Can anyone clarify my DAC conundrum?

[nautilus]

Let me start by explaining why I'm asking this question. A couple years ago when I first went to college I didn't know much about audio gear and ended up replacing my broken Sony receiver with a low end Onkyo TX-SR503. Over the last couple years I bought some Wharfedale Diamond 9.2's, a Bob Carver Dominator D-12 sub and some AH-D2000 cans, at this point the receiver is the obvious weak point. Anyway, I'll get to the point here I want to get into vinyl and instead of spending $100 on a decent Phono preamp I figured this would be a good excuse to upgrade the receiver to one with a built in phono input like the Denon AVR-2802 (which I found a good deal on) but all of my music comes from my macbook pro via optical in mp3 or lossless format and the DAC in the Denon is 24bit 96khz opposed to 24bit 192khz in the Onkyo. So here's the question since the CD standard samples audio at 44.1 kHz there should be no audible difference from mp3's? And the higher sampling frequency could only make a difference in sources like a SACD?

I'm obviously pretty ignorant about this so any information would help, including a link to a better explanation (which I couldn't find). Also are there any variables that I'm leaving out like oversampling?
Thanks for reading all of this,
-Mike

 

[hciman77]

Quote:

Originally Posted by nautilus /img/forum/go_quote.gif ...music comes from my macbook pro via optical in mp3 or lossless format and the DAC in the Denon is 24bit 96khz opposed to 24bit 192khz in the Onkyo. So here's the question since the CD standard samples audio at 44.1 kHz there should be no audible difference from mp3's? And the higher sampling frequency could only make a difference in sources like a SACD?

1) MP3 ripped from CD will normally be sampled at 44.1, resampling probably wont do any harm and certainly will not add anything, 24 bits willl give you 8 bits of padding, you dont really gain any information, resampling at 96 or 192 doesnt really help because you have already lost everything above 22.05 Khz. Some reckon that upsampling lowers noise by moving aliases out of the audible range, I dont know how much audible effect that really has. A recent AES paper suggested that it was extremely difficult to detect the effect of downsampling DVD-A and SACD audio down to 16/44.1 standard, take that any way you like.

2) SACD only outputs an analog signal as far as I know, i.e you dont get to bypass the players internal DAC, I may be wrong.

3) My external DAC is a 20 bit DAC but accepts 24 bits and accepts signals at
32K 44K and 48K - my DVD player outputs a 24Bit /44.1 signal from CD or MP3. Damned if I can tell a difference between this arrangement and when I feed it a signal from my 16/44.1 CD player or the 16/44.1 signal decoded by the CD players internal DAC.

 

[nautilus]

Thanks, so essentially it shouldn't make a difference right?

 

[hciman77]

Quote:

Originally Posted by nautilus /img/forum/go_quote.gif Thanks, so essentially it shouldn't make a difference right?

I would doubt it.

 

[facelvega]

I wouldn't worry about the difference in stats in this case, which as hci points out, won't impact cd sound. However, different dac chips and implementations do sound different from one another, so really the only way to get an idea of the difference is to hear for yourself or try to find model-specific reviews.

 

[hciman77]

Quote:

Originally Posted by facelvega /img/forum/go_quote.gif ...different dac chips and implementations do sound different from one another

This has always puzzled me. Clearly some CD players do sound different from others, we have blind tests to prove it, but how much is to do with the DAC mechanism ?. When you look at the specs and measurements on modern CD whatsits the Digital stage is almost always exemplary. Apart from manufacturers like Wadia who build-in 3DB roll-offs the FR on CD players is as flat as the proverbial pancake certainly you almost never see more than +/- 0.5db from 10hz to 20HZ. Jitter has been done to death but apart from some *extreme* examples you just dont get jitter above 500ps or with sidebands much above -100db. It occurs to me you can mess up an audio signal much more easily once it hits the analog domain...

 

[facelvega]

Quote:

Originally Posted by hciman77 /img/forum/go_quote.gif It occurs to me you can mess up an audio signal much more easily once it hits the analog domain...

For sure the amping of the output stage and the power handling make up a great deal of the impact, but I'd also speculate that the dac sections do too, in ways less easy to measure. I like to think in these terms: that the human ear, though easily fooled in some ways, is sensitive enough in others that it can pick up differences between components that achieve identical nominal accuracy, so that even two extremely good components can still seem quite different. But for sure this is vastly less noticeable in sources than in transducers, no question.