[vscratch]

In the audio settings for a lot of games I get a few things which, while I do have a very basic understanding of what some of them mean, don't know if I should change from default or not.

I'm talking about "buffer lenght"/"Latency", "Sample Rate" and "number of sound channels". Is there an 'optimal' setting for each one? Does it depend on the audio source or sound card?

EDIT: Also "Bit Depth" and "Replay Gain" from my media player settings.

I am googling these terms and learning about them, but it rarely gives me a good idea of what values to give them in practice.

 

[vscratch]

bump

 

[NA Blur]

Buffer Lenght:  How much ahead of time processing your computer will do to the audio.  Think of this as the buffer when watching youtube videos.  The more it buffers the higher quality in general because it has it loaded and does not have to do continuous signal processing live.  The longer the buffer usually the better, but your PC must have enough memory to handle the more audio stored in the buffer.
 
Sample Rate:  Just like it sounds the sample rate of a system or piece of audio is how many samples per second it was recorded / played at.  The higher the rate the more samples.  The more samples the higher the quality as the signal forms more closely to the actual signal.  In addition the more samples the more storage the audio file will take up.  Most CDs are recorded at 44.1 kHz ( note this is hz not kbps ).  This means they play and are typically recorded at a sample rate of 44.1 thousand samples per second.  An mp3 might be as low as 128 kbps.  This is 128 thousand bits per second.  A CD for example has 1,411 kbps as the sample/bit rate or about x4 times the sample rate of an MP3.
 
Sample rate and bit rate are not synonyms, but both can be thought of as signal processing rates.
For more on this see:
http://thestereobus.com/2008/01/12/sample-rate-and-bitrate-the-guts-of-digital-audio/
 
Bit Depth:  This is the number if bits used to record /playback a signal.  The higher the bit depth the more sampling occurs during each piece of the signal.  You can imagine this as how many parts of the sine wave are being captured.  A low bit depth may only give you the peaks of a signal whereas a high bit depth ( 16 bit ) samples many parts of the wave capturing the peak, trough, and a few points in between.  As you may have guessed this will directly correlate to the noise in a system.  If only the peaks of a wave are captured there is little room for the system to see a zero volume area on the curve.  Think of bit depth as a direct control of the signal to noise ratio or how well the system can reject / detect noise.  Not all devices can support higher bit rates.  For example many audio devices are only 16-bit devices and telling your PC to run at 24-bit depth may introduce noise because the PC may try to sample more of the wave then the equipment is capable of. 
http://en.wikipedia.org/wiki/Audio_bit_depth