[Community of A/V Enthusiasts]

Is it just me?

I've found that when it comes to DACs, I prefer the basic, non-upsampling mode. Anything else sounds too etched or sharper than normal and not real. The only analogy I can think of is tracing a beautiful picture hundreds of thousands of times. It will look like the original and then, nothing like the original.

It will have lines all its own and though resembling the original it simply will not fool anyone. That's not to say it will look bad, just different, and maybe even pleasing to someone. I also wonder if all of this upsampling to the nth degree is like the old horsepower wars of old: the more, the better, or so we old timers thought. High Rez downloads might benefit from upsampling but to me ears, via coax from my DVD player to my Music Hall DAC, the basic setting is all I need (but it does oversample 8X).

Just like separates tend to sound better by virtue of the thought and dedication it takes to fully implement them, maybe, just by separating the DAC from the transport allows designers to do the same and build a DAC with better parts without the size and constraints of a CDP.

Or I'm just whistling Dixie.

Either way, I feel that my Music Hall is just a well built DAC that does what it needs to and lets me sit back and enjoy the music.

Sorry about the long intro, but I'd like to hear what others feel about this.

Next Post >

Comment

You need to be a member of SECRETS CAVE to add comments!

Join SECRETS CAVE

Comment by Michael Nemoy on March 25, 2010 at 11:35pm: modern DACs operate in 32-bit

The problem is not in the rounding errors (there are none), but with the filters being applied after SRC.

Comment by Tim Patchett on March 25, 2010 at 11:33pm: Brian:
Thanks for the clarification. I've read similar info elsewhere and it's good to hear that I'm not alone on this.
My prior CDP was a Consonance CD-120 linear with a 16 bit, non over/upsampling with no opamps or digital filters and a J-FET output. I loved it for a few years but after trying the Music Hall DAC, it seemed to do all the good things that the Consonance does but with better extension on both ends, with better detail rendering, and a clearer, deeper presentation. All that, and no loss of that '16 bit' magic with the basic setting being used.
If I understood what you wrote, what the Music Hall does do, at the basic setting, clears up the jitter better than what the Consonance does but doesn't introduce the negative effects of upsampling since I don't use it.
I think I got it.
My head hurts now.
By the way, I just read a glowing review of the Virtue Piano CDP which uses a TDA 1542 (16 bit) but with upsampling and the reviewer loved it so much, he bought it. And it's under $600. I guess there's always more food
for thought.

Comment by Brian Alvarez on March 25, 2010 at 11:06pm: In actuality upsampling to a frequency that is not a multiple of the original, SRC (Sample Rate Conversion) such as from 44.1khz to 96khz or 192khz. Will introduce distortion. It's a mathematical inevitability. Doing so from 44.1 to 88.2 or 176.4khz will not as there's no floating point calculations with the inevitable quantization.

The main reason most new DACs use upsampling with SRC (and usually also over sampling as well) is to separate the clocks from the data receiver and the DAC. The receiver has to lock on to the clock sent via
the transport and the instability(jitter) that is inherent in SPDIF and Toslink transmission. By taking the data
from the transport receiver, buffering it, then SRC conversion the data is then re-clocked with a high
quality clock at the fixed rate that is the same as the SRC. This helps to remove the transport Jitter from
the signal. The SRC will introduce it's own issues but these in most cases are outweighed by the reduction in
jitter.

The reason the SRC is done to usually 96 or 192khz is that since the advent of DVD players, most DAC chips
are designed to primarily operate at multiples of 48khz. Since 48khz is the sample rate for video devices
both professional and consumer.

Many people (Myself, Ken Ishiwata of Marantz, and others) feel that for Red Book play back it's hard to beat
the sound of the older DAC chips specifically meant for 44.1khz. They also have the advantage of being much
larger in physical CMOS size than the newer chips which means they have more current. I've yet to find a DAC
that sounds as good as my 1999 California Audio Labs CL-10 based around now ancient 20bit oversampling
44.1khz DACs.

That may change though when I try out the Benchmark DAC1 USB this weekend. With my Meridian DAC, and others I always thought that upsampling softened the attack and drive of the music in exchange for more depth and a better sense of space.

Comment by Michael Nemoy on March 10, 2010 at 12:54pm: Most of high-end DACs use upsampling. The question is WHICH upsampling.
Swiss Anagram is quite frequently used.

I upsample all CD-quality material to 96/24 and POLYPHASE filter using SoX.
http://leute.server.de/wilde/resample.html

To my ears that sounds much better on almost every album

Comment by John E. Johnson, Jr. on March 9, 2010 at 10:06am: Upsampling really only has the purpose of making it easier for the filter to roll off the frequency response above 22 kHz (the Nyquist frequency for Redbook CD). It does not produce higher resolution anymore than 480i TV programs being upsampled to 1080p.

Comment by Jared on March 9, 2010 at 10:02am: How very Zen of you Piero!

It is great that you have the option to use/not use the up-sampling feature. I would try it with various recordings, you may find some benefit and some don't.

Comment by Piero Gabucci on March 9, 2010 at 9:40am: Tim you said the most important thing, you enjoy your Music Hall. So many think there is always better out there and are unsatisfied.