Oh and the code (now with the VOX ADPCM) is here!

https://github.com/reillypascal/data2audio

#Databending #Rust #Programming #Coding #Telecom #Telecommunications #ADPCM

There's definitely going to be a blog post about this, with some more backstory on ADPCM and the VOX format

I've been reading windytan's blog (https://www.windytan.com/2012/11/the-sound-of-dialup-pictured.html) lately and really enjoying the deep dives into how specific audio things work, and I think I want to do some of that.

#Databending #Rust #Programming #Coding #Telecom #Telecommunications #ADPCM

I found the original VOX ADPCM specification (https://multimedia.cx/mirror/dialogic-adpcm.pdf), and when that didn't sound quite like Audacity's implementation, I looked at the FFMPEG source (https://ffmpeg.org/doxygen/7.0/adpcm_8c_source.html#l00553)

For some reason, FFMPEG seems to do it a bit differently than the specification. I'll have to look into that more and see what's going on, but for the time being, it works!

#Databending #Rust #Programming #Coding #Telecom #Telecommunications #ADPCM

I implemented my own VOX ADPCM decoder!

I had mentioned (https://reillyspitzfaden.com/posts/2025/05/databending-part-4/#adpcm) wanting to use that format to interpret raw data as audio, and the Rust Symphonia crate (https://crates.io/crates/symphonia/0.3.0) didn't have the VOX variant and assumed the incoming data was properly formatted as an ADPCM file, rather than random data, so it was actually easier to do myself.

Here's how it sounds:

#Databending #Rust #Programming #Coding #Telecom #Telecommunications #ADPCM

@silsinn9821 According to this documentation it's a YMF288/OPN3. I don't think the first five #touhou games ever used custom #ADPCM which is the most notable removal of this cheaper #OPNA imo, so probably should be fine for #PC98Touhou I guess? That is if Wikipedia is correct that the RSS and custom ADPCM are separate (even if both are using ADPCM) ​:sagume_think:​

@Pawlicker

"Only a Synth deals in absolutes....."

My "nemesis", as I refer to them, said that to me once.
I disagreed at the time, citing the persistent fluctuations that exist with analogue synthesis systems, but we held accord on that there was argument to be had for digital (DSP) synthesis methods and absolutes.
(a copy log of that particular conversation can be found HERE)
But how much information is really needed for a workable "absolute"?

https://joindiaspora.com/posts/fb017730d2fb0135bdf10242ac110007 https://friendica.feneas.org/display/76158462-425f-a8aa-c5ad-5fc554362525

When document lies:

https://github.com/NeoChen1024/IMA_ADPCM/blob/master/adpcm.c#L73

#ADPCM #C

正在學習 IMA ADPCM 演算法的我： :fishthink:
#ADPCM