INTERACT FORUM

Please login or register.

Login with username, password and session length
Advanced search  
Pages: [1]   Go Down

Author Topic: 24 bit operation  (Read 825 times)

brookb

  • Regular Member
  • Recent member
  • *
  • Posts: 31
  • Yep, I'm a music loving geek.  :-)
24 bit operation
« on: September 28, 2002, 11:08:08 pm »

Hi,

I've been a MJ user for a while now.  I've noticed this problem for a while, and its never been fixed.  I've avoided it by using 16 bit resolution on a 24 bit sound card.  I'd really like to use the 24 bit resolution of my 24 bit card, and I'd like some feedback from MJ.

I'm using a LynxOne audio card - a professional 24 bit/ 44.1,44,96ksample unit with analog and digital I/O.  This is a great sounding card, and I really want to use it with MJ, but there are a couple of issues:

1) When I select 24 bit playback in MJ, the music plays back fine, except that cross-fade does not work.  Typically, the last few seconds of a song are simply not played, and the next song abruptly starts.  However, if I switch MJ to use 16 bit output, everything works fine.  Unfortunately, 16 bit output just doesn't sound quite as good.

2) I'm wondering how the DSP functions work with 24 bit audio.  For example, MP3 songs are probably quantized to 16 bit output, with each sample representing one of 65536 different levels.  However, if I use Replay Gain, I lose a lot of that resolution.  For example, a 3dB attenuation would effectively reduce output to 15 bit resolution (32768 levels).  Other reductions that are not exact multiples of 2 will result in not only loss of total bits, but potential resolution loss.  So one question is: does MJ dither 16 bit output when the DSP alters gain and EQ?  If not, a noticable reduction in quality will result.   Another question is: does MJ convert MP3s to 16 bit wav when 24 bit is enabled, or does it convert natively to 24 bit?

The flip side of this, though is that when using a 24 bit audio card with the DSP on, you can eliminate the need to dither, and you can use the extra resolution of the 24 bit card to find appropriate levels between each of the 65536 previous 16 bit levels.  So the question here is: does MJ use 24 bit (or higher) resolution math when doing DSP to produce a 24 bit output after performing DSP on 16 bit input?  Does it correctly interpolate using full 24 bit resolution on DSP functions?   If so, you can expect much better sound on 24 bit cards when using DSP.  (except for problem #1)

Logged
Pages: [1]   Go Up