Author |
Message |
Tim Kleinert
Joined: Mar 12, 2004 Posts: 1148 Location: Zürich, Switzerland
Audio files: 7
G2 patch files: 236
|
Posted: Wed Mar 07, 2012 4:43 am Post subject:
Anti-aliased osc sync and user-definable waveforms Subject description: Two remaining G2 problems solved |
|
|
Hi folks,
As the title says, I solved two remaining G2 problems yesterday: generating aliasing-free oscillator synchronisation and aliasing-free user-definable waveforms.
Unfortunately these solutions are quite expensive in DSP usage, but they work.
I published the circuits in the building block section.
Anti-aliased osc sync:
http://www.electro-music.com/forum/post-355719.html#355719
Anti-aliased user-definable waveforms:
http://www.electro-music.com/forum/post-355718.html#355718
Nothing this red box can't do.
cheers to all,
Tim
EDIT: Warning! These circuits push the limits hard and can malfunction on the demo software under certain conditions. The G2 hardware has no problem though.
EDIT 2: I detected a problem when sweeping the osc sync range more rapidly (occasional subtle audio clicks). No idea what causes this. Otherwise it works well (no "bird tones" when pitching into high registers).
EDIT 3: For a comparison between anti-aliased and non-anti-aliased, simply deactivate the 36dB filter. In higher registers, the difference can be huge. |
|
Back to top
|
|
|
mosc
Site Admin
Joined: Jan 31, 2003 Posts: 18197 Location: Durham, NC
Audio files: 212
G2 patch files: 60
|
|
Back to top
|
|
|
Tim Kleinert
Joined: Mar 12, 2004 Posts: 1148 Location: Zürich, Switzerland
Audio files: 7
G2 patch files: 236
|
|
Back to top
|
|
|
Roland Kuit
Joined: Sep 29, 2003 Posts: 1090 Location: The Netherlands/Sweden
Audio files: 8
G2 patch files: 127
|
|
Back to top
|
|
|
Roland Kuit
Joined: Sep 29, 2003 Posts: 1090 Location: The Netherlands/Sweden
Audio files: 8
G2 patch files: 127
|
Posted: Sun Mar 11, 2012 3:02 am Post subject:
|
|
|
And don't forget PLL sync. Normally this is done with a S&H module. Now I used this logic trick and a delay module. The funny part is, you can ditch the syncing osc in this pch too:
Description: |
|
Filesize: |
99.14 KB |
Viewed: |
17031 Time(s) |
|
Description: |
|
Download |
Filename: |
PLL Sync 10.pch2 |
Filesize: |
1.56 KB |
Downloaded: |
4115 Time(s) |
|
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Wed Mar 21, 2012 7:28 pm Post subject:
|
|
|
Hi Tim.. nice work again...
Now i really do an own TK patches folder
I havent used the G2 for a while, at least havent done new patches...just getting back to it now and defently have to dig a bit deeper in all the sophisticated building blocks you have done in the last years..
Have you really solved all the impossibilities of the G2 now?
dorremifasol mentions the "juicy filter" border in your
acompaning thread in the buildings block section.
Actually that is still a limit somehow where i havent found a conclusive solution yet too. where one building block becomes the filter of choice when it gets demanding.
Actually no digital system has shown a filter yet that really matches the analog ones. But i guess its possible to get closer as it is in the G2 with the right model..Just,
a difficult thing where you probably need to have some gain staging and saturation curves happening you cant get out of the internal modules so easily.
I wonder whether the techniques you have developed can provide some alternative ways to address that topic. One could try to mangle the audio.. Or model overtones with pitch detection and dynamic envelope following to model resonance clusters.. I actually think that there is a bit the problem.
In an analog filter you dont have equal poles, and these poles are not just slightly out of tune, when the trasistors dont match they add distortions and they interact with the other poles. Depedend on the circuit of cause..
In general the internal overdrive behavior on feedback loops of the G2 is a bit critical. And there is maybe a point where your number crunching routines on sample level might provide some solutions??? I dont know enough about your patches yet ..just realized that you somehow really patch litltel software subroutines somehow, directly adressing infrastructural limits that way.
Anyway.. maybe an idea for an own thread.. to collect where we are at in the filter department of the G2 in the year 2012.
What is the best alternative filter design made for the g2 sofar?
DIY or combination of exsisting filter module approaches..
Modeling approaches? Or just using sine oscilators and simulate teh resonance instead performing it?
You demonstrate pretty well that the G2 is not finally explored yet..
Maybe there are still tricks to learn in the filter department too.
What do you guys think? An extra thread to collect filter ideas?..
Maybe as a challenge to over think what we have in our libs already, or as motivation to try again to get it a step further?
I actually would like one day to finish my 303 emulation for the g2 project. Or get one step further now.
I have gathered some new insights during the last years by working with a real one again.
greats
Sven |
|
Back to top
|
|
|
Tim Kleinert
Joined: Mar 12, 2004 Posts: 1148 Location: Zürich, Switzerland
Audio files: 7
G2 patch files: 236
|
Posted: Thu Mar 22, 2012 6:03 pm Post subject:
|
|
|
Filters with embedded nonlinearities are hard because you really have to oversample them to get a good result. And oversampling is impossible on the G2. I can't make it calculate faster than 96k. Also, with IIR filters (just like any type of feedback), floating point vs fixed point becomes a real big issue. And the G2 is fixed point arithmetic unfortunately. (I've also come to suspect that this fact is also the reason why it's almost impossible to get a really good sounding DIY reverb patch, even with designs that max out an entire expanded G2.) |
|
Back to top
|
|
|
3phase
Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Fri Mar 23, 2012 3:08 am Post subject:
|
|
|
Would it be possible with your box of tricks to do oversampling for a lofi version of a filter? like running on quarter rate?
Or is it just not possible on any rate because of structural limits? |
|
Back to top
|
|
|
Tim Kleinert
Joined: Mar 12, 2004 Posts: 1148 Location: Zürich, Switzerland
Audio files: 7
G2 patch files: 236
|
Posted: Fri Mar 23, 2012 6:23 am Post subject:
|
|
|
3phase wrote: | Would it be possible with your box of tricks to do oversampling for a lofi version of a filter? like running on quarter rate?
Or is it just not possible on any rate because of structural limits? |
I'll think about it.
My G2 "solutions" usually suddenly pop up in my mind in a flash of insight after seeding my subconscious with the problem at hand. This can take days, months or years. I had a hunch about the possibility of anti-aliased DIY oscillators some time ago, after I discovered that the G2 delay lines (buggy as they may be) have decent fractional sample interpolation. But only after getting other things to work (in this case a 96k DIY sawtooth counter with accurate pitch tracking and resolution) did every thing fall into place one morning in a flash (I was under the shower ).
So I've seeded my subconscious with this filter thing this morning. Let's see what happens. I can't guarantee anything though. |
|
Back to top
|
|
|
dorremifasol
Joined: Sep 28, 2006 Posts: 814 Location: Barcelona, Spain
Audio files: 7
G2 patch files: 49
|
Posted: Fri Mar 23, 2012 6:28 am Post subject:
|
|
|
The shower, great source of inspiration!
When I need to compose a melody in a rush for the work, it usually comes in the shower at home.
I suppose it's because in the shower one is away from other distractions, and relaxed. _________________ Cheers,
Albert |
|
Back to top
|
|
|
Baby
Joined: Apr 24, 2006 Posts: 45 Location: Laos Angeles
Audio files: 2
G2 patch files: 15
|
|
Back to top
|
|
|
Tim Kleinert
Joined: Mar 12, 2004 Posts: 1148 Location: Zürich, Switzerland
Audio files: 7
G2 patch files: 236
|
Posted: Fri Mar 23, 2012 11:22 am Post subject:
|
|
|
Baby wrote: | My filter solution.. | Neat. Would you mind posting it in the building blocks section? |
|
Back to top
|
|
|
dasz
Joined: Oct 16, 2004 Posts: 1644 Location: victoria, canada
Audio files: 29
G2 patch files: 56
|
Posted: Fri Mar 23, 2012 6:24 pm Post subject:
|
|
|
Shower and walks is what do it for me, and often I get both at the same here in Vancouver, Canada (rain showers while on a walk).
Nice Work, tim!
Dasz |
|
Back to top
|
|
|
varice
Joined: Dec 29, 2004 Posts: 961 Location: Northeastern shore of Toledo Bend
Audio files: 29
G2 patch files: 54
|
Posted: Sat Mar 24, 2012 8:00 pm Post subject:
|
|
|
Baby wrote: | My filter solution.. |
I have a similar filter solution using a modified Moog moogerfooger Low Pass Filter and modified G2X:
http://electro-music.com/forum/topic-12266.html _________________ varice |
|
Back to top
|
|
|
pyrosonic
Joined: Jul 12, 2008 Posts: 383 Location: Kent,Oh. USA
|
|
Back to top
|
|
|
|