Dedicated Sound or Onboard?

[imported_BigT383]

I'm pretty well versed in computer hardware, and I have my own opinions/suspicions about these questions (almost typed: questinos), but I just wanted to get some others' opinions on this.

Back in the day, I had an Audigy A3D card. Loved it. Then the company went under, before making any A3d-compatible drivers for Win2k (I switched from win98 to win2k around beta 3, I hated win98 so much). Then I switched to the Sound Blaster Live. When I got an NForce2 board, I moved over to the soundstorm, which rocked, but now that I have my NForce3 I'm just using the NForce3 onboard, which I know is nothing special. So I've got a few questions:

1. Do you think I should be using the onboard sound or my old SbLive? I have a digital link to my speakers so the crappy onboard DACs don't get used.
2. How are Creative's drivers for the X-Fi? I know the Audigy drivers were horrible.
3. Besides CPU usage, what would be the benefit of moving to the X-Fi or another dedicated sound card? (Remember I'm using a digital output)
4. How much can using the onboard really affect CPU usage? I mean, I have an Athlon 64 3500+, which hopefully I'll soon be upgrading to a 4400+. Stereo sound can't really impact performance that much, can it?

 

[JEDIYoda]

Originally posted by: BigT383
I'm pretty well versed in computer hardware, and I have my own opinions/suspicions about these questions (almost typed: questinos), but I just wanted to get some others' opinions on this.

Back in the day, I had an Audigy A3D card. Loved it. Then the company went under, before making any A3d-compatible drivers for Win2k (I switched from win98 to win2k around beta 3, I hated win98 so much). Then I switched to the Sound Blaster Live. When I got an NForce2 board, I moved over to the soundstorm, which rocked, but now that I have my NForce3 I'm just using the NForce3 onboard, which I know is nothing special. So I've got a few questions:

1. Do you think I should be using the onboard sound or my old SbLive? I have a digital link to my speakers so the crappy onboard DACs don't get used.
2. How are Creative's drivers for the X-Fi? I know the Audigy drivers were horrible.
3. Besides CPU usage, what would be the benefit of moving to the X-Fi or another dedicated sound card? (Remember I'm using a digital output)
4. How much can using the onboard really affect CPU usage? I mean, I have an Athlon 64 3500+, which hopefully I'll soon be upgrading to a 4400+. Stereo sound can't really impact performance that much, can it?

Depending on the game--YES!! Onboard sound can be a real pita to use while gaming!

 

[YOyoYOhowsDAjello]

If you want to use a digital output, a Chaintech AV-710 would do the trick, but you're not going to get surround sound in games from digital like you did with soundstorm. The Creative cards don't do that either.

There are a couple cards that output DDL like soundstorm does though.

What kind of speakers do you have?

 

[Concillian]

Originally posted by: JEDIYoda

Depending on the game--YES!! Onboard sound can be a real pita to use while gaming!

Bah, most games are not CPU limited at the resolutions normal people play at. If they are, the FPS is usually well over the point where it really matters.

You will see claims that game X sees a yy% performance increase, but these claims are usually at totally unrealistic settings (640x480 noAA noAF) to make it seem like there is a huge gain by using an add-on soundcard. Also, typically the FPS are already quite acceptable, so while you may get a gain from an average of 100 FPS to 120 FPS, who's really going to notice that kind of a difference?

Crank up the settings and the performance difference between on-board sound and an x-fi starts melting away as the CPU becomes less and less important and the GPU becomes more and more of a limiting factor.

 

[Shadrack]

Originally posted by: Concillian

Originally posted by: JEDIYoda

Depending on the game--YES!! Onboard sound can be a real pita to use while gaming!

Bah, most games are not CPU limited at the resolutions normal people play at. If they are, the FPS is usually well over the point where it really matters.

You will see claims that game X sees a yy% performance increase, but these claims are usually at totally unrealistic settings (640x480 noAA noAF) to make it seem like there is a huge gain by using an add-on soundcard. Also, typically the FPS are already quite acceptable, so while you may get a gain from an average of 100 FPS to 120 FPS, who's really going to notice that kind of a difference?

Crank up the settings and the performance difference between on-board sound and an x-fi starts melting away as the CPU becomes less and less important and the GPU becomes more and more of a limiting factor.

I very seriously doubt you are a gamer with any gaming experience.
What you are say just is not true at all.
These claims are based on realistic setting not rookie gaming settings.
and its obviously you have no working knowledge of frame rates and what they actually means when playing a game.
There is a huge difference amongst hardcore gamers between 100 and 120 fps...
I`m not being mean but you really are out of your league when discussing gaming and the effects that onboard sound has while gaming.

 

[imported_BigT383]

I'm pretty sure I get what Concillian is saying...

Obviously if you get a max of 120 FPS but a minimum of 15 then yes you can definately tell, but even the most advanced gamer can't tell the difference between a minimum framerate (not average or maximum) of 100FPS to a minimum framerate of 120FPS... The visual system in the brain only processes at about 85 updates/second. Everything over that should be unnoticable. It's when you dip below that that things get problematic.

 

[Concillian]

Originally posted by: Shadrack

There is a huge difference amongst hardcore gamers between 100 and 120 fps...
I`m not being mean but you really are out of your league when discussing gaming and the effects that onboard sound has while gaming.

Most people don't have monitors sensitive enough to display a difference between 100 and 120 FPS (my monitor runs at 85Hz, for example, and very few people are using >100 Hz CRTs. The best LCDs on the market are only around 12-14 ms average for a typical color change (even though rated lower, those are only on transitions in a very specific color range), which translates into a quickest displayable (accurate) color change of around 70-80 FPS on a pixel.

When using FRAPS to analyze framerates, it's very clear to see what is affected when you drop the CPU speed. Generally the FPS drops in the higher FPS regimes by lowering CPU speeds and in the lower FPS regimes by lowering video card speeds.

For an Average frame rate of 60 FPS, your max is probably in the 100 - 150+ range. It doesn't matter, even if you're hardcore, in this range, primarily due to display hardware limitations. What you care about are the minimums in the 20s, 30s or 40s. This is where you NOTICE slowdowns. This is also typically where CPU makes very little difference.

When I analyze things for my own optimazation of visual quality and framerates, I will use FRAPS data and simply chop off any data above 85 FPS (the refresh rate I use on my CRT... it's simply not physically possible to detect changes above this, because the monitor won't display them) This adds sensitivity to the changes in minimum frame rates.

Here is an example of what I see when I do this kind of thing (though I generally take smaller steps in CPU speed when looking at a specific game that I play):

http://www.theinquirer.net/?article=23973

Scroll down to the section titled "Memory and CPU"

Between 2.4 GHz and 1.6 GHz the only real difference is a section where the FPS is already very high. It's not until the CPU speed is dropped ridiculously low that it starts affecting the bulk FPS in a way that is noticeable. If you are sensitive to low framerates, you'll notice it most at the point where it drops to the 30s, and this is clearly not at all affected by the CPU speed in a speed range typical to a modern CPU in this example.

This is a typical result from most games I've done detailed testing on. There are certainly exceptions. I haven't done testing on CS:S, but that game supposedly is much more driven by CPU speed, so should show a larger impact if you decrease available CPU by using on-board sound.

Most gamers (not professional gamers, but people playing for enjoyment) will increase visual quality until framerates are unacceptable (then move back to acceptable). In these cases, CPU is rarely a limitation, and on-board sound will give you little issue. In cases where you are worried about every last FPS, then of course it's an issue. But from the tone of the original post, I assume he's more in the first category than the second.

The final conclusion will depend on the specific game in question. Based on the linked review, BF2 shouldn't show much difference until your minimum FPS is getting into the 50s or 60s. Most people would consider that pretty acceptable.