2013/02/26 21:19:46
sharke
Interesting article about the non-linear nature of our hearing and what this means for audio compression:

http://arstechnica.com/sc...akes-mp3s-sound-worse/
2013/02/26 21:40:59
cclarry
Interesting article...
2013/02/26 22:12:48
craigb
I'd use wave files, but they take too long to create in Notepad...
2013/02/27 03:10:30
FastBikerBoy
I can confirm that too. Every time I play someone one of my songs which are always in mp3 format they say "That's crap". Not very scientific but proof none the less.
2013/02/27 03:12:19
bapu
craigb


I'd use wave files, but they take too long to create in Notepad...

But Notepad has no bugs, roight?


2013/02/27 04:33:20
craigb
bapu


craigb


I'd use wave files, but they take too long to create in Notepad...

But Notepad has no bugs, roight?


Ya, works great!  At least if you avoid the newest "X" versions that is.
2013/02/27 10:19:29
Moshkiae
Hi,

Hahaha!!! Love it!

I wanna add to that article!!!!

There is also a difference in each person that makes this equation even tougher ... and I BELIEVE that it also accounts for a person's tastes and ability to listen to more, and different musics, where many of its details are totally different and less linear than the majority of music that we spend our time listening to it.

I kinda wonder, and have thought, that this is the area where "conditioning" works best ... you get used to hearing things like this, and eventually you do not think, or believe that anything else is possible, thus anything else is considered bad or not correct.

I like to joke that this is the socialist style of thinking, where individuality is not allowed!
2013/02/27 12:20:36
bitflipper
So imagine now that a new codec replaces MP3 as the most-ubiquitous online format. Now fast-forward 20 years...MP3 hardware players are no longer manufactured, and the old software players won't work on Windows 19. Your massive collection of MP3s sits silently on the DVDs you backed them up to back in 2017 - but there are no more DVD players, either. 
2013/02/27 12:25:57
soens
.
2013/02/27 12:27:43
drewfx1
Unfortunately it's misleading and misinterpreted.

The recently published underlying paper has been very popular with audiophools and other people who don't really understand it because things tend to be written about it in such a way that they seem to imply that humans beat math/physics. The truth is that what has been demonstrated is that humans do in fact beat a particular limitation involving a type of linear processing -  but this was something that was basically known but apparently unconfirmed.

The basic idea is that there is a limit (known as the Gabor limit) in time vs. frequency resolution when performing a linear process (an FFT) and that it has now been demonstrated that humans can perform better than this limit.

But the idea that human brain does non-linear processing has been known/assumed for a loooooong time by people who work with human audio perception. Like say, for instance, people who do things like perceptual audio coding for things like mp3's.

Humans just don't use really a linear process like an FFT to perceive things. This isn't a new idea; it just (apparently) hasn't been scientifically confirmed before in this specific case.

But none of this matters to the listener, because though FFT's are used in audio processing, the limit doesn't adversely affect the audio itself - it's just puts a limit on how you can process it using linear math. You can either optimize for frequency resolution or time resolution, but not both at the same time.


What is happening here is that people who don't really have any background in or understanding of any of this stuff are starting from an interesting confirmation of what was long suspected about human audio perception being non-linear and jumping from that to ridiculous conclusions about non-existent impacts on real world audio.

Carry on.
© 2026 APG vNext Commercial Version 5.1

Use My Existing Forum Account

Use My Social Media Account