|
|
|
11-17-2018, 07:20 AM
|
#41
|
Human being with feelings
Join Date: Apr 2011
Location: Germany
Posts: 1,186
|
Adding the peak bands with a double click works very well for me. But I would have three wishes, too:
1. adding lowshelfs with a double click below 100 Hz and highshelfs above 5 kHz. Is implemented, but does not work!? I found out. Low shelf with double click at 20 - 30 Hz and high shelf at 10 - 15 kHz.
2. it would also be nice if I could change the slope of the low and highcuts with the mouse wheel and by dragging up and down at the touch point.
3 The analyzer is a bit too slow for my taste. It would be nice if I could adjust it a bit faster.
But even without my wishes this is a first-class EQ for me. Thanks a lot!
Last edited by mawi; 11-17-2018 at 06:59 PM.
Reason: I read the manual. :)
|
|
|
11-17-2018, 07:23 AM
|
#42
|
Human being with feelings
Join Date: Feb 2017
Posts: 305
|
It's super green!
Thank you much for sharing!
|
|
|
11-17-2018, 07:45 AM
|
#43
|
Human being with feelings
Join Date: May 2014
Location: Germany
Posts: 643
|
Fantastic stuff, JSFX at its best!!!
Here you can find a small extension based on the original post (still with 2x bug, sorry):
I added an offset option, just to align spectrum with EQ curve.
This makes the range options obsolete (but still there)
Max spectrum is also enabled, still need to see how to reset it:-)
Have fun with it.
Last edited by TBProAudio; 11-17-2018 at 11:26 PM.
|
|
|
11-17-2018, 02:05 PM
|
#44
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,746
|
Quote:
Originally Posted by nitsuj
SSE optimized branch? Nice. Let me know if it's a bug in my code that you've spotted. I'd like to get bugs ironed out before I ReaPack it.
BTW, great job on JSFX. Making it fast enough to do this kind of processing is really cool.
|
Glad you like JSFX, thanks for making such great scripts!
The bug I saw is almost certainly on our side (if it works on the release but fails in my branch, then the problem is in my branch). The SSE implementation isn't anything fancy with SIMD, just about getting rid of old x87 code (which is especially important on x86_64 where the calling convention dictates xmm0 as return values, etc).
Also: I did improve the compile speed of ReEQ in the latest +dev pre-releases (on my i7 RMBP it was taking 1.2s to compile, which I reduced down to 120ms... which is still quite a bit longer than I'd like it to be.)
|
|
|
11-18-2018, 12:14 AM
|
#45
|
Human being with feelings
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,787
|
Quote:
Originally Posted by nitsuj
- I've used showmenu() for the menus. Technically it's not documented for JSFX so I don't know how well it works on other platforms.
|
Could you elaborate ? Does it work ? What add-on does provide that function ? How did you find it ? How did you get the specs ?
-Michael
|
|
|
11-18-2018, 03:51 AM
|
#46
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by mschnell
Could you elaborate ? Does it work ? What add-on does provide that function ? How did you find it ? How did you get the specs ?
-Michael
|
Yes, it works. And judging by the feedback here it's good for multi-platform. I think it's part of the interface that's exposed for Reaper scripting in general. It's documented for lua but not for eel2.
I found it on a forum post:
https://forum.cockos.com/showthread.php?t=164227
|
|
|
11-18-2018, 03:53 AM
|
#47
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by Justin
Also: I did improve the compile speed of ReEQ in the latest +dev pre-releases (on my i7 RMBP it was taking 1.2s to compile, which I reduced down to 120ms... which is still quite a bit longer than I'd like it to be.)
|
That's a seriously good improvement. 10 x ? Yes please. Looking forward to getting this version.
|
|
|
11-18-2018, 04:03 AM
|
#48
|
Human being with feelings
Join Date: Nov 2007
Location: France
Posts: 919
|
Nice plugin, I also notice that
- If I double the click the mouse in the lower part where the boxes are, bands are created.
|
|
|
11-18-2018, 08:25 AM
|
#49
|
Human being with feelings
Join Date: Jun 2007
Location: Terra incognita
Posts: 7,670
|
Thanks for a nice EQ, nitsuj. Menus and all seem to work fine here on Window 10 PC.
Quote:
Originally Posted by nitsuj
Best thing is to use Reaper's performance view to see how much CPU the DSP is taking. I'd be happy to see feedback to see how it's performing on different machines.
|
I made a test project with 10 tracks. One ReEQ instance on each track, every instance had low and hi cut plus one peak. No fx GUIs open. Windows 10, i7 4770k @ 3.5GHz.
Interestingly the CPU went up when the project was stopped and came back down when started again.
I have saved the project for a possible future reference.
edit: this test run with 5.962+dev1117/x64, I can compare later with regular 5.962.
Last edited by xpander; 11-18-2018 at 04:32 PM.
|
|
|
11-18-2018, 12:27 PM
|
#50
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by xpander
Interestingly the CPU went up when the project was stopped and came back down when started again.
|
Thanks for the performance feedback. That's really odd though. I wouldn't expect it to take any processing time if the project is stopped. I'll check myself.
|
|
|
11-18-2018, 04:10 PM
|
#51
|
Human being with feelings
Join Date: Jan 2014
Posts: 27
|
Thanks for the scripts! They are truly awesome!
My main critique atm, would be that it's a CPU hog. There also seem to be a weird bug where it actually uses less CPU when playing (!?). I haven't noticed this behaviour with other plugins. Though it might be my "theme" changes that is causing it, so take that with a grain of salt
If anyone one wants to check out the "theme'ing" options, the files can be found here:
https://www.dropbox.com/s/n9bm05q843...J_TvM.zip?dl=0
Last edited by tvm79; 11-18-2018 at 04:22 PM.
|
|
|
11-18-2018, 04:37 PM
|
#52
|
Human being with feelings
Join Date: Jun 2007
Location: Terra incognita
Posts: 7,670
|
Quote:
Originally Posted by nitsuj
Thanks for the performance feedback. That's really odd though. I wouldn't expect it to take any processing time if the project is stopped. I'll check myself.
|
Sorry, I forgot to mention I did the test run with Reaper v5.962+dev1117. Regular 5.962 will give a bit different numbers, but the similar behavior. Something else I noticed with this test. If I open Reaper fresh but don't start the project immediately, sometimes the base FX CPU is about 9%, sometimes around 20%. With the project running it then goes to 20+ and finally when playback is stopped, it starts to rise towards 40%?
Just to let you know, in case my system is for some reason not stable enough currently to run tests like this. Wouldn't want you to chase ghosts which could be on my end...
|
|
|
11-18-2018, 05:57 PM
|
#53
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,746
|
Quote:
Originally Posted by nitsuj
That's a seriously good improvement. 10 x ? Yes please. Looking forward to getting this version.
|
Down to 44ms in the latest prerelease (see the prerelease forum for more info)
|
|
|
11-18-2018, 06:00 PM
|
#54
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,746
|
Quote:
Originally Posted by xpander
Sorry, I forgot to mention I did the test run with Reaper v5.962+dev1117. Regular 5.962 will give a bit different numbers, but the similar behavior. Something else I noticed with this test. If I open Reaper fresh but don't start the project immediately, sometimes the base FX CPU is about 9%, sometimes around 20%. With the project running it then goes to 20+ and finally when playback is stopped, it starts to rise towards 40%?
Just to let you know, in case my system is for some reason not stable enough currently to run tests like this. Wouldn't want you to chase ghosts which could be on my end...
|
The increase in CPU when stopped could be a denormal issue. JSFX is meant to automatically deal with that but maybe it’s not doing a great job of it. Or maybe it’s pdc related — does that same thing happen when the oversampling is off?
|
|
|
11-19-2018, 12:37 AM
|
#55
|
Human being with feelings
Join Date: May 2014
Location: Germany
Posts: 643
|
Yes, it was the denorm thing. Just find the small update here (still wo 2x OS bug).
Update 1.2: denorm fixed, click into graphics resets max values
|
|
|
11-19-2018, 01:46 AM
|
#56
|
Human being with feelings
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,787
|
Quote:
Originally Posted by nitsuj
Yes, it works. And judging by the feedback here it's good for multi-platform. I think it's part of the interface that's exposed for Reaper scripting in general. It's documented for lua but not for eel2.
I found it on a forum post:
https://forum.cockos.com/showthread.php?t=164227
|
Wonderful !!
-Michael
|
|
|
11-19-2018, 04:02 AM
|
#57
|
Human being with feelings
Join Date: Jul 2009
Posts: 1,231
|
Been running this for a few mixes now. It's very good! I have a small quibble that could perhaps be resolved:
- Could we have an option to not have the EQ range influence the display range of the frequency analyzer? The signal has to be quite hot to make use of the analyzer when working at lower range-settinga, like 6-12dB.
|
|
|
11-19-2018, 06:43 AM
|
#58
|
Human being with feelings
Join Date: Dec 2015
Location: holland
Posts: 29
|
feedback
hallo 1s thnks again for this lovely eq there for some feedback
i use a dell optiplex 4gig ram on windows7
64bit reaperdaw
project 24 track
eq used on masterbus
1.. No tracks playing and eq floating on master bus... un muted all tracks reads 8,5% cpu
2.. no tracks playing and all tracks muted...it reads 36,5 % cpu use
3.. all tracks playing and unmuted eq floating on master reads 10,5 % cpu use
4.. all tracks playing but muted eq on master bus floating reads 36.5% cpu use
5..all tracks playing and unmuted with the eq on the masterbus reads 10,5% cpu use..
find that strange you woud expect the other way around..
hopfully helpfull info
thnks 1s again and greetings from holland
|
|
|
11-19-2018, 09:16 AM
|
#59
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by TBProAudio
Yes, it was the denorm thing. Just find the small update here (still wo 2x OS bug).
Update 1.2: denorm fixed, click into graphics resets max values
|
What did you do to fix the denorm issues?
BTW, here's the code to fix the oversampling 2x. It's in the top original zip file now.
Code:
/*
* Set oversampling rate including PDC
*/
function set_oversample(os) (
// We'll only switch to oversampling if the sample rate
// is low enough to benefit.
srate < 88200 && os ? (
DO_OVERSAMPLE = 1;
SAMPLE_RATE = srate * 2;
pdc_delay = get_FIR_pdc();
pdc_bot_ch = 0;
pdc_top_ch = 2;
) : (
DO_OVERSAMPLE = 0;
SAMPLE_RATE = srate;
pdc_delay = 0;
pdc_bot_ch = 0;
pdc_top_ch = 2;
);
);
|
|
|
11-19-2018, 09:23 AM
|
#60
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by ramses
- Could we have an option to not have the EQ range influence the display range of the frequency analyzer? The signal has to be quite hot to make use of the analyzer when working at lower range-settinga, like 6-12dB.
|
The range for the analyzer and the range for eq are unrelated code-wise. If you play sound then alter the eq range you shouldn't see the analyzer display alter. Maybe I didn't understand correctly?
|
|
|
11-19-2018, 11:03 AM
|
#61
|
Human being with feelings
Join Date: May 2014
Location: Germany
Posts: 643
|
Quote:
Originally Posted by nitsuj
What did you do to fix the denorm issues?
|
Here it is: svf_filter.jsfx-inc
You may add this to zdf_svf1 as well.
Code:
function zdf_svf0(v0)
instance(rbj, ic1eq0, ic2eq0, a1, a2, a3, m0, m1, m2, nlp, lp1, lp2, lp3, lp4, lp5, lp6, lp7, onepole, z1_0, og, cutoff, op0, op1)
local(v1, v2, v3)
(
v0+=0.00000001; // DeNorm fix
nlp > 0 ? (
I will later send the full package (including 2x OS fix)
|
|
|
11-19-2018, 11:06 AM
|
#62
|
Human being with feelings
Join Date: May 2014
Location: Germany
Posts: 643
|
New update 1.3: fix for 2x OS bug
|
|
|
11-19-2018, 11:08 AM
|
#63
|
Human being with feelings
Join Date: May 2014
Location: Germany
Posts: 643
|
Quote:
Originally Posted by nitsuj
The range for the analyzer and the range for eq are unrelated code-wise. If you play sound then alter the eq range you shouldn't see the analyzer display alter. Maybe I didn't understand correctly?
|
TB 1.3 version connects spectrum and EQ range and makes menu entries range obsolete.
The right button value "offset" aligns spectrum and EQ display.
|
|
|
11-19-2018, 12:33 PM
|
#64
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by TBProAudio
TB 1.3 version connects spectrum and EQ range and makes menu entries range obsolete.
The right button value "offset" aligns spectrum and EQ display.
|
Feels like I ought to get ReEQ and ReSpectrum into Github and ReaPack soon!
As a side, the reason I had the spectrum and EQ range separate is so that you can see what the filters do at full spectrum range - all the way down to -200db. It's very interesting to see what other plugins do down there in terms of introducing noise and harmonics. I'm a sucker for analysing plugins.
Just trying to understand the denorm fix. Does it need to be done for each filter or could it be done at the top of the @sample section to spl0 and spl1?
Last edited by nitsuj; 11-19-2018 at 12:34 PM.
Reason: Another question
|
|
|
11-19-2018, 12:39 PM
|
#65
|
Human being with feelings
Join Date: Jul 2009
Posts: 1,231
|
Quote:
Originally Posted by TBProAudio
TB 1.3 version connects spectrum and EQ range and makes menu entries range obsolete.
The right button value "offset" aligns spectrum and EQ display.
|
Ah, that explains it, I was using the TB 1.3 version. And I stumbled on the righmost button right after I wrote the question. Alright, so this might work, thanks for clearing it up.
|
|
|
11-19-2018, 01:30 PM
|
#66
|
Human being with feelings
Join Date: Oct 2017
Location: Larisa, Greece
Posts: 3,827
|
WOW!Thanks a lot for sharing those great scripts!
|
|
|
11-19-2018, 02:01 PM
|
#67
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,746
|
As far as the denormal issue goes -- I assume ext_no_denormal is being set to avoid noise being added to the analyzed input?
Anyway -- the addition of 0.0000001 or whatever helps, but ideally the JSFX code generator should be taking care of this. I'm investigating why it is not. Edit: made some slight tweaks the the denormal-prevention code, which fixes (coming in 5.97 or whatever).
Something I noticed in svf_filter:
Code:
nlp > 0 ? (
v3 = v0 - ic2eq0;
v1 = a1 * ic1eq0 + a2 * v3;
v2 = ic2eq0 + a2 * ic1eq0 + a3 * v3;
ic1eq0 = 2 * v1 - ic1eq0;
ic2eq0 = 2 * v2 - ic2eq0;
v0 = m0 * v0 + m1 * v1 + m2 * v2;
);
nlp > 2 ? (
v3 = v0 - lp1.ic2eq0;
v1 = lp1.a1 * lp1.ic1eq0 + lp1.a2 * v3;
v2 = lp1.ic2eq0 + lp1.a2 * lp1.ic1eq0 + lp1.a3 * v3;
lp1.ic1eq0 = 2 * v1 - lp1.ic1eq0;
lp1.ic2eq0 = 2 * v2 - lp1.ic2eq0;
v0 = lp1.m0 * v0 + lp1.m1 * v1 + lp1.m2 * v2;
);
These would be better off nested (the optimizer is not smart enough to detect that nlp wouldn't change, and it will still run each comparison in sequence even if the first one fails...)
Last edited by Justin; 11-19-2018 at 02:55 PM.
|
|
|
11-19-2018, 02:17 PM
|
#68
|
Human being with feelings
Join Date: Nov 2015
Location: Switzerland
Posts: 1,966
|
uuuh seems like ReEQ is getting a lot of attention... will it maybe be the new native ReaEQ?
|
|
|
11-19-2018, 03:25 PM
|
#69
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
There's an updated ReJJ.zip in the first post. Changes:
* Filters in SVF now cascade using nested conditions (optimization, thank you Justin for spotting)
* Double click on bottom numeric buttons reset values to default
* Double click at bottom margin no longer creates filter bands
I'm holding off on the denorm fix until Justin can confirm if it's something that JSFX ought to be handling. EDIT: apparently it's fixed in JSFX.
Apologies to folks tweaking it already...it's a bit of a moving target whilst I'm fixing things.
Last edited by nitsuj; 11-19-2018 at 03:43 PM.
|
|
|
11-19-2018, 04:21 PM
|
#70
|
Human being with feelings
Join Date: Nov 2015
Location: Switzerland
Posts: 1,966
|
Thanks for the update. Btw I tested ReEQ on my other windows 10 machine and there are no weird lines being drawn while upscaling the interface
|
|
|
11-19-2018, 06:24 PM
|
#71
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,746
|
Quote:
Originally Posted by nitsuj
I'm holding off on the denorm fix until Justin can confirm if it's something that JSFX ought to be handling. EDIT: apparently it's fixed in JSFX.
|
Can try in 5.962+dev1119 (at least in my limited tests on macOS x86_64, that resolved the issue).
I saw an odd issue testing on armv7l linux: when the UI isn't visible the CPU use doubled. I'm baffled by this. Is there anything in the ReEQ code that detects when the @gfx code runs and changes the processing depending on whether it has run recently? If so can you point me in the right direction to see why the slowdown happens?
|
|
|
11-20-2018, 12:30 AM
|
#72
|
Human being with feelings
Join Date: May 2014
Location: Germany
Posts: 643
|
1.4:
- Filters in SVF now cascade using nested conditions (optimization, thank you Justin for spotting)
- Double click on bottom numeric buttons reset values to default
- Double click at bottom margin no longer creates filter bands
1.3:
- fix the oversampling 2x
1.2:
- denorm fixed, click into graphics resets max values
1.1:
- added offset spectrum display, better EQ/Spectrum alignment
|
|
|
11-20-2018, 03:37 AM
|
#73
|
Banned
Join Date: Dec 2016
Location: England
Posts: 2,432
|
great energy flow! keep it up!
right now,there are 2 justins jamming !! =
i think 'authors' would need to give permissions,to make anything native.. be though.
|
|
|
11-20-2018, 04:23 AM
|
#74
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by Justin
I saw an odd issue testing on armv7l linux: when the UI isn't visible the CPU use doubled. I'm baffled by this. Is there anything in the ReEQ code that detects when the @gfx code runs and changes the processing depending on whether it has run recently? If so can you point me in the right direction to see why the slowdown happens?
|
That is odd. There shouldn't be any dependencies between inner DSP code and the UI code. There's shared state of course for the filters and modes but nothing that would give the effect you're seeing as far as a know.
You could try replace the @sample section with some idle work that isn't related to the rest of the code to see if still happens.
|
|
|
11-20-2018, 05:01 AM
|
#75
|
Human being with feelings
Join Date: Aug 2009
Location: NL
Posts: 1,458
|
Thanks for sharing! This looks like it's going to be a staple for me.
Also appreciate that you listed the source of the filters.
|
|
|
11-20-2018, 06:32 AM
|
#76
|
Human being with feelings
Join Date: Oct 2007
Location: home is where the heart is
Posts: 12,110
|
Quote:
Originally Posted by nitsuj
As a side, the reason I had the spectrum and EQ range separate is so that you can see what the filters do at full spectrum range - all the way down to -200db. It's very interesting to see what other plugins do down there in terms of introducing noise and harmonics. I'm a sucker for analysing plugins.
|
I like the separate settings for spectrum and EQ range too, please keep it this way (what concerns me).
|
|
|
11-20-2018, 07:07 AM
|
#77
|
Human being with feelings
Join Date: Nov 2017
Posts: 292
|
Quote:
Originally Posted by sai'ke
Thanks for sharing! This looks like it's going to be a staple for me.
Also appreciate that you listed the source of the filters.
|
You're welcome. And thanks for your latest Filther. Just love the MS-20 non-linear - running a synth bass through that is like catnip for me.
|
|
|
11-20-2018, 07:54 PM
|
#78
|
Human being with feelings
Join Date: Jun 2007
Location: Terra incognita
Posts: 7,670
|
Quote:
Originally Posted by Justin
Can try in 5.962+dev1119 (at least in my limited tests on macOS x86_64, that resolved the issue).
|
Pretty much same here with 5.962+dev1119 on Win10 x64. I noticed just a tiny creep up after the stop, but after a bit longer wait, it seemed to average to about 3-4% higher CPU than when running (using the same 10 instance test as earlier). Also, the FX CPU is now almost 10% lower than with the previous version and same test project.
Quote:
Originally Posted by Justin
I saw an odd issue testing on armv7l linux: when the UI isn't visible the CPU use doubled.
|
And same here too. Noticed it with earlier version already, but got too busy and without time to really verify didn't want to make hasty conclusions...
|
|
|
11-21-2018, 12:39 AM
|
#79
|
Human being with feelings
Join Date: Feb 2009
Location: Reaper HAS send control via midi !!!
Posts: 4,032
|
Quote:
Originally Posted by nitsuj
You're welcome. And thanks for your latest Filther. Just love the MS-20 non-linear - running a synth bass through that is like catnip for me.
|
Two megamasters of jsfx/lua talking. Thanks to all of genious, generous, supportive, open minded, friend developers here and elsewhere!
|
|
|
11-21-2018, 01:04 AM
|
#80
|
Human being with feelings
Join Date: Feb 2009
Location: Reaper HAS send control via midi !!!
Posts: 4,032
|
Quote:
Originally Posted by Justin
Glad you like JSFX, thanks for making such great scripts!
|
Thanks to Reaper, thanks to all jsfx/lua script authors! For many years nobody needs Steinzeitdaws like Cubase any more, where you have to download some GB or TB (now), for some updates? Alone the fact Reaper having all internal features open in the form of actions is fantastic. This is why genious developers like saike can create new .lua scripts like hackey trackey and hackey patterns, which can access all those million Reaper actions, also directly from those new interfaces, in this example the magic jeskola buzz tracker interface, which is the best tracker interface of all time. Reapers open philosophy + another genious entering the scene, saike, brings another century result, bringing the 90s music making style directly into the heart of Reaper. Now, why should someone say, no I want to download a few GB for some nonsense tiny updates for a Steinzeit DAW like Cubase? Or next day, another genious knocks at the door, telling, hey I made ReEQ, ReSpectrum, here, have fun. Wowy, mowy. Next day, saike is bored, knocking, telling, hey I made some experiments with filters, Filther, here, have fun. Wowy, mowy. The islands of Cubase/ProTools getting more and more lonely. The sinking islands. The next film in the cinema, showiing 90 mins, Cubase users downloading 10 GB downloads for updating for tiny changes, nobody wanted. Like Windows, you want to switch off your computer, OS starting some nonsense updates, telling you have to wait... No. Linux.
Last edited by TonE; 11-21-2018 at 01:11 AM.
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 04:16 AM.
|