Hey there,

My current audio setup consists of a Mac Mini, with an analog and an optical output, a Yamaha receiver (RX-V430RS) and a couple of speakers. The question that kept on popping up in my head (those annoying little voices..) is how much better will the overall quality be, if I use the digital output instead of the analog one to connect my computer to the receiver.

If I use a digital output on my soundcard, the sound output from the computer should theoretically be perfect (aside from resampling issues and the like). Is this correct?
Will the DAC in the receiver definitely be better then one on most soundcards (where should the signal be converted to an analog form)?


But either way, optical is the way to go for me, as it enables me to send surround (dolby / dts) signals from the computer to the receiver. Also it is definitely cooler.


thanks for you time,
phil

I used to worry about questions like that... the answer I never wanted to hear is: if you don't hear a difference, why bother? (ooh, I still hate that answer... )

The other answer would be a guess - I don't know about the Mac Mini's audio quality - but in general I guess it's safe to say that DA-conversion in the receiver will be better than it is on an onboard chip.

Besides, there are numerous advantages... you get DTS/DD through the same optical line, and the Mac and the stereo set will not be electrically coupled...

The answer is that the DAC on most professional or semi-pro soundcards will be better than that in the receiver. This doesn't apply to those on many gaming cards or to most MB built-in DACs, however.

It will be better exaclty in proportion to the amount of difference you can hear between your soundcard's DAC and your receiver's, in a blind, level-matched comparison. Which might well be: none at all.

Additionally, if you use any DSP functions on your receiver, and you feed it an analog signal, the receiver is taking it through another round of A/D/A conversion before output. Whereas if you feed it a digital signal, it stays digital all the way until the final D/A output.

As others have said it depends if you can tell the difference

Personally I use the digital out because I have crappy onboard sound which is not likely to have as good a DAC as my amp. I'm not saying I'd tell the difference but just knowing that theoretically my £800 amp should be better than the £4 onboard sound makes me do it.

Thank you for your answers.

A short test, which I had already conducted beforehand, concluded that I can detect no difference between the two methods, so my question was a hypothetical one anways. Damn curiosity and such.
In my case I will be using optical for the simple reason that I can send a surround signal to my receiver with it, and that my voodoo gut feeling* dictates that the DAC in my receiver is superior to the one crammed up in my Mac Mini.


By the way, the Yamaha DAC is rated at 96kHz and 24-bit. While these are only numbers, I still believe that I will be buying new speakers before I decide on a new DAC solution.

*my voodoo gut feeling is devoid of all scientific reasoning and logic

You are right. It should be perfect apart from various issues.

The onboard analog output might have some background noise.

You could use RMAA to measure the two outputs and see for sure the difference, though that would take some effort if you don't have windows available and/or something that could record the output of your system for analysis on a Windows machine in RMAA.

I get something like that


How do I fix it? It's totally annoying, like a hissing/humming sound of my case and fans and things.

Not much you can do about it. If it's fans you might see an improvement powering them off the 12V drive plugs from your PSU rather than motherboard headers (the adapters are cheap), but beyond that you are fairly powerless.

What I worry about with digital out on a PC is how much 'crap' the PC software pulls?

More specifically, currently I run the digital out of my SB Audigy directly into my professional DAC. In theory a good setup (and it sounds fine), but:

- If I reduce volume on the PC, how high-quality math (dithering?) is used to achieve this?
- If I DONT reduce volume on the PC, do I get the exact bit output of the WAV source?

Maybe the technology that does this is 'kernel streaming' (like winamp and foobar2k support) but I haven't looked into this too deeply yet.

I am still not sure whether I should be happy or annoyed about the fact that, when using the digital output, I cannot change the total system volume under osX.

While I also find it easy to mistrust the "voodoo", which happens in audio applications and hardware, volume control is something I would not be overly worried about. It seems to be a trivial task that every video and music application is able to conduct on digital material, and therefore I would assume that Creative's implemenation of it isn't all too inferior and will only degrade the signal-to-noise ratio.

You would be surprised.

The 'proper' way to do it, I believe, is to upsample a signal to say 4x its spec (so 16bit -> 24 bit, 44Khz -> 192Khz), then divide by volume value, then go back to original spec (16/44?) and then dither.

The FAST and also EASY and therefore CHEAP way to do it is to divide by volume and round off the result to the nearest integer value.

(compare how an anti-aliased image or font looks compared to a non-anti-aliased one "jagged edges". The principle is similar with audio)

Most DACs take more than 16/44; at least 24/96. Just send 24/96 output. Don't need to worry about dither then I don't think.

My DAC doesn't

Just as a followup in case some day someone happens to be searching for this information..


According to C'T (a german computer magazine), the Mac Minis sound card has the following specifications. Please forgive my crude translations, sadly I am not a pro at audio-lingo.

Analog Output:
Signal-to-Noise Ratio (Rauschabstand): 70dBA
Dynamic (Dynamik): 71dBA
Distoration (Klirrfaktor): 0,042%
Frequency Divergence (Frequenzgangabweichung): 4,8dB
Line-In Dynamic: 72dBA
Line-In Distoration: 0,038%

Digital Output/Input:
44,1, 48 and 96kHz


ps // I really like the noise level of my Mac Mini. The tests conducted placed the Mac Mini between 0,2 and 0,3 Sone (idle and load).

With these specs, I'd go for digital SPDIF output ... a frequency divergence of 4,8 dB with a rather low SNR of 70 dBA (most analog tapedecks reached that value) doesn't seem to be better than what your receiver has to offer.

Does the Mac Mini's sound controller perform internal resampling or does oit allow for bit-true output ?

That is definelty what I am going for.


Sadly, I have absolutely no idea. I presume that this is handled by OSX' Core Audio? Perhaps someone knows how this audio engine handles resampling?

To me the biggest indication would be this: if your receiver converts all analog inputs to digital (ADC, very likely), processes them, and then converts back to analog, combined with the rather mediocre DAC performance of the Mac Mini, I would definitely go SPDIF out to the receiver.

The choice is between:
1) Analog connection
Mac DAC -> receiver ADC -> receiver DAC

2) Digital connection
MAC SPDIF -> receiver DAC

#2 is much cleaner on paper, only 1 conversion from digital to analog vs. 2 digital to analog and one analog to digital. That is how I have my setup (different equipment of course).

This is a very good point. My receiver has a 48Khz DAC.All analog inputs get sampled so the DSP stuff can be applied, even though I try to avoid dsp when listening to stereo sound from my computer.

Rather than use my soundcard,which samples to 48, I use the SSRC plug-in to do the sampling, and send the signal via spdif to the receiver DAC.If analog was used, the card would resample, convert to analog with who knows what quality.The analog signal would then be resampled.Lots of unecessary processing this way.

As far as how it sounds, my digital cable box could be a/b'd from analog to digital.There was a volume difference but the digital signal sounded much cleaner to my ear, possibly the fault of the DAC in the cable box.

In general, less processing is probably better.

Well, you could know by measuring it, or finding a review of the card


A volume difference could well be the reason for hearing quality differences; this is I think the biggest catch with a/b testing.