Hello! I have been "lurking" this forum for a while, but haven't posted. I'm not sure of all the etiquette yet, so please let me know if I'm off base with anything.
I am relatively new to DSP. I have done the Max MSP / PureData thing for a couple years, and finally breaking out into other things as I'm up against the limitations of those frameworks. I am trying to make a MiniMoog Model D, and wanted to write my own implementation of the ladder filter. I was heavily inspired by Urs Heckmann's blog posts from 2016 (One-pole unlimited / monster). This really got the wheels turning for me, and I like the simplicity of the "s" terms as an accumulator for the capacitors. My degree is in Electrical Engineering, so I'm well aquatinted with the idea of the capacitor state as a function of accumulated charge. While this model may be slightly reductive, the advantage seems to be how simple the control is, particularly with setting the cutoff frequency through the "g" term. The blog articles only allude to the full implementation, and since I opted for a non-linear model with global feedback (positive feedback overdrive, not just the negative feedback resonance), I spent quite a bit of time figuring out how to set up a multi-dimensional NR solver. I was documenting this as a went, and realized I basically had half a "research paper", and that it wouldn't be too much more work to format it into something presentable. I also figured I may as well publish it somewhere in hopes that it would assist someone trying to do something similar, and answer some of the questions that I had throughout this process.
For anyone interested, I have sort of a draft copy of the paper hosted here. If you are previewing directly from GitHub, you might have to hit the "more pages" button a couples times - its 13 pages long. I'd welcome any feedback, if anyone notices anything that is not quite right. I also wanted to give Urs Heckmann and Andy Simper a heads up about this, as I cite them and their work several times. I don't have a way to reach either of them, but I know they are quite active here. Thank you both for your work, your papers and recorded talks were an invaluable resource for me!
I am also wondering if anyone has any ideas of the best place to "publish" this? My personal "portfolio" website is on the fritz, and will need to be redone before I host anything there, so I was thinking something along the likes of ResearchGate, ArXiv.org, or Academia.edu. Most of the DSP papers I've read are hosted by specific institutions like CCRMA, DAFX, and ADC, which I currently have nothing to do with (apart from enjoying their content/publications). Any advice would be appreciated!
My next steps are that I am currently working on a C++ implementation of this model (Max external first, then maybe JUCE).
I am relatively new to DSP. I have done the Max MSP / PureData thing for a couple years, and finally breaking out into other things as I'm up against the limitations of those frameworks. I am trying to make a MiniMoog Model D, and wanted to write my own implementation of the ladder filter. I was heavily inspired by Urs Heckmann's blog posts from 2016 (One-pole unlimited / monster). This really got the wheels turning for me, and I like the simplicity of the "s" terms as an accumulator for the capacitors. My degree is in Electrical Engineering, so I'm well aquatinted with the idea of the capacitor state as a function of accumulated charge. While this model may be slightly reductive, the advantage seems to be how simple the control is, particularly with setting the cutoff frequency through the "g" term. The blog articles only allude to the full implementation, and since I opted for a non-linear model with global feedback (positive feedback overdrive, not just the negative feedback resonance), I spent quite a bit of time figuring out how to set up a multi-dimensional NR solver. I was documenting this as a went, and realized I basically had half a "research paper", and that it wouldn't be too much more work to format it into something presentable. I also figured I may as well publish it somewhere in hopes that it would assist someone trying to do something similar, and answer some of the questions that I had throughout this process.
For anyone interested, I have sort of a draft copy of the paper hosted here. If you are previewing directly from GitHub, you might have to hit the "more pages" button a couples times - its 13 pages long. I'd welcome any feedback, if anyone notices anything that is not quite right. I also wanted to give Urs Heckmann and Andy Simper a heads up about this, as I cite them and their work several times. I don't have a way to reach either of them, but I know they are quite active here. Thank you both for your work, your papers and recorded talks were an invaluable resource for me!
I am also wondering if anyone has any ideas of the best place to "publish" this? My personal "portfolio" website is on the fritz, and will need to be redone before I host anything there, so I was thinking something along the likes of ResearchGate, ArXiv.org, or Academia.edu. Most of the DSP papers I've read are hosted by specific institutions like CCRMA, DAFX, and ADC, which I currently have nothing to do with (apart from enjoying their content/publications). Any advice would be appreciated!
My next steps are that I am currently working on a C++ implementation of this model (Max external first, then maybe JUCE).
Statistics: Posted by pscorbett — Sat Jan 27, 2024 8:58 pm — Replies 0 — Views 26