music DSP research
audio plugin development
My first collaboration as a freelancer on a big synth plugin. I led DSP engineering here.
The sound engine hosts a full, high-quality, component-level model of the Buchla Music Easel synthesizer, with extra added features (e.g., polyphony, MIDI sync, extra modulations).
I took the lead of the DSP development and performed all of the circuit modeling.
The plugin has instantly turned into a best seller, won the MIPA 2018 award as part of V Colleection 6, and has received very positive reviews.
Arturia Minifilter V and Mini V filter
Especially since Minifilter V was released as a free Christmas gift for a couple of weeks in late 2016 – a circumstance that was sufficiently covered by the specialized press – I got an overwhelming amount of feedback, and almost unanimously positive.
Minifilter V has been such a success in terms of downloads that Arturia server admins still remember those give-away days with horror.
Arturia guitar amp simulators
Farfisa V also contains a similar guitar amp model without vibrato.
These two algorithms manage to reproduce the main electrical, mechanical, and acoustic effects of the amps + microphones with minimal CPU usage and more-than-reasonable sound quality.
Arturia Stage-73 V preamp and tone controls
I have developed a state-space model of the original suitcase-model preamp circuit, where the audio path essentially consists of an imperfectly-buffered active Baxandall-type equalizer, and the control path contains a vibrato LFO with output opto-isolators.
For the stage model, I have provided reproductions of two tightly input-coupled passive tone circuits.
Arturia B-3 V preamp and expression pedal
The plugin contains a model I have developed of the non-vibrato channel of the original tube preamp, which is directly interconnected with the capacitive expression pedal.
Effects and various DSP for Arturia
I have developed several effects and pieces of DSP that are used in Arturia products, such as
- an overdrive, modeled after a famous Japanese guitar pedal from the late 1970s,
- an analog delay, inspired by a BBD-based guitar pedal from the 1970s,
- a wah pedal, emulating the original tenderhearted device from the 1960s,
- a flanger, of the subtractive thru-zero type,
- a bitcrusher,
- a limiter,
- a parametric EQ,
- a step sequencer,
- a multi-point envelope generator,
- a IEC 60268-18-conformant digital peak programe meter.
Doctoral disseration, Aalto University, Espoo, Finland, November 2014.
Concise, down-to-earth, yet scientifically rigorous and up-to-date introduction to white-box VA modeling.
It touches on traditional circuit simulation and MNA, state-space methods, WDFs, and ZDF techniques, then details practical applications of those principles in my previous research.
S. D’Angelo and V. Välimäki, IEEE/ACM Trans. Audio, Speech, and Lang. Process., vol. 22, no. 12, pp. 1825–1832, December 2014. DOI: 10.1109/TASLP.2014.2352495.
In-depth linear response analysis of the ladder filter generalized to any number of stages.
Various alternative parameterization strategies are considered, and a low cost linear digital model that correctly reproduces the frequency response is proposed.
Generalized Moog Ladder Filter: Part II–Explicit Nonlinear Model through a Novel Delay-Free Loop Implementation Method
S. D’Angelo and V. Välimäki, IEEE/ACM Trans. Audio, Speech, and Lang. Process., vol. 22, no. 12, pp. 1873–1883, December 2014. DOI: 10.1109/TASLP.2014.2352556.
A new ZDF technique is proposed that is fully non-iterative, perfectly accounts for linear behavior, requires minimal topology modifications and no trasformation of nonlinearities.
It is then applied to discretize a large-signal model of the ladder filter generalized to any number of stages, leading to a lightweight and high-accuracy implementation with strong stability guarantees.
J. Parker and S. D’Angelo, in Proc. 16th Intl. Conf. Digital Audio Effects (DAFx-13), pp. 278–285, Maynooth, Ireland, September 2013.
Simple and efficient implementation of both the control and audio paths, as well as their peculiar optocoupler-based interconnection.
Dynamic and nonlinear effects are accurately reproduced and stability is preserved.
R. C. D. de Paiva, S. D’Angelo, J. Pakarinen, and V. Välimäki, IEEE Trans. Circ. Systems–II: Express Briefs, vol. 59, no. 10, pp. 688–692, October 2012. DOI: 10.1109/TCSII.2012.2213358.
Simple WD opamp model, and explicit WD models of single and antiparallel diodes based on the Lambert W function.
They are usd to simulate two common guitar distortion circuits with high accuracy and minimal CPU usage.
S. D’Angelo and L. Gabrielli, in Proc. 21st Intl. Conf. Digital Audio Effects (DAFx-18), pp. 107–112, Aveiro, Portugal, September 2018.
A cheap and tonally-transparent time-domain method to extend the duration of any audio signal without sounding static.
Developed to help the guys at MIND Music Labs demonstrate their ELK MusicOS at NAMM 2018, which eventually ended up being nominated as one of the best new products by the MIDI Manufacturers Association.
S. D’Angelo, J. Pakarinen, and V. Välimäki, IEEE Trans. Audio, Speech, and Lang. Process., vol. 21, no. 2, pp. 313–321, February 2013. DOI: 10.1109/TASL.2012.2224340.
Two high-accuracy and delay-free vacuum tube triode models for use in WDFs, based on a mixed physical-interpolative approach.
Tests show lower CPU usage and richer harmonic response than previous models. The algorithm is flexible enough to let the interpolation scheme be easily modified.
Wave-Digital Polarity and Current Inverters and Their Application to Virtual Analog Audio Processing
S. D’Angelo and V. Välimäki, in Proc. Intl. Conf. Acoustics, Speech, and Signal Process. (ICASSP 2012), pp. 469–472, Kyoto, Japan, March 2012. DOI: 10.1109/ICASSP.2012.6287918.
Clever yet fundamental (IMHO) addition to basic WDFs, where two new adaptors allow to easily interconnect subnetworks using different polarity and sign conventions.
They are trivial to implement, fully compatible with existing WD structures, perfect to avoid remodeling asymmetrical subnetworks, and proven to never cause stability issues.
S. D’Angelo, in Proc. 22nd Colloquium on Music Informatics (XXII CIM), pp. 59–63, Udine, Italy, November 2018.
Essential references and practical advices to help DSP programmers implement efficient real-time circuit simulation algorithms.
This was an invited paper to a national conference that has a long and important tradition.
S. D’Angelo and V. Välimäki, in Proc. Intl. Conf. Acoustics, Speech, and Signal Process. (ICASSP 2013), pp. 729–733, Vancovuer, Canada, May 2013. DOI: 10.1109/ICASSP.2013.6637744.
Detailed derivation of a large-signal model of the ladder filter and development of a digital model that outpeforms previous attempts in nonlinear behavior.
Largely made obsolete by more recent research I have conducted.
I was born in 1987 and have always been interested in computers and music. I studied classical piano as child and electric bass in my high school years. I played hundreds of gigs.
In 2008 and 2010 I got B.Sc. and M.Sc. degrees in computer engineering from Politecnico di Torino, Turin, Italy. In the meanwhile I got involved in free/open source music software projects, started one of my own, and participated to two editions of the Google Summer of Code program as a mentor. Also, in 2008 I was an intern in an electronic engineering company developing embedded software, and in 2010/2011 I worked in a research institute on a big EU-funded project.
Then I started my doctoral studies, which led to me getting the D.Sc. (Tech.) degree in signal processing and acoustics from the Aalto University, Espoo, Finland, in 2014. Since then I am actively involved in scientific conferences, peer review, collaborations with universities, etc.
From April 2015 to December 2017 I have been leading DSP research and development at Arturia in Grenoble, France.
Since March 2018 I work as a freelancer from Italy for companies that need cutting edge music DSP technology. I try to combine scientific rigor, industry experience, creativity, and empathy. I tend to favor long-term collaborations.
You can take a look at a more formal CV.