First, some background:

I've been using the digital coax out (1/8 miniplug) on my SBLive Value, along with several "headphone" extension cords (Radio Shack vintage), ending in a 1/8 to RCA adapter, plugging this into the digital coax in on my JVC home theater receiver. Obviously this is all consumer lever equipment, no high falootin audiophile stuff in my living room. The quality has always been perfectly acceptable (although I went to the digital connection because the 50' run of analogue signal picked up a really nasty hum...).

Unfortunately, my new wireless PCI card really doesn't like the SBLive - it would seem the dreaded resource sharing issues with the SBLive have finally reared their ugly heads. I have removed the offending card and am now using integrated audio (Creative CT5880 PCI sound controller ), again, digital coax connection. I am not happy with this solution; skips, quick freezes, only one program can access the "soundcard" at a time, etc. So I am looking for a new soundcard.

My question (if you're still reading...:
If the soundcard is passing a digital signal to the DAC on the JVC receiver, does it matter which soundcard I use? Won't they all sound the same as the receiver is doing the actual conversion to analogue?

The sampling rate info I've seen is somewhat confusing, although I get the basic idea that higher is better and the more headroom with which to process the signal the better, etc. I can convert my experience with higher bitrate colour processing over okay, but the specifics of up and down sampling get too twisted for my feeble head. Would it make sense to match the sample rates of my receiver's DAC with the soundcard's to avoid spending too much on a high preformance PC card that just gets downsampled at the receiver? Or does the higher sample rate (before conversion to analogue) allow for smoother output when downsampled (like in Photoshop...) Hope that was clear - nothing like someone who doesn't really understand something trying to "verbalize" concepts...

The features I want in a card would include:
- digital coax out
- digital optical in (receiver can output optical, I'd like to record sound this way)
- good music performance much more important than games (I play 90% music, 10% games)
- that said, it would be nice to output a digital 5.1 signal to the receiver (more likely in a future game than DVD as I have a DVD player in the HT setup), I understand Audigy (1+2?) will not do this - DRM??
- plays nice with the rest of my computer
- reasonable price

The Audigy2 seems to meet most of my criteria, however, I am a little concerned that it too will give the wireless card a hissy fit. What other cards meet my needs?

Well, this seems to have turned into a product recomendation thread. I don't know if that is frowned upon here; if so, please feel free to ignore the 2nd half of my post.

Again, the meat: If the receiver DACs are processing the digital signal from the soundcard, will different cards sound different?

Yes, you might notice a slight difference, because, to put it in simple words, some cards have a lossy digital output (and input too). This is the case of the Audigy 2, and the SB live too, but since this is consumer equipment, you should not notice any difference.
Some cards from Terratec and M-Audio have lossless digital in/out. But in all the recent cards, I'm lost. Note that you can't play games losslessly, or mix different audio sources losslessly. You must find a card with both optical and coaxial, because coax/opt converters are expensive (80$).

How u claim a card's digital output is lossy?



[Edited] Remove unnecessary words.

Thanks for the reply. What does the above mean? I would certainly look for (at minimum) coax out and optical in, if not both I/O.

Would you say that, if the difference is nearly non-noticable, that I should therefore look (shocked gasps from audiophile crowd) at things other than sound quality when making my decision? IE: CPU usage, I/O ports, etc?

Can you elaborate on what you mean by a "lossy" digital i/o?

Lossy digital-out explained: the classical example of this is the Soundblaster Live.
The EMU chip on that thing has a fixed internal clock at 48kHz. All audio going through the card must be resampled, since this chip cannot be slaved to an external clock.
That means, 44,1kHz CD-data is converted to 48kHz on the way in/out. Even when you feed the input with 48kHz (e.g. from a DAT-deck) digital signals, they will be resynced to internal clock.
Resampling is a lossy process, although according to Pio2001's listening tests, it is hardly audible.

If I understand your explanation well, you're only playing from digital out, and not using the analog outputs?
This is a good read then: http://www.hydrogenaudio.org/forums/index....&t=6891&hl=2448

So everything that being resample is consider lossy digital output, then I have two questions here:-

1. Audigy 2 did not resample everything to 48kHz, it support 96kHz and 192kHz internally, is it still consider a lossy digital output?

2. Based on this theory, can I say dCS Purcell is a lousy product since it is a upsampler?

OK, so it's not really a lossy digital output, but rather a lossy (inherent) resampling process. Perhaps we should call a spade a spade and clarify that there is no "loss" in a digital output, but losses will be inherent in a resampled card's digital audio path before the output...

Well, I shouldn't have put it in simple words after all.

A soundcard like Creative soundcards and current AC97 soundcards resample anything they play to 48 kHz (maybe 96 kHz for the most recent ones), and resampling is lossy. Thus, the soundcard is lossy, including the digital out. It can't play losslesly a digital file into the digital output, unless you use some weird tricks, like stating that your digital file is a Dolby Digital file.

Example : here's a 22 kHz sine, sampled at 44.1 kHz, recorded in loopback with the SB live soundcard, through the coaxial digital output/input
Not only is it impossible to set the volume equal to the source (when it's just lower, you move the record volume one step up, and it becomes louder than the source), but the "digital copy" looks completely different than the original. Thus it's not lossless.

About mixing, I meant that if mixing sources from different sample rates is possible, some of them are necessarily resampled, unless they are mixed with an analog mixer. But then you need one DAC per track.

Thanks for the clarification. A properly implemented SPDIF should support 44.1KHz output, so the problem lies in Creative's lack of native 44.1KHz playback support. I guess a properly supported SPDIF is what you would call "not lossy"...

Whoa ! Was I so long to answer ? There were 4 other posts in between !



This killer sample sounds like seven quiet tones when played losslessly : http://www.hydrogenaudio.org/forums/index....t=ST&f=1&t=9772

It is a revelator for resampling and leveling processes. In this post, Halcyon reports distortions with the Audigy 2 Platinum EX, showing that it doesn't play digital files without processing them.
http://www.hydrogenaudio.org/forums/index....9772#entry98389

The problem is that although the Audigy 2 does support 44.1KHz playback through it's DACs, it's digital SPDIF output only supports 48KHz and 96KHz sampling rates. Thus, anything with a 44.1KHz native sampling rate gets upconverted before being output to the SPDIF plug.

I have separate discussion with Halycon on this at 3DSS Forum, check it out HERE.

As mentioned in the post, the IMD only affected the frequency above 20KHz, thus the noise u all hear I believe is the level that out of the card's acceptable range. Nothing to do with IMD at all.

Then how do you explain that the problem nearly disappear when the sample is resampled (and not leveled) before playback ?

I browsed the original thread for references about the problem being solved by resampling :

Halcyon / Audigy 2 Platinum EX DirectSound / http://www.hydrogenaudio.org/forums/index....9772#entry98389
Halcyon's setup : http://www.hydrogenaudio.org/forums/index....t=75#entry99616

LPTB / Soundmax integrated chipset / http://www.hydrogenaudio.org/forums/index....t=25#entry98481

JonPike / Audigy 1 / http://www.hydrogenaudio.org/forums/index....t=50#entry99544

In a computer, the signals have to be resampled somewhere in the data path, because there are many signals being mixed.

So you have some 8bit, 22kHz windows sounds, mixed with your 16bit, 44kHz winamp stream. The easiest (only?) way to mix them is to make them all have the same resampling rate. This is most likely done by kmixer. But AFAIK, newer cards do all the mixing in hardware, so kmixer is just a "front-end" to the hardware, not an actual software resampler.

Things get even trickier when you want to play a DVD. You can't send out a DD/DTS signal together with a PCM signal. So will the DVD stream block any other streams?

So however you cut it, most likely your path will include one or more resampling steps. As long as you can't hear them, you don't need to worry about it.

From your link, this is what I found:

"I can get rid of this last trace (I think) of distortion by reducing the Master volume control to c. 50% and the same for Wave/mp3 playback volume (both in Creative Surround mixer). These are not exact figures, as I didn't have time to test all percentile value combinations one by one. I still have to connect the Audigy 2 to my main testing rig to verify whether the distortion was just masked or truly eliminated. At least it's not horribly audible anymore."

Is this anything IMD related cause by resampling related? No, I don't think so.

It is more likely the level itself to me.



[Edited] In bold.

Right, analog (or digital ?!) clipping also affects this sample. It really can't stand any kind of lousy processing.