Press "Enter" to skip to content

The "Simple Math" Myth of Sample Rate Conversion

I just saw this on a forum. I’ve seen something like this said many times. I won’t say who said it because he’s a good engineer and my intention is not to poke at him.

“I’ve been sticking with 88.2, to me the simple math (divide by 2) is a little more reassuring then 96 when down-sampling to 44.1.”

*facepalm*

Another guy (elsewhere) said:

“If you are recording music you should stick to 44.1k which is cd redbook standard. Recording at 48k and sample rate converting to 44.1 later only introduces artifacts and degrades the sound quality. Save 48k for when doing sound for video.”

*double facepalm*

Honestly, that makes no sense. Please people, get this “Simple Math” MYTH out of your head!!

You don’t have to fully understand (or care) how sample rate conversion works, but you do need to realize that really smart people have already figured it out and MATH is not a difficult task for a F*cking COMPUTER!!!

The quality of conversion is what matters most, and it’s pretty damn important to capture as much as possible from the start. Not all converters are created equally well, and some will sound better operating at a certain sample rate.

@RandyCoppinger and I recently chatted about how to accurately test which sample rate our converters sound best at. We haven’t done the test yet but once our goggles and lab coats come in we probably will. Currently I use 44.1, and he uses 48k. Frankly I’m not yet working on anything good enough to spend twice the disk space and processing power on.

What sample rate do YOU prefer and why?


3 Comments

  1. Petri Suhonen
    Petri Suhonen October 3, 2010

    I’m with 44.1 because my ears are completely happy with it 🙂

  2. Giovanni Rozzi
    Giovanni Rozzi October 4, 2010

    I prefer 48kHz just for that extra high end on most of my recordings (you can always remove it but you can’t add it if it wasn’t recorded). But besides that little bit of difference between 44.1kHz and 48kHz it still comes down to how you recorded it at the source (granted your sampling clock isn’t dieing) and strength of the song.

  3. Marcos Ficarelli
    Marcos Ficarelli November 18, 2010

    I use 44.1, but not to spare my disk. Instead, I record with 24-bit (and would with 32 if I had an interface). Later, processed tracks will be in 32-bit and also mixes. I didn’t try 64-bit music, but I should try.
    There is no difference between 44.1 and 48. I really found difference only working with 192, and this works mainly for live recording of singers, piano or acoustic guitar (maybe contrabass).
    I believe that translating 48 to 44.1 will create artifacts, but not big enough to bother in a big mix.
    Final: “It’s the Bit”

Leave a Reply

Your email address will not be published. Required fields are marked *