or Connect
Head-Fi.org › Forums › Equipment Forums › Sound Science › Sampling Rate Conversion Loss
New Posts  All Forums:Forum Nav:

Sampling Rate Conversion Loss

post #1 of 4
Thread Starter 

What happens when audio is converted to different sampling rates? For example, a 44.1 KHz song playing on a media player set to output 48 KHz (sampling rate of movies) to a DAC, or the opposite with a 48 KHz movie playing on a device outputting 44.1 KHz.

There should be a loss when converting between these, no? How can one tell if hardware converts without consent?

post #2 of 4

http://src.infinitewave.ca/

 

http://src.infinitewave.ca/help.html

 

Your question is rather broad.  There are artifacts to conversion.  Depending on how it is done it can vary from pretty awful, to nearly a non-issue. 

 

The above links show tests of many different conversion software.  The second link explains the results you are seeing so they can make sense.  

 

To know in any detail one would need to specify the equipment you are asking about.  Perhaps look at the links, narrow your question a bit more and someone can give some better answers.

 

This might help too.

 

http://wiki.benchmarkmedia.com/wiki/index.php/Sample-rate

post #3 of 4
Thread Starter 
Quote:

 

To know in any detail one would need to specify the equipment you are asking about.  Perhaps look at the links, narrow your question a bit more and someone can give some better answers.

 

I remember an old PC sound card which's DSP only worked with 48 KHz. I also believe the Xonar Phoebus exclusively uses 48 KHz when the 3D mode is activated. There could be more out there that doesn't support native 44.1 KHz, and thus knowing a way to identify these offenders would be useful. I'm especially interested in knowing the sampling rates of consoles and their games. 

post #4 of 4

Newer sound cards (including onboard HD audio codecs) usually have hardware support for 44.1 kHz sample rate, at least if you do not use DSP effects. Some may have slightly worse performance than at 48/96/192 kHz, though, for example, the ASUS Xonars have higher noise floor by 6-7 dB, perhaps because of the re-clocking by the CMI8788 chip. Notable examples of older hardware that was limited to 48 kHz include AC97 onboard audio, and various Sound Blaster Live and Audigy cards.

 

In the likely case that you are using Windows, note also that the Windows audio stack may still convert the sample rate in software anyway, sometimes even more than once. It is not uncommon to see this problem in recorded samples and RMAA test results posted by people (see the "24bit vs 16bit, the myth exploded!" thread for a possible recent example).

 

For more information on sample rate conversion, see also this older Sound Science thread.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Sampling Rate Conversion Loss