Nice. So now any, say 4 part, chord could be sent to a chain with 4 suitably configured instances, as long as each part (finger) is on a separate channel. (Not much use for live playing, but good for step sequencing)
Nice. So now any, say 4 part, chord could be sent to a chain with 4 suitably configured instances, as long as each part (finger) is on a separate channel. (Not much use for live playing, but good for step sequencing)
My thinking exactly. For live you could use this new plug-in I just came up with. It receives MIDI on any channel, and distributes Note On/Off to channels 1..16.
poly_mono_midi_router:
Code:
desc:Dynamic poly/mono MIDI router
// (c) Theo Niessink 2013
// License: GPL - http://www.gnu.org/licenses/gpl.html
@block
while(
midirecv(ofs, msg1, msg23) ? (
status = msg1 & 0xF0;
cc = msg23 & 0x7F;
note = cc + 1;
velocity = msg23 >> 8;
// Note On
status == 0x90 && velocity ? (
i = 0;
while(
i < 16 ? (
!buf[i] ? (
buf[i] = note;
msg1 = status | i;
n += 1;
0;
) : (
i += 1;
);
) : (
msg1 = 0;
);
);
) :
// Note Off
status == 0x80 || (status == 0x90 && !velocity) ? (
i = 0;
while(
i < 16 ? (
buf[i] == note ? (
buf[i] = 0;
msg1 = status | i;
n -= 1;
0;
) : (
i += 1;
);
) : (
msg1 = 0;
);
);
) :
// All Notes Off
status == 0xB0 && cc == 123 & n ? (
i = 0;
loop(16,
buf[i] = 0;
midisend(ofs, status | i, msg23);
i += 1;
);
msg1 = n = 0;
);
msg1 ? midisend(ofs, msg1, msg23) : 1;
);
);
I guess this would need some more thought for things like Sustain (CC#64) and Pitch Wheel...
Yeah, pretty much everything except Note On and Note Off. The only problem is that REAPER already broadcasts All Notes Off and Pitch Wheel reset on all 16 channels, so some special handling is required to prevent excessive broadcasting.
All effects are stock JS and ReaPlugs, except the MIDI router.
http://forum.cockos.com/showpost.php...5&postcount=44 <--------- you'll have to put this JS code, in a text file called 'tale poly-mono router JS', into your REAPER/effects/tale folder for the attached project to work.
Cool! Can I add the MP3 to the audio samples on the mono_synth mini website?
Oh God! If you like (it's so rough)
I'm really glad you're finding time to develop this JS synth, so sure
I'd really like it if we could start a thread or page somewhere with track templates like the ones I have started above, and your 'poly_blep' project, with sounds taking full advantage of parameter modulation and sidechaining the synth with JS and ReaPlugs. Maybe a zip file on the mini website?
I'm really glad you're finding time to develop this JS synth, so sure
Thanks. I have added it for now, but if you want me to remove it again (or replace it with a less rough version?), just let me know.
Quote:
Originally Posted by hamish
I'd really like it if we could start a thread or page somewhere with track templates like the ones I have started above, and your 'poly_blep' project, with sounds taking full advantage of parameter modulation and sidechaining the synth with JS and ReaPlugs. Maybe a zip file on the mini website?
Sure. What shall we include in the ZIP? My poly_blep.rpp, and all of your track templates?
Sure. What shall we include in the ZIP? My poly_blep.rpp, and all of your track templates?
Yep, that's what I'm thinking, but I'd like to extract your poly_blep.rpp tracks as .RTrackTemplate with a naming convention ie: LD, BS, DR, CH, AR, PD for respectively Lead, Bass, Drum, Chord, Arp and Pad presets.
Then zip them all in one file, which you put on the mono_synth 'mini website'. Unless you have a better idea, which would not surprise me!
First, let me state up front that I'm a rank beginner at programming and JS so be gentle if this isn't the brightest question. I've been studying midi_queue.jsfx-inc the past couple days and I think I have my head around most of it, at least conceptually, except for this:
Code:
head < tail && head[] <= ofs ? (
Quote:
Originally Posted by Tale
1. midiq_remove() spots the MIDI message, and because its offset is equal to the current offset, it removes the message. It returns true to let you know you should call midiq_remove() again.
Could you help a poor hillbilly out and clue me in on why it's <= instead of ==?
Could you help a poor hillbilly out and clue me in on why it's <= instead of ==?
I guess you are right in thinking that == would also work in this case, because ofs always gets incremented by only 1. However, <= is more robust, should any (future) code ever increment ofs with more than 1.
Ah. So it makes sure no events are skipped. Cleaver stuff. Say ofs was incremented by more than 1 for some reason, you would lose sample accuracy at that point?
Thanks for taking the time to explain this. I feel like I'm drinking from a fire hose trying to work through things on my own. These little breakthroughs give me a small ray of hope I may understand it enough to be useful some day.
hello Tale. been looking at your mono synth code with an eye on antialiasing. vfine stuff! and i am humbled by how nicely documented it all is. big thanks for making it available.
i am particularly intrigued by the poly blep bits. i am wondering if that code could be adapted to smooth out arbitrary audio input, or if it is too specific to the oscillator logic? alas, the paper that describes how poly blep works is behind a paywall. and my rusty math skillz would need some work in any case. would you hazard an opinion on whether poly blep for arbitrary input is in fact possible before i go off chasing moonbeamz?
and fwiw, what makes me ask this question is some results i'm seeing from comparing your poly blep oscillators with more typical oversampled antialiasing. for example, this shows a triangle wave sweeping from ~2k-4k. the purple is 4x oversampled, and the yellow is your poly_tri():
yellow has higher peaks, but purple doesn't go as low, so more or less a wash considering the oversampled wave takes 4x the cpu. but here the purple is 4x oversampled *and* poly_tri():
quite a significant improvement, for very little extra cpu. so if poly blep could also be used on arbitrary audio input it could be a big win.
thanks for your consideration.
enjoy! /dan
ps- for comparison, here is a non-oversampled plain triangle wave in purple with poly_tri() in yellow:
so poly_tri() is in fact making a big improvement by itself.
I don't think you can use polyBLEPs for arbitrary audio input. The basic idea is to replace the non-bandlimited transient (from +1 to -1) with a more smooth, quasi-bandlimited polynomial transient. Arbitrary audio will contain transients other than just from +1 to -1, and I think that each of these transients would need their own polynomial. Also, you would have to look ahead a few samples, because the polynomial should start just before the transient.
I don't think you can use polyBLEPs for arbitrary audio input. The basic idea is to replace the non-bandlimited transient (from +1 to -1) with a more smooth, quasi-bandlimited polynomial transient. Arbitrary audio will contain transients other than just from +1 to -1, and I think that each of these transients would need their own polynomial. Also, you would have to look ahead a few samples, because the polynomial should start just before the transient.
thanks for the link. yup, my rusty math needs work. :^) though Urs comment was intriquing since i used to do computer graphics code and just used a bit of Catmull's work to do cubic interpolation in wildwavefu. i'd be happy to accept a bit of latency to smooth arbitrary tranients. so i'll think on that. but i'm confused if you say poly bleps only do +1 to -1 transients what's going on with the poly_tri() wave? i see you are somehow integrating a square wave there. so i imagine the poly blep logic is smoothing things where the tri changes direction somehow. no need to reply here. enjoy! /dan
but i'm confused if you say poly bleps only do +1 to -1 transients what's going on with the poly_tri() wave? i see you are somehow integrating a square wave there. so i imagine the poly blep logic is smoothing things where the tri changes direction somehow. no need to reply here.
I think I have read somewhere that it is possible to come up with a polynomial for the triangular wave's transients, but it will be different from the one used for sawtooth/square/rectangle. Anyway, my math skills are too weak to come up with such a polynomial, so I just integrated a polyBLEP square to obtain a triangle. The same goes for the full-wave rectified sine, which I obtain by integrating a saw.
+1 for a polysynth.
This thing is awesome. The osc sounds superb.
I'm just getting into synth creation and i SO wish i could learn how to implement these type of idea into new synths.
Please keep developping this and if you need help ( tho i cannot code ) i will be more then happy to give you a hand
I tried to modulate the pulse width with reaper's own parameter modulation and it modulates the noise level instead.
Just taught you should know
Ha ha, I tried to use parameter modulation to get FM (also no way), however I totally agree about the good sounding osc, also the envelopes are excellent, all up it's a brilliant little monosynth
Hi Theo,
could you add a slider to control the number of output channels, please?
With this incarnation I would have to either;
a) Use 16 synths to avoid dropping any notes
b) Modify the plugin for each application (easily done, but not terribly efficient)
Cheers!
Dax.
Quote:
Originally Posted by Tale
My thinking exactly. For live you could use this new plug-in I just came up with. It receives MIDI on any channel, and distributes Note On/Off to channels 1..16.
poly_mono_midi_router:
Code:
desc:Dynamic poly/mono MIDI router
// (c) Theo Niessink 2013
// License: GPL - http://www.gnu.org/licenses/gpl.html
@block
while(
midirecv(ofs, msg1, msg23) ? (
status = msg1 & 0xF0;
cc = msg23 & 0x7F;
note = cc + 1;
velocity = msg23 >> 8;
// Note On
status == 0x90 && velocity ? (
i = 0;
while(
i < 16 ? (
!buf[i] ? (
buf[i] = note;
msg1 = status | i;
n += 1;
0;
) : (
i += 1;
);
) : (
msg1 = 0;
);
);
) :
// Note Off
status == 0x80 || (status == 0x90 && !velocity) ? (
i = 0;
while(
i < 16 ? (
buf[i] == note ? (
buf[i] = 0;
msg1 = status | i;
n -= 1;
0;
) : (
i += 1;
);
) : (
msg1 = 0;
);
);
) :
// All Notes Off
status == 0xB0 && cc == 123 & n ? (
i = 0;
loop(16,
buf[i] = 0;
midisend(ofs, status | i, msg23);
i += 1;
);
msg1 = n = 0;
);
msg1 ? midisend(ofs, msg1, msg23) : 1;
);
);
I guess this would need some more thought for things like Sustain (CC#64) and Pitch Wheel...
Resonance is simply a CC (Control Change) message, exactly like sustain (which is CC#64).
Incidentally, resonance is plugin-specific and here's a list of common CCs:
Hi Theo, I've played with the synth a little today the 20th December version and the plug is still not returning 'last touched' to the 'Param' button correctly on many sliders. Examples - Volume seems to return OK, Tuning is OK, Attack returns 'Decay', Decay retuns 'Sustain' and so on.
On many plugins this is a handy way to invoke an envelope or Parameter Modulation, which is currently broken in Simple mono synth.
Hi Theo, I've played with the synth a little today the 20th December version and the plug is still not returning 'last touched' to the 'Param' button correctly on many sliders. Examples - Volume seems to return OK, Tuning is OK, Attack returns 'Decay', Decay retuns 'Sustain' and so on.
On many plugins this is a handy way to invoke an envelope or Parameter Modulation, which is currently broken in Simple mono synth.
Hey, that is weird. Actually I think that is a bug in REAPER, but I will look into it, and see if I can fix it at my end.
v20130404
Can I add the MP3 to the audio samples on the mono_synth mini website?
I have added it for now, but if you want me to remove it again (or replace it with a less rough version?), just let me know.
Linear Mode
