Author Topic: Guan  (Read 10655 times)

32 Bit Floating-Point?

  • Guest
Guan
« on: 26 Jun '03 - 23:53 »
Does XMPlay have 32 Bit Floating-Point?

Topic Floating-Point lol

  • Guest
Re: Guan
« Reply #1 on: 26 Jun '03 - 23:58 »
Sorry for the Topic=Guan hehe it should have been Topic=Floating-Point  :laugh:

Anyway i mean Floating-Point in general not just 32 bit.

Jace

  • Posts: 842
Re: Floating-Point
« Reply #2 on: 27 Jun '03 - 00:56 »
Quote
Sorry for the Topic=Guan hehe it should have been Topic=Floating-Point  :laugh:

At least you could have corrected it on the second go ;D

Olego

  • Posts: 557
He he he
« Reply #3 on: 27 Jun '03 - 04:04 »
:laugh:

I'm pretty sure I read that XMPlay uses 64 bit integers, and the quote from BASS is "32 bit floating-point decoding and processing".  So I guess XMPlay also uses 32 bit floats, unless that quote from the BASS site is outdated...

~Olego~
« Last Edit: 27 Jun '03 - 04:09 by Olego »

Ian @ un4seen

  • Administrator
  • Posts: 26102
Re: Guan
« Reply #4 on: 27 Jun '03 - 16:00 »
All internal decoding/processing is in 32-bit floating-point, and it's only converted to the user-chosen output format at the end... I'm not sure if that's exactly what you were asking, but there you go anyway :D

Guan

  • Guest
Re: Guan
« Reply #5 on: 3 Jul '03 - 14:05 »
Does 64Bit floating point improve sound quality?

Ralesk

  • Posts: 652
Re: Guan
« Reply #6 on: 4 Jul '03 - 03:34 »
You have to have an alien parasite sitting in your ear to notice the difference between 24 and 32, already.

Ian @ un4seen

  • Administrator
  • Posts: 26102
Re: Guan
« Reply #7 on: 4 Jul '03 - 14:25 »
:laugh:

Internally, the FPU does perform the calculations in 64-bit, but the results are stored by XMPlay in 32-bit... storing 64-bit results would be major overkill - the differences are so tiny that there's no way anyone can possibly hear any benefit (especially considering soundcards don't go above 24-bit :)), and is not worth the performance hit it would incur.