arjf

6 ms·

1.
by arjf 3mo ago
Most likely microphone lower range limitation. The fundamental frequency is out of the microphone’s range, so you only see the harmonics.
2.
by arjf 3mo ago
32.7 Hz is C1, lowest C on a standard piano (A440 tuning). No inherent limitation of the technique, the resonators can be tuned to much lower frequencies. But indeed you‘ll need a better mic than the iphone’s to get frequencies below 50Hz i
3.
by arjf 3mo ago
Thanks - adding all of that to my to read / to do list :-)
4.
by arjf 3mo ago
Thanks - to be fair the electric piano seems to me like a relatively favorable case for this approach.
5.
by arjf 3mo ago
A tracking resonator bank should self-tune to any frequencies... so as long as the density of resonators is adequate, after convergence, it should paint a representative picture of the tone profile. Then you can try funny chords, or see how
6.
by arjf 3mo ago
I think the output of a tracking resonator bank is only the basis for higher level analysis that will produce results suitable for specific applications (see my comment on frequency component tracking and prediction/feedback).
7.
by arjf 3mo ago
I used the Exponentially Weighted Moving Average (aka low-pass filter) because it has a very nice iterative form and is very computationally efficient. My objective was low-latency for real-time systems (so no looking into the future either
8.
by arjf 3mo ago
My approach is to let the "bins" collide at the filter bank level (really let nearby tracking resonators agree on the dominant frequency in the neighborhood), and use the bank's instantaneous state as input for a frequency co
9.
by arjf 3mo ago
Sounds really interesting! Could you share some description of the algorithm used for chord detection? What model of tonality are you using for pitch/chord naming?
10.
by arjf 3mo ago
Spectral analysis has indeed been around as a concept for centuries and there have been apps based on the FFT for decades, so definitely nothing new there. What I have implemented however, while based in known concepts and techniques, allow
11.
Show HN: Resonate – Low-latency, high-resolution spectral analysis(alexandrefrancois.org)
46 points by arjf 3mo ago|31 comments
12.
by arjf 1y ago
Actually digging into SWIFT a bit more, the formulas differ by more than just the heuristic for alpha (unless I missed something) so the analysis in the SWIFT paper does not apply directly to(or maybe even at all).
13.
by arjf 1y ago
At least I might have introduced a bit of creative noise in some AI coding models :-P
14.
by arjf 1y ago
Awesome - The code (Swift or C++) in the Oscillators package is probably the best place to look for implementation details https://github.com/alexandrefrancois/Oscillators ping me if you have any question
15.
by arjf 1y ago
yes - the sample app has demo of single resonator (so frequency bin equivalent) frequency estimation/tracking based on phase shift and also Doppler velocity computation (the code for these is in the Swift package, equations in the upco
16.
by arjf 1y ago
Just want to call out the resources listed at the bottom of the Resonate website: - The Oscillators app demonstrates real-time linear, log and Mel scale spectrograms, as well as derived audio features such as chromagrams and MFCCs https:&
17.
by arjf 1y ago
The Sliding Windowed Infinite Fourier Transform (SWIFT) has very similar math, and they provide some analysis in the paper. I use a different heuristic for alpha so I am not sure the analysis transfers directly. In my upcoming paper I have
18.
by arjf 1y ago
It only requires more computation if you really need to compute the full FFT with all the bins, in which case the FFT is more efficient... With this approach you only compute the bins you really need, without having to pre-filter your signa
19.
by arjf 1y ago
Interesting - thanks for sharing!
20.
by arjf 1y ago
This formulation is close to that of the Sliding Windowed Infinite Fourier Transform (SWIFT), of which I became aware only yesterday. For me the main motivation developing Resonate was for interactive systems: very simple, no buffering, no
21.
by arjf 1y ago
Indeed... I honestly don't remember where or how I sourced the value, and why I did not use the "correct" one - I will correct in the next release of the package. Thanks for pointing it out!
22.
Show HN: Resonate – real-time high temporal resolution spectral analysis(alexandrefrancois.org)
76 points by arjf 1y ago|30 comments