INTERACT FORUM

Please login or register.

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

Author Topic: function cost in processor usage  (Read 1199 times)

LisaRCT

  • Guest
function cost in processor usage
« on: October 31, 2003, 01:05:03 pm »

It appears that much of the multimedia functions take a fair amount of processor power to work.  This can, in spite of today's more powerful processors, press the limits to the point of some functionality deteriorating somewhat.

This brings to mind questions regarding the 'processor cost' vs the benefit that you may have gained.

For example . . .  from the performance I see on my PC (2GB AMD-XP, 1Gb DDR, Win2K), I am of the impression that ASIO consumes a bit more processor than Direct Sound or WaveOut.
Add on top of that playing APE files and their need for processor to decompress for play, the use of direct-X plug-in's, running Media Server, and using MC's 'Output Format'  feature to send 24/96 audio to my Terratec audio card. . . .  and you apparently have a decrease in  system performance.  Nevermind adding in visualizations or even w00t.

SO, I can't help but wonder . . .
Is ASIO worth it?
Is there really any gain in audio quality?

Is running MediaServer and more taxing than running off a mapped drive?

Is running 'Output Format' at all worthwhile?  
I can't create what is NOT there, so is there any gain?
Is the gain worth the costs?
Etc . . . . .
Etc . . .
Etc.

Does anyone have any test results or evidence as to the costs in processor and memory load for various functions?

How about ways to achieve a comparable quality using less resources?

This is a question I have seen discussed regarding disk space vs audio format vs sound quality . .. .  .  but I have not seen any discussion of processor resources.

Or is it all just negligable and matters not?
Logged

JimH

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 72439
  • Where did I put my teeth?
Re:function cost in processor usage
« Reply #1 on: November 02, 2003, 07:47:46 am »

I think that APE decoding takes less CPU than MP3 decoding.

Media Server is a very light load.  I can run several streaming connections at once from one server, all using APE.

I believe that ASIO should take less power.

Visualizations can be a load.  Some more than others.

Have you used the task manager to monitor what's going on?  Other applications may be a factor.
Logged

LisaRCT

  • Guest
Re:function cost in processor usage
« Reply #2 on: November 02, 2003, 11:31:45 am »

Actually Jim, I am not trying to resolve any problem, but rather just seeking some guidelines.  Was kinda hoping some folks had come across some research or side-by-side test results in their travels
Logged

marketability

  • Regular Member
  • World Citizen
  • ***
  • Posts: 177
  • eat!
Re:function cost in processor usage
« Reply #3 on: November 02, 2003, 11:36:10 am »

I've done some tests

http://www.satbythesea.com/vacard_cputest.pdf

My findings for what they are worth...
Playback of MP3 is less CPU intensive than APE (I use max compression APE so that may explain why)
ASIO drivers use less CPU than DirectSound

I struggle to imagine anything better in quality than ASIO as it sends the same signal as you would get between a CD transport and a DAC (I think). It certainly avoids Microsoft software (like mixers, DSP, etc) getting in the way of the audio signal

I use APE > ASIO > HiFi DAC or Lossless WMA > ASIO > DAC

APE is a big CPU overhead if I use my virtual/software soundcards (which converts to MP3 on the fly) in order to stream the output to networked audio

 :-\I struggle to tell the diffrerence between my HiFi DAC (£2,700) and my soundcard DAC (M-Audio 24-96 - £150) !!!!! sad isn't it - boys toys and besides the DAC has lights and buttons!

I can definately tell a significant difference in quality with APE over MP3 and so for me the CPU overhead is worth it

As a HiFi geek the less in the way of the signal the better - so ASIO and an external DAC lights my candle

Hope it helps!

CPU is pretty important - I struggle to run 3 zones playing back APE on a PIV 2.4GHz with 1.5Gb RAM!
Logged

JimH

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 72439
  • Where did I put my teeth?
Re:function cost in processor usage
« Reply #4 on: November 02, 2003, 02:22:50 pm »

I've done some tests

http://www.satbythesea.com/vacard_cputest.pdf

My findings for what they are worth...
Playback of MP3 is less CPU intensive than APE (I use max compression APE so that may explain why)
Normal compression is preferred.  Max gives just a few percent more savings and it's probably not worth the time it takes.  Both are perfect quality.
Logged

marketability

  • Regular Member
  • World Citizen
  • ***
  • Posts: 177
  • eat!
Re:function cost in processor usage
« Reply #5 on: November 02, 2003, 02:40:02 pm »

I agree Jim - shame I only noticed after ripping 20,000+ files at Max!

I keep meaning to do some more comparative tests to help people on deciding the best format for them - maybe one wet weekend...
Logged

LisaRCT

  • Guest
Re:function cost in processor usage
« Reply #6 on: November 02, 2003, 10:45:20 pm »

Wonderful graph Marketability, thank you.
This is the kind of dialog I was hoping to generate.
I use APE on High compression (not highest) and was concerned that as I try to put my old AMD 550 K6 into service as a server (until something better comes along :D  )
The 24/96 card does sound awesome thru ASIO
Logged
Pages: [1]   Go Up