Re: KDE's farting error sound
From: Arthur Hagen (art_at_broomstick.com)
Date: 06/10/04
- Next message: Arthur Hagen: "Re: Recording from line input on SoundBlaster Audigy - Solved!"
- Previous message: Dr. Popper: "RAID MIRROR SW, PCSERVER AND BOOTING"
- In reply to: baskitcaise: "Re: KDE's farting error sound"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Thu, 10 Jun 2004 07:49:06 -0400
baskitcaise <baskitcaise@hotmail.com> wrote:
> Arthur Hagen adjusted his tin foil beanie and asbestos underwear to
> write:
>
>> While adjusting the buffers and PCI latency can help, it also causes
>> sound cut-offs, especially with the SBLive! 5.1, which really likes
>> having the bus
>> more than it should. It also doesn't seem to affect KDE, which
>> continues to play alerts a long time (1/4 seconds or more) after the
>> fact. It's especially noticable if I press backspace on an empty
>> console line. The sound arrives much much later than it should.
>>
>> Regards,
>
> Hi Arthur,
>
> Mmmm... I don`t suppose the SB/live card is on a Via chipset board?
Nope. Intel 815P (Asus TUSL2-C).
> I seem to remember the last Via board ( most Via boards used to show
> this fault ) I used would show this type of error ( in Windows as
> well ) unless I set the latency in the bios to 32 for the PCI bus and
> moved the card to different slots to find one that it liked.
Yeah, I used to have a Via board (Asus P3V4X) before this, and the problem
is that the SBLive! wants to hog the bus for longer than the other cards,
while the Via chipset doesn't have anything extra to go on and seems happy
to give the cards the extra cycles they want even when it means that other
cards won't get served at all. Thus, when the SBLive! grabs 64 cycles
instead of the default 32, you need to compensate by lowering the overall
latency by 32/[number of other PCI cards].
Incidentally, the same problem occurs with a couple of software ATA raid
cards, which change their PCI latency to 64. The combination of one of
these and a SBLive! on a Via-based system can be deadly.
Not the problem here, though.
> Now on Nforce and the delay you indicate in the console does not
> appear here, but as you say I think the quality of the sound has gone
> down since I have been using the 2.6 kernels, I have adjusted the PCI
> latency for the SB/live to 32 in powertweak and messed with the buffer
> size which does appear to change the quality a little, however I do
> not think I have hit the "sweetspot" yet.
>
> Also bear in mind I am an "Old Fart" and the ears have taken a bashing
> so am not the best of judges when it come to quality, what I might
> think is OK might sound terrible to a proper audio enthusiast :)
The SBLive! 5.1 sounds terrible to a proper audio enthusiast, period! It's
the forced resampling to 48kHz that is the main culprit, especially for CD
playing. The newest Audigy isn't much better -- although it can do lots of
different sampling frequencies up to 192kHz, that only applies when using
the input/output directly and not processing the sound in any way. If you
play a CD and do anything to the sound (including the volume control) except
feeding it out, it's still going to be resampled, and sound horrible to
audio buffs. What the Creative cards are great for is audio effects and
games -- bang for your buck, literally. And that's even good-sounding if
the game source sound is 48kHz :)
For music listening, though, you're usually better off picking a sub-$50
sound card with a Via Envy chip, no fixed resampling and software buffers.
The noise level might be higher, but it won't grate on the ears like the
resampled sound will.
Regards,
-- *Art
- Next message: Arthur Hagen: "Re: Recording from line input on SoundBlaster Audigy - Solved!"
- Previous message: Dr. Popper: "RAID MIRROR SW, PCSERVER AND BOOTING"
- In reply to: baskitcaise: "Re: KDE's farting error sound"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
