ASIO4All stopped me from getting rid of Chaintech AV710
Jan 15, 2005 at 5:51 AM Thread Starter Post #1 of 7

wali

1000+ Head-Fier
Joined
Nov 22, 2003
Posts
1,100
Likes
13
Out of curiosity I purchased a chintech AV710. I received it a couple days back and went to install/set it up according to ‘graphic guide…’ thread. Everything went smoothly and I did not encounter any trouble with Audigy2 also installed (w/ all it's bloatware) in the same computer. Then after many hours of listening some issues become apparent:

-Low volume, forcing me to turn the amp all the way up resulting in very low hiss
-Incoherent representation—the music sounded as if it is being reproduced with a lot of effort
-Subtle fluctuations in music, as if the output signal is not consistent
-I also felt a strong “cognitive dissonance”, basically a feeling that something is wrong with the audio reproduction of this soundcard


I was about to give up on AV710, but after reading a few posts I installed ASIO4ALL and set ASIO as output in foobar, removing the resampler. surprisingly the situation improved instantly. Not only with ASIO I don’t have to enable the resampler in foobar (no CPU overhead) at the same time the system run much more stable. The volume situation got fixed and most importantly the soundfloor dropped considerably (impressive for a computer based source).

So, what do I think of AV710 now? Well, I can simply say that $20 had never sounded so good. AV10, when setup properly, produces a balanced output, reasonably detailed presentation combined with a decent soundstage. Most of the music which I listened to using this card sounded pretty good; however, when it came to Jazz this card just couldn’t reproduce the bass with full authority and also it lacked the full dynamic range of bringing to life a jazz session and all of its instruments.

I would strongly recommend that ASIO output should be used with this card instead of KS ( but I don’t have any technical reason, excpt subjective listening experience). Not only forgoing the resampler is an attractive choice at the same time—in my case at least, the ASIO route meant the difference between discarding this card and being impressed by it.
 
Jan 15, 2005 at 7:09 AM Post #3 of 7
ASIO4All IS NOT ASIO.. it's just a Kernel Streaming wrapper, so you don't gain anything going with it.. disabling resampler, you could have done that even without ASIO4All..
 
Jan 15, 2005 at 8:21 PM Post #4 of 7
Quote:

Originally Posted by Glassman
ASIO4All IS NOT ASIO.. it's just a Kernel Streaming wrapper, so you don't gain anything going with it.. disabling resampler, you could have done that even without ASIO4All..


Wrong!

From ASIO4ALL web-page:

Universal ASIO Driver For WDM Audio

ASIO4ALL is an ASIO driver which allows ASIO output with foobar to function. KS only works with resampler--in my case at least.

/Lay off the caps next time, it is annoying and makes you look idiotic.
 
Jan 16, 2005 at 7:36 AM Post #5 of 7
depends on how you set it up, in case of WDM it's just like using WaveOut plugin in foobar.. resampling is done by kmixer.. seems like ASIO4All now does resampling, but he's talking about 44.1 exclusively..

sorry for the caps in previous post, wasn't a good day for me then..
 
Jan 16, 2005 at 7:45 AM Post #6 of 7
Quote:

Originally Posted by wali
Universal ASIO Driver For WDM Audio


Then it's not really ASIO. ASIO and WDM are two different things. WDM is like WAVout and Direct Sound, whereas ASIO is a professional playback output.

Also, lay off Glassman, he is very knowledgeable in this stuff. I have learned much from him. I am pretty sure ASIO4ALL, in the way you described it, is not really true ASIO. ASIO and WDM is two differnt ways of outputting music.
 
Jan 16, 2005 at 1:14 PM Post #7 of 7
ASIO4All is an ASIO to WDM (kernel streaming) wrapper that allows using ASIO with cards that only have WDM drivers. Achievable latencies are quite close to those of native ASIO implementations, and the output is usually just as bit perfect, however the additional overhead leads to noticeably higher CPU usage with high sample rates, particularly if the card doesn't provide any hardware acceleration and resampling is used with a slower CPU. (This is not the first time I observed increased CPU usage on a normally low-usage process with another process drawing a lot of CPU time. I guess memory access is normally optimized via MMX/SSE but this no longer works well if the process gets starved on memory bandwidth.)
 

Users who are viewing this thread

Back
Top