NPR Quiz – can you hear the difference?

[Michael Canavan]

A minor nitpick: of all the compression artifacts that are audible in MP3 encoding, aliasing is not one of them.

Aliasing is a pretty major thing for a lot of algorithms we use in audio production, but when discussing MP3 encoding it's not.

OK well, dirt in the high end. It's very obvious with the 128bit versions, and to me anyway it sounds very similar to the warbley sound of aliasing in the high end of samples from years ago. I suppose artifacts is the correct term.

[Shooshie]

A minor nitpick: of all the compression artifacts that are audible in MP3 encoding, aliasing is not one of them.

Aliasing is a pretty major thing for a lot of algorithms we use in audio production, but when discussing MP3 encoding it's not.

It's been a long time since I heard low-quality mp3s, and I'm talking about things posted in the mid-1990s, but I'm pretty sure aliasing was a problem in those. When you hear high-pitch modulations like someone turning a radio dial in the movies, that's aliasing. It's usually caused by rate conversion during the compression process.

Examples: http://www.audiorecording.me/dithering- ... study.html

Shooshie

[philbrown]

I have a semi-related question. I did an electronic new age CD a few years ago. The nature of what I was trying to do called for a lot of whispy high frequency synth material. The 24 bit masters sound great (to me), but dropping to 16 bit caused some harshness in those high frequencies, not bad, but noticeable on direct comparison. Harshness might be a bit strong- maybe just "less smooth". As far as terminology- is that aliasing, or would you call that something else? I'm most likely mis-using the term aliasing when I blanketly use it to describe harshness in high frequencies.

[Robert Randolph]

I have a semi-related question. I did an electronic new age CD a few years ago. The nature of what I was trying to do called for a lot of whispy high frequency synth material. The 24 bit masters sound great (to me), but dropping to 16 bit caused some harshness in those high frequencies, not bad, but noticeable on direct comparison. Harshness might be a bit strong- maybe just "less smooth". As far as terminology- is that aliasing, or would you call that something else? I'm most likely mis-using the term aliasing when I blanketly use it to describe harshness in high frequencies.

There's no way to know what caused the issue, or what you heard, without more information. It could have been aliasing, or something else.

The only thing for sure, is that a simple bit-depth reduction would not cause aliasing. If there was a sample rate reduction, or use of some other process, then it could possibly be aliasing.