[Csnd] more efficient lush synth pad?
| Date | 2019-03-13 01:34 |
| From | Ed Cashin |
| Subject | [Csnd] more efficient lush synth pad? |
Hi! In csound I am happy to find I can make a synth pad sound that I imagined. However! To avoid overwhelming my laptop, I have to turn ksmps way up (1024) or turn JACK buffering way up. I suspect I'm maybe missing some opportunities to use tables in order to make things easier computationally. I would really appreciate hearing any suggestions on how to make the code below leaner while maintaining the lush and dynamic sound. saws-pad.csd by Ed Cashin 2019, based on PolyphonyControl.csd Written by Iain McCurdy 2008, Rewritten 2012 (originally entitled 'MonoOverlaps.csd'). <CsoundSynthesizer> <CsOptions> ;UNCOMMENT THE FOLLOWING LINE OF CODE IF YOU DO NOT HAVE A MIDI KEYBOARD CONNECTED ; -odac -dm0 -M0 -+rtmidi=virtual ;-odac -3 -+rtaudio=jack -+rtmidi=alsa -M hw:5,0 -dm0 -odac -3 -+rtaudio=jack -+rtmidi=alsa -M hw:5,0 -dm0 -B 4096 </CsOptions> <CsInstruments> sr = 88200 ksmps = 1024 nchnls = 2 0dbfs = 1 FLcolor 255, 255, 255, 100, 200, 200 FLpanel "Polyphony Control", 500, 250, 0, 0 ; ON | OFF | TYPE | WIDTH | HEIGHT | X | Y | OPCODE | INS | STARTTIM | IDUR gkExit,ihExit FLbutton "Exit Example", 1, 0, 21, 120, 25, 375, 200, 0, 999, 0, 0.001 ;VALUE DISPLAY BOXES WIDTH | HEIGHT | X | Y idatt FLvalue " ", 60, 20, 5, 25 iddec FLvalue " ", 60, 20, 5, 75 idsus FLvalue " ", 60, 20, 5, 125 idrel FLvalue " ", 60, 20, 5, 175 ; MIN | MAX | EXP | TYPE | DISP | WIDTH | HEIGHT | X | Y gkatt, ihatt FLslider "Attack Time", 0.01, 12, -1, 23, idatt, 490, 25, 5, 0 gkdec, ihdec FLslider "Decay Time", 0.01, 8, -1, 23, iddec, 490, 25, 5, 50 gksus, ihsus FLslider "Sustain Level", 0, 1, 0, 23, idsus, 490, 25, 5, 100 ;SUSTAIN LEVEL CANNOT BE ZERO FOR mxadsr OPCODE gkrel, ihrel FLslider "Release Time", 0.01, 12, -1, 23, idrel, 490, 25, 5, 150 ;COUNTERS MIN | MAX | STEP1 | STEP2 | TYPE | WIDTH | HEIGHT | X | Y | OPCODE gkPolyLimit, ihPolyLimit FLcount "Polyphony Limit", 1, 127, 1, 1, 2, 120, 25, 5, 200, -1 FLsetVal_i 0.27, ihatt FLsetVal_i 4, ihdec FLsetVal_i 0.42, ihsus FLsetVal_i 1.77, ihrel FLsetVal_i 6, ihPolyLimit FLpanel_end ;END OF PANEL CONTENTS FLrun ;RUN THE WIDGET THREAD! ;NUMBER OF ACTIVE NOTES ;(NOTES IN RELEASE STAGE REGARDED AS INACTIVE) gkactive init 0 ; output audio gaL init 0 gaR init 0 instr 1 ; NOTE TRIGGERING INSTRUMENT gkactive init i(gkactive) + 1 ;INCREMENT NOTE COUNTER if gkactive > i(gkPolyLimit) then turnoff endif krel release ;IF NOTE HELD = 0, IF NOTE RELEASED = 1 ktrig trigger krel,0.5,0 ;WHEN RELEASE FLAG CROSSES 0.5 UPWARDS, I.E. NOTE HAS BEEN RELEASED... if ktrig==1 then gkactive = gkactive - 1 ;...DECREMENT ACTIVE NOTES COUNTER endif icps cpsmidi ;READ CYCLES PER SECOND VALUES FROM MIDI KEYBOARD INPUT iamp ampmidi 1 ; ATT | DEC | SUS | REL aenv madsr i(gkatt), i(gkdec), i(gksus), i(gkrel) ;MIDI SENSING ADSR ENVELOPE WITH LINEAR SEGMENTS kfatt = gkatt kfsus ntrpol gksus, iamp, 0.8 kfco madsr i(kfatt), i(gkdec), i(kfsus), i(gkrel) ;MIDI SENSING ADSR ENVELOPE WITH LINEAR SEGMENTS ;OUTPUT OPCODE AMP CPS FUNCTION_TABLE asig vco2 0.3*iamp, icps, 0 ;AUDIO OSCILLATOR ispread init 0.25 ; Hz ichaos init 0.6 asaw1 vco2 iamp / 4, icps - ispread, 0 asaw2 vco2 iamp / 4, icps + ispread, 0 klfo1 lfo 0.4, 0.05, 0 klfo2 lfo 0.3, 0.15, 0 anoise1 noise ichaos, 0 anoise2 noise ichaos, 0 asaw3 vco (iamp / 4) * anoise1, icps + klfo1, 0 asaw4 vco (iamp / 4) * anoise2, icps + klfo2, 0 asub vco2 iamp / 2, icps / 2, 12 kfoldmod1 lfo 3, 0.3 kfoldmod2 lfo 4, 0.2 asub1 fold asub, 9.5 + kfoldmod1 asub2 fold asub, 10.5 + kfoldmod2 aL = asig + asub1 + asaw1 + asaw3 aR = asig + asub2 + asaw2 + asaw4 kfiltmod1 lfo 500, 0.2 kfiltmod2 lfo 400, 0.1 kfco = 1 aL moogladder aL, (7000 + kfiltmod1) * kfco, 0.3, 1 aR moogladder aR, (7000 + kfiltmod2) * kfco, 0.3, 1 gaL clip gaL + (aL * aenv), 0, 0.7 gaR clip gaR + (aL * aenv), 0, 0.7 endin instr 99 arL, arR reverbsc gaL, gaR, 0.85, 12000, sr, 0.5, 1 aoutL ntrpol gaL, arL, 0.6 aoutR ntrpol gaR, arR, 0.6 outs aoutL, aoutR gaL = 0 gaR = 0 endin instr 999 ;QUIT exitnow ;EXIT CSOUND IMMEDIATELY endin </CsInstruments> <CsScore> f 1 0 32768 10 1 f10 0 65536 10 1 0 0 0 0 0 0 0 0 .05 i 99 0 3600 ; reverb </CsScore> </CsoundSynthesizer> |
| Date | 2019-03-13 10:00 |
| From | "Jeanette C." |
| Subject | Re: [Csnd] more efficient lush synth pad? |
Mar 13 2019, Ed Cashin has written:
...
> I suspect I'm maybe missing some opportunities to use tables in order to
> make things easier computationally. I would really appreciate hearing any
> suggestions on how to make the code below leaner while maintaining the lush
> and dynamic sound.
...
Hi Ed,
I'll not work on your code directly. But here are some ideas:
there is the "padsynth" GEN routine to create chorusing/beating sounds.
It's the technique developed by Nasca Octavian Paul for ZynAddSubFX
(Yoshimi).
You could also try for a chorus, either using the flanger effect or
modulating one or more delay lines with LFOs.
Not sure how much this would do: but the randh noise has a frequency
parameter, so you can turn it down. Something like this might help in
variations:
knoise randh kmodamp, kmodfreq
kmodnoise port knoise, islewtime
Other than that I *THINK* that vco2 is quite efficient generating a
sawtooth.
You could try using a sine wave with oscil for your "subharmonic". It
might not change the sound drastically and could decrease CPU usage a
bit.
Best wishes,
Jeanette
--
* Website: http://juliencoder.de - for summer is a state of sound
* SoundCloud: https://soundcloud.com/jeanette_c
* Youtube: https://www.youtube.com/channel/UCMS4rfGrTwz8W7jhC1Jnv7g
* GitHub: https://github.com/jeanette-c
* Twitter: https://twitter.com/jeanette_c_s
Don't worry, you're gonna be alright,
But Cinderella's got to go <3
(Britney Spears)
Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here |
| Date | 2019-03-13 15:13 |
| From | thorin kerr |
| Subject | Re: [Csnd] more efficient lush synth pad? |
| Attachments | padoscil.csd |
Hi Ed, I've found lush pad sounds using a combination of vco2init, vco2ft, and oscbnk. oscbnk lets you have control over the number of oscillators, so, you have some control over efficiency (albeit trading it for 'lushness'). I've attached an example, having a go at the sound you were happy with. See what you think. Thorin On Wed, 13 Mar. 2019, 11:34 am Ed Cashin, <ecashin@noserose.net> wrote:
|
| Date | 2019-03-13 16:06 |
| From | Ed Cashin |
| Subject | Re: [Csnd] more efficient lush synth pad? |
Hi, Jeanette. Thanks very much for the suggestions! I'll respond to selected quotes below. On Wed, Mar 13, 2019 at 6:00 AM Jeanette C. <julien@mail.upb.de> wrote: ... there is the "padsynth" GEN routine to create chorusing/beating sounds. Thanks. I tried that opcode quite some time ago but should revisit it. You could also try for a chorus, either using the flanger effect or That kind of modulation is different from the ones I have now, but I did add the ensembleChorus UDO by Bhob Rainey at one point. It's a useful tweak, but it adds to the computational complexity, as it cannot be used instead of the rest of the instrument. Not sure how much this would do: but the randh noise has a frequency That's a great idea and something I will try! Thanks. You could try using a sine wave with oscil for your "subharmonic". It I preferred the current sound with the folds on the triangle, but maybe that's a compromise to consider. Ed Cashin <ecashin@noserose.net> |
| Date | 2019-03-13 16:08 |
| From | Ed Cashin |
| Subject | Re: [Csnd] more efficient lush synth pad? |
Hi, Thorin! Those are super suggestions, and I appreciate the example. I’m not sure why it gives me a lot of dac buffer underruns at first, but before it ends I do hear a promising and similar pad sound. I’ll definitely follow up on this approach. Thanks very much. On Wed, Mar 13, 2019 at 11:14 AM thorin kerr <thorin.kerr@gmail.com> wrote:
Ed Cashin <ecashin@noserose.net> |
| Date | 2019-03-13 21:56 |
| From | thorin kerr |
| Subject | Re: [Csnd] more efficient lush synth pad? |
Hmm, there are a lot of tables initialised by vco2init at the beginning. Just speculating, but do you have extra flags like --realtime or --num-threads in play? Thorin On Thu, 14 Mar. 2019, 2:08 am Ed Cashin, <ecashin@noserose.net> wrote:
On Thu, 14 Mar. 2019, 2:08 am Ed Cashin, <ecashin@noserose.net> wrote:
|
| Date | 2019-03-13 22:58 |
| From | Bernt Isak Wærstad |
| Subject | Re: [Csnd] more efficient lush synth pad? |
I think moogladder is quite expensive, so perhaps try to not have an instance on each voice and move it to the global instrument instead? Or try with something like lpf18 instead On Wed, 13 Mar 2019 at 22:56, thorin kerr <thorin.kerr@gmail.com> wrote:
Mvh.
Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here
Bernt Isak Wærstad |
| Date | 2019-03-14 00:17 |
| From | Ed Cashin |
| Subject | Re: [Csnd] more efficient lush synth pad? |
I had changed the sample rate to 44100, but the buffer problems stop if I add 10 to the start times and change back to 44100. On Wed, Mar 13, 2019 at 5:56 PM thorin kerr <thorin.kerr@gmail.com> wrote:
-- Ed Cashin <ecashin@noserose.net> |
| Date | 2019-03-14 00:18 |
| From | Ed Cashin |
| Subject | Re: [Csnd] more efficient lush synth pad? |
I'll try a different filter, yes. (I do want to keep each note filtered independently if possible, though.) On Wed, Mar 13, 2019 at 6:58 PM Bernt Isak Wærstad <berntisak@gmail.com> wrote:
-- Ed Cashin <ecashin@noserose.net> |
| Date | 2019-03-15 11:45 |
| From | "Dr. Richard Boulanger" |
| Subject | Re: [Csnd] more efficient lush synth pad? |
Ed, Really nice instrument - thanks for sharing. -dB _____________________________________________ Dr. Richard Boulanger Professor of Electronic Production and Design Professional Writing and Music Technology Division Berklee College of Music ______________________________________________ OFFICE: 1126 Boylston St., Suite 201 (EPD), Suite 208 (Boulanger) ______________________________________________ President of Boulanger Labs - http://boulangerlabs.com Author & Editor of The Csound Book - http://mitpress.mit.edu/books/csound-book Author & Editor of The Audio Programming Book - http://mitpress.mit.edu/books/audio-programming-book ______________________________________________ about: http://www.boulangerlabs.com/#about about: http://www.csounds.com/community/developers/dr-richard-boulanger/ music: http://www.csounds.com/community/developers/dr-richard-boulanger/dr-richard-boulanger-music/ ______________________________________________ email: rboulanger@berklee.edu facebook: https://www.facebook.com/richard.boulanger.58 On Tue, Mar 12, 2019 at 9:35 PM Ed Cashin <ecashin@noserose.net> wrote:
|
| Date | 2019-03-15 11:46 |
| From | "Dr. Richard Boulanger" |
| Subject | Re: [Csnd] more efficient lush synth pad? |
Thorin, A nice alternative OscilPad instrument. Thanks for sharing! -dB _____________________________________________ Dr. Richard Boulanger Professor of Electronic Production and Design Professional Writing and Music Technology Division Berklee College of Music ______________________________________________ OFFICE: 1126 Boylston St., Suite 201 (EPD), Suite 208 (Boulanger) ______________________________________________ President of Boulanger Labs - http://boulangerlabs.com Author & Editor of The Csound Book - http://mitpress.mit.edu/books/csound-book Author & Editor of The Audio Programming Book - http://mitpress.mit.edu/books/audio-programming-book ______________________________________________ about: http://www.boulangerlabs.com/#about about: http://www.csounds.com/community/developers/dr-richard-boulanger/ music: http://www.csounds.com/community/developers/dr-richard-boulanger/dr-richard-boulanger-music/ ______________________________________________ email: rboulanger@berklee.edu facebook: https://www.facebook.com/richard.boulanger.58 On Wed, Mar 13, 2019 at 11:13 AM thorin kerr <thorin.kerr@gmail.com> wrote:
|