1,362
Default

Quote:
Originally Posted by sai'ke View Post
Yeah, that occurred to me too and that was how I initially wanted to implement it, but the reason for the way it is implemented now is that vel overrides the dynamics variable with velocity and is incompatible with using the RMS option. Key doesn't override the dynamics variable (it would be nice to have both key and dynamics), but is additive to whatever pitch is being played last. Hence it is compatible with using the RMS following option.

That's interesting. I will look at it again tonight, see if I can see this behavior happen. I'll also have a close look at the video. I mean, usually it does take a while for the CPU meter to adjust to the new preset. This is normal I think, and is because it is probably some moving average, but if it really stays, then that might indicate an issue in filther.

I will report back!

Hope everyone gets well soon.
Thank you.

Hm, why does KEY not override dynamics? Does it not change the cutoff key dependently plus the dynamical cutoff changes in MIDI or RMS mode?
Ah ... you mean you did this to avoid RMS and VEL together. But this could also be assured by greying out the opposite button. If someone enables VEL, the RMS gets grayed and in reverse ... but - wait a moment! Does this even make sense? Hm, ... I guess this could work. I don't know. (⊙_⊙)

And Bri1: Two weeks ago I made an extension the very first time in my life. It was amazing because since the early 90ies I was confronted several times with programming in school - but I hated it so much then. In my opinion this was only for boys and super boring. But I had to learn IF, WHILE, FOR, DO, THEN, FUNCTION etc (in Basic) if I didn't want to get a bad grade. So I new about it, but never ever thought I would use it anytime for anything.
Then I wrote this VLC extension and it was like diving into a new world. I lived inside this little thing. (I guess Sai'ke does it too, but inside Filther.) At a certain point I was sure this extensions finally runs smoothly. I tested with all purposes and it never showed any problems. So I posted it - and at the very same moment someone answered: "IT does not work!"
What? Yes! And the first time in my life I realized that it is probably impossible to oversee everything for a programmer. In my case it was a simple space character in the user's path which I never had noticed as an error if he didn't had pointed at it. So I really know how important and helpful it is, if users test stuff out.

And as I already mentioned. This is the least I can do to help Sai'ke with his thingy. (Please don't get this wrong )
__________________
☆.。.:*・°☆.。.:*・°☆.。.:*・°☆REAPER//✿◔‿◔)°☆.。.:*・°☆.。.:*・°☆
Eliseat is offline   Reply With Quote
Old 01-31-2019, 06:15 AM   #154
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
I tested with all purposes and it never showed any problems. So I posted it - and at the very same moment someone answered: "IT does not work!"




So I really know how important and helpful it is, if users test stuff out.

And as I already mentioned. This is the least I can do to help Sai'ke with his thingy. (Please don't get this wrong )

heh- yes- it's a very fine line for devs- testing stages upto the releasing stage.
i'm sure some coders are very satisfied with their results -so they share--- but-- are they really doing full testings before doing so>?
here's an extreme eg: car manufacturers- people would be crashing a whole lot more if these guys were not so vigorous in testing stages!! think about it.

seems not so because: i have only spent about 2 years checking this forum,and i see a lot of 'issues' - like literally daily (totally unrelated to filther) but also related-- as like sai'ke himself is even finding he has to stretch his imaginations and practical code to 'workaround' a limitation he has found while creating....this is not really sai'ke faulting -but cockos not realizing these limits would be met- easily,by home developers <><>?
i actually thought reaper had 'no limits'- but i see it all very differently now- it does- and many other problems since day 1 of using it.. (rant!)


it reflects back poorly on user develops,when the programme they are using to create- is failing them-- and failing hard.. that's being honest.
cockos have spent literally years developing this programme- and look at it!! yep it's wonderfull- but actually broken 't fk for a lot of 'users' - or,completely un_logical for both new and advancing users..
sort it out ffs.


no offense intended!! to any creator!! > but hopefully, points are 'considered'.
Bri1 is offline   Reply With Quote
Old 01-31-2019, 08:06 AM   #155
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Yeah, greying out RMS could be implemented, but I thought that this way would be more intuitive, since all the buttons in the dynamics window affect what dynamic is listening for (RMS or MIDI), while the pitch following is completely unrelated to dynamics.

Greying out RMS seems counter-intuitive to me. It seems less likely to me that one would make the mental connection with the yellow curve then. But I don't know, I'm not a UX expert
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]
sai'ke is offline   Reply With Quote
Old 02-01-2019, 05:41 PM   #156
sai'ke
Human being with feelings
 
sai'ke's Avatar
 
Join Date: Aug 2009
Location: NL
Posts: 1,453
Default

Parallel filter is an option now (only basic for now, no mid/side business (yet))

Also cleaned up the changelog.

Also added a randomize button (SHIFT + R), but I'm not a huge fan. Most of the stuff obtained with randomize sounds downright terrible.
__________________
[Tracker Plugin: Thread|Github|Reapack] | [Routing Plugin: Thread|Reapack] | [More JSFX: Thread|Descriptions|Reapack]

Last edited by sai'ke; 02-01-2019 at 06:50 PM.
sai'ke is offline   Reply With Quote
Old 02-02-2019, 03:00 AM   #157
sinkmusic
Human being with feelings
 
sinkmusic's Avatar
 
Join Date: Feb 2006
Location: decepticon mothership in a hidden place inside a mountain
Posts: 3,754
Default

Thank you !
sinkmusic is offline   Reply With Quote
Old 02-02-2019, 04:39 AM   #158
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default


would it anger the godz by requesting another lfo?
the godz would please the citizens greatly with this gift !
it is unique.
Bri1 is offline   Reply With Quote
Old 02-02-2019, 08:58 AM   #159
kneipp123
Human being with feelings
 
Join Date: Apr 2017
Posts: 96
Default

hi there, very cool plugin! randomize not working for me. how do i use it? is it because i have something else mapped to shift R?
kneipp123 is offline   Reply With Quote
Old 02-02-2019, 08:59 AM   #160
kneipp123
Human being with feelings
 
Join Date: Apr 2017
Posts: 96
Default

also the gui is cut off in the right corner for me, any chance of a resize? this happens with a lot of plugins for me, anyone know an easy fix?
kneipp123 is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 06:58 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.