gbh444g

16 ms·

31.
by gbh444g 5y ago
I've been trying to find the "true and correct" visualization of music that would reflect its "inner symmetry". I started with an assumption that music is a much simpler entity than drawings, simply because it'
32.
by gbh444g 5y ago
Yeah, it was mostly mp3 and ogg files. If you're suggesting flac would be better, I don't think flac adds much clarity over a decent 320 kbps mp3.
33.
Analysing bird songs with Wigner transform(soundshader.github.io)
76 points by gbh444g 5y ago|22 comments
34.
by gbh444g 5y ago
Compressing that boilerplate into one sentence: "under pressure, contact-electrification occurs between its conductive and dielectric layers, creating an electric charge." A relevant off-topic: there is a living "meta materia
35.
Wigner Distribution Function for Dummies(soundshader.github.io)
2 points by gbh444g 5y ago|0 comments
36.
by gbh444g 5y ago
I wanted to add that the nature of those thin horizontal layers and the patterns they form is similar to diffraction. A laser beam passing thru a thin slit forms a pattern with the sinc-wave distribution of intensity, and so is FFT of a rec
37.
by gbh444g 5y ago
Applying the Hann window function eliminates all the spectral leakage, but it also makes the image rather dull and precise, very similar to CWT. You've made me realise that the intricate patterns seen on violin spectragrams are the res
38.
by gbh444g 5y ago
Larger FFT window has undesired side effects because the estimated frequencies are averaged over the entire window. Moreover, the FFT output always spans from 0 Hz to 24 kHz (with 48 kHz sample rate), so to zoom into the 0..6 kHz region we&
39.
by gbh444g 5y ago
I did experiment with CWT in past [1] and was disappointed, to be honest. Not only it's grossly slow and complicated, it hardly gives more fidelity than plain FFT, it has the "window problem" which makes the low freqs too blu
40.
by gbh444g 5y ago
This sounds about right. I assume your (n+1)/2 is really n+1/2. The idea, like you've said, is to get Y[k+1/2] values where Y = FFT[X]. Whether this is mathematically sound is another question. I presume that it is, for
41.
by gbh444g 5y ago
I was thinking this is the case, until I stumbled upon a stackoverflow question that explains how to recover the phase data from overlapping FFT frames. The key word here is "overlapping".
42.
by gbh444g 5y ago
Hello HN! Author here. I was thinking to call the post "The underappreciated complexity of musical sounds" but decided to stick with the DFT one as it would probably get more attention. This is a small discovery I came across this
43.
Highres Spectrograms with the DFT Shift Theorem(soundshader.github.io)
124 points by gbh444g 5y ago|69 comments
44.
by gbh444g 5y ago
Feature request: if you replace the "techno" sounds with piano notes, you'll get a potentially interesting melody.
45.
by gbh444g 6y ago
I think the argument here is that without e2ee, the chats are available to whoever pays enough (e.g. advertisers). So in return he will get some ads and sms spam.
46.
by gbh444g 6y ago
Is it a regexp? I'm thinking if we could come up with a especially dense writing style that needs quadratic time to fully understand a sentence. Like you read a sentence of N words and get only a surface level meaning. Then you re-read
47.
by gbh444g 6y ago
No. But with enough effort and reasonable assumptions, you can restore most of the sound. It would be enough for speech, but not enough for high quality music.
48.
by gbh444g 6y ago
Mr. Polgar would know he's wrong if he tried to make just 100 geniuses out of average children. Most people are unable to hold their attention on the same object for longer than a minute.
49.
by gbh444g 6y ago
The process looks similar to a CNC machine replicating a predefined shape. Except that here the material is able to grow itself. In this case it's not surprising that a CNC machine always creates about the same shape from a rather amor
50.
by gbh444g 6y ago
Watching the initial stage where this amorphous blob of cells just kept multiplying, made me think that the blob of cells is just that - the bioprinter material. The shape is probably set by the printer, i.e. the parent body, and that blob
51.
by gbh444g 6y ago
I have a deep suspicion that this is mostly about risk management. With just 50 employees, what you have is essentially a board of directors that have enormous leverage over you.
52.
Ask HN: Personalised HN Feed?
1 points by gbh444g 6y ago|0 comments
53.
by gbh444g 6y ago
Does Brave support uMatrix and uBlockOrigin?
54.
by gbh444g 6y ago
We usually call this an RCE exploit.
55.
by gbh444g 6y ago
Or maybe he's another Edison. I mean, he's likely the nominal inventor: other people do the research, but only his name appears on papers.
56.
by gbh444g 6y ago
FANG pays 450-550 to someone with 10 yoe. GS-15, on the other hand, requires a PhD and even if it's reachable within 10 years, pays only 180.
57.
by gbh444g 6y ago
There's also the "executive schedule" [1] for management, but it doesn't pay much better: $220K for the highest paid ES-1, and a PhD is a requirement. In fact, the president of the US makes only $400K. [1] https:/&
58.
by gbh444g 6y ago
So a computer is just an electrical current.
59.
by gbh444g 6y ago
In addition to that, Stephen Wolfram hypothesized that QC won't be possible because with the growth of QC state, the energy required for error correction will grow just as fast, so the theoretical exponential speedup will require expon
60.
by gbh444g 6y ago
I'm a pragmatic person. If QC mental model is so complex that I can't learn it in a few weeks, then very few in the world can and QC will remain a subject of scientific papers understood by a few.

More ›