Digital audio articles I wrote

[zepplock]

I was reading a lot of Facebook posts from people completely lost in the digital sound world. So I wrote two articles explaining digital audio myths and typical issues.
The articles are written with non-scientific people in mind since I failed to explain even simple things (for example Nyquist freq, etc). Instead, I'm using a lot of dead-simple analogies. This is basically "Explain Like I'm 5" of digital audio. The goal is to stop people from spreading snake-oil fud around.

Please provide feedback.

Part 1: http://vova.org/articles/digital-audio-myths-part1
Part 2: http://vova.org/articles/digital-audio-error-correction-part2

 

[pozz]

These are hilarious. Well done.

 

[Martin]

Your analogies do not display correctly on an iPhone unfortunately making the articles unreadable.

 

[zepplock]

Your analogies do not display correctly on an iPhone unfortunately making the articles unreadable.

I'll check it, thanks.

 

[zepplock]

I'll check it, thanks.

Fixed.

 

[zepplock]

I'm thinking about doing a piece on cables next.

 

[Vincent Kars]

We have a shoebox full of cotton balls. They took it and put it into ~3x larger container - say a 5 lb bucket. Same amount of cotton, in a different container.

Nope, highres has an increased dynamic range and increased frequency range. You do have more cotton balls hence need a larger box.

it gets error rejected or error corrected by Ethernet, TCP/IP and other network protocols.

Nope, USB audio (isochronous mode) allows for error detection but not correction as it is a quasi real time stream.
Streaming over Ethernet or WiFi is often done using UDP instead of TCP/IP. Again error detection is possible but no retry hence no correction.

Most of the time reported differences (audiophile switches, ethernet cables, etc) are a matter of listening with one's eyes instead of one's ears.
Maybe you can conjectured up a nice analogy to demonstrate the difference.

 

[zepplock]

Vincent, if you take 16/44 audio file and upscale to 24/96 - you don't get more dynamic range.

 

[Vincent Kars]

Correct
I overlooked that you were talking about upsampling hence means "fake" highres and not about "true" highres.
Sorry

 

[zepplock]

Vincent, as for UDP vs TCP, most streaming services, Roon, etc are all using TCP these days.

 

[mansr]

Streaming over Ethernet or WiFi is often done using UDP instead of TCP/IP.

UDP is only used for "live" radio/TV streams and telephony applications requiring low latency. Everything else uses TCP. Spotify, Tidal, Netflix, Youtube, and so on all use TCP. In-home streaming using UPnP/DLNA or Roon is also TCP.

 

[Vincent Kars]

My knowledge get's a bit dated

 

[LightninBoy]

I love the cotton ball in a shoe box vs altoid box analogy. Great way to describe lossless. I'd expand on that further and contrast it against what would happen with the cotton balls with lossy compression.

However, the #1 most important thing to establish is that digital music is just *data*. And that storing, retrieving, and transmitting data with near 100% accuracy is a solved problem. Most of our modern society is built on that fact. This foundation can then be used to debunk many myths around cables, green markers, etc. Once folks really understand this, most of the other malarkey becomes self evident.

#2 - its important to know that when errors do occur in audio data transmission (say because of a faulty USB cable for example), the resulting audio distortions are not subtle and typical manifest as pops and static, skips, etc. Any "veils", "air", or "pace, rhythm and timing (PRAT)" is embedded in the data itself, so as long as the technical components are storing, retrieving, and transmitting the data without error (which almost all do all the time, see #1), they are delivering all the possible air, PRAT, bass impact, or any other audiophile buzzwords.

 

[zepplock]

Good points, I'll highlight these in the article. Maybe I'll also link back to this thread for people to provide feedback.

 

[zepplock]

I love the cotton ball in a shoe box vs altoid box analogy. Great way to describe lossless. I'd expand on that further and contrast it against what would happen with the cotton balls with lossy compression.

However, the #1 most important thing to establish is that digital music is just *data*. And that storing, retrieving, and transmitting data with near 100% accuracy is a solved problem. Most of our modern society is built on that fact. This foundation can then be used to debunk many myths around cables, green markers, etc. Once folks really understand this, most of the other malarkey becomes self evident.

#2 - its important to know that when errors do occur in audio data transmission (say because of a faulty USB cable for example), the resulting audio distortions are not subtle and typical manifest as pops and static, skips, etc. Any "veils", "air", or "pace, rhythm and timing (PRAT)" is embedded in the data itself, so as long as the technical components are storing, retrieving, and transmitting the data without error (which almost all do all the time, see #1), they are delivering all the possible air, PRAT, bass impact, or any other audiophile buzzwords.

Updated, Thanks!