mathstodon.xyz is one of the many independent Mastodon servers you can use to participate in the fediverse.
A Mastodon instance for maths people. We have LaTeX rendering in the web interface!

Server stats:

2.7K
active users

#dsp

4 posts4 participants0 posts today
Continued thread

Okay #math and #dsp folks. Is there a noise function that I can initialize to generate a mode between to values?

Say I have a function f(x) which generates either 1 or 0 each time I call it. I call it N times, and sum the ones and zeros and then divide by N, I want the result to be "P" (which is a number 0 <= P <= 1)

A uniform distribution gets me 0.5, I want a distribution that centers around P.

Thoughts? Plan of attack? References?

In news surprising to no one, blood tests show I haven't been getting adequate nutrition.

#RaiseTheRate #DSP #Poverty #HealthCare

tenor.com/en-GB/view/protein-p

Alt text

A GIF of an extreme close-up shot of "Yoga Instructor" (ヨガインストラクター), a kangaroo character with as yet no name, in the Japanese and US animated series Aggretsuko. His eyes are glowing and sparkling as he says "Pro-teyyn".

One way to get more variety when transforming data into audio is to change the encoding. Today I'm implementing the VOX ADPCM telephone codec — which I especially like — in Rust to accomplish this! reillyspitzfaden.com/posts/2025/0... #DSP #Audio #Rust #Programming #Coding #ElectronicMusic

Databending Part 5 — Listening...

reillyspitzfaden.comDatabending Part 5 — Listening to Telephone Codecs

DSP Project Update Time 🥳

Couple of weeks back got some FFT code into it 👍
This let me do 'sample morphing' where by you gave it two chunks of audio, a bunch of FFT gets built for both, and over N "frames" it'll morph between then sounds smoothly.

Also added FFT based wave form generation; whereby I can just poke phase/magnitude values into FFT bins and get sounds out the other end, which is fun

Continued thread

Quelle Surprise! Department of Housing & Public Works is at it again.

They have put my public housing application on hold again (prob b/c I asked why the community group they say they referred me to a) didn't get a referral and b) never accept unrequited referrals).

It’s like playing Guess Who (is lying? Trick question, maybe both are).

Now again demanding the paperwork I had already given them in triplicate, both printed, scanned and in electronic form.

Continued thread

If people are interested, the code should now be at a point that others can use it (if you're comfortable using cargo to run the code): github.com/reillypascal/data2a

If people have suggestions on distributing a cross-platform CLI tool binary using Rust, I'd be interested — that and adding some different sample formats when importing (e.g., VOX ADPCM/NMS ADPCM are some of my next goals for this

Overall I'm enjoying Rust a lot, and making CLI tools seems to be a good way for me to have manageable but interesting projects

Convert any file to audio (often called “databending”) - reillypascal/data2audio
GitHubGitHub - reillypascal/data2audio: Convert any file to audio (often called “databending”)Convert any file to audio (often called “databending”) - reillypascal/data2audio
Continued thread

To top it off, QSTARS has made it clear that, despite my efforts, securing alternative accommodation is unlikely.

I feel invisible, hardly even a statistic, much less a person.

QLD Housing Centre continue to refuse assistance and gaslights me & others in desperate need. Lying to our faces.
 
So, the next time you hear a politician claim they’re helping individuals like us, remember—it’s often far from the truth.

Last night, took a random sample from a discord I'm on and ran it through my DSP project to see what it would sound like.

Opened the resulting wav in RX9 and saw the wav looked mad; huge DC offset in the samples 🤔

Turns out the original sample had a small DC offset (0.03%) and time stretch + grains amplified it to 17%!

So, anyway, I've now got some DC offset filtering code aka a simple first order high pass to take out super low frequencies (sub-20hz or so) - problem solved👍

Crowdsourcing #DSP creativity!

You could help @gnuradio 's block documentation wiki a lot if you could try out our instructions for how to add example flowgraphs to Block Doc pages!
wiki.gnuradio.org/index.php?ti

The documentation team (and countless contributors) have created a host of 555 block documentation pages to date. Not every single one of these has an example flowgraph that really illustrates the useful things you could do with that block.
Could you help?
#gnuradio #sdr #docs #docstodon

I've been learning about resampling and I've got some nice spectrograms. if you want to take a sample and play it back at a different pitch, there's a surprising amount of work involved to make it sound good

i'm doing a similar test as described here:
jeskola.net/xs1/content/test/
15khz tone pitched up through an octave and then down. some of the notes are above the nyquist frequency (half the sample rate) so shouldn't be there at all ideally.

first, zero-order hold i.e. skipping or repeating samples as necessary. unusable unless you're making chiptune.
second, linearly interpolating between samples, still very bad

Wavelet-Based Spectrum Analyzer! 🎶

FFT has long been the go-to method for visualizing audio spectra, but what if there’s a faster and more efficient alternative? Enter the Haar Wavelet Transform: a technique that provides logarithmic frequency resolution while being even more computationally efficient than FFT-based analysis.

Stay tuned for a deep dive into how wavelets can be used for real-time spectral analysis—no FFTs required!

Some preview: gist.github.com/ashafq/99d468d

#dsp#wavelets#audio