[linux-audio-user] External DAC for mastering use?

Mark Knecht mknecht at controlnet.com
Thu Mar 18 12:41:26 EST 2004

Anahata wrote:
> On Thu, Mar 18, 2004 at 06:49:03AM -0800, Mark Knecht wrote:
>>I will also point out that you would not, I think, be able to hear the 
>>direct effects of your choice of dither in this setup. All audio you 
>>send to this unit would have to be dithered so all you could do would be 
>>make a comparitive judgement.
> While that may be true, it's the same problem whether the DAC is
> internal or external, if it's 16 bit in both cases.

Of course! But that wasn't my point. My point was he should really be 
using a 24-bit external DAC. Of course, it's more money...

>>Over the top of all of these points, I think you want to be pretty 
>>careful choosing your D/A. They do sound different.
> So the choice is between:
> - an internal DAC that may be of indifferent quality and pick up
>   noise from the PC's internal electrical environment.
> - an external DAC that will be electrically isolated from the PC and
>   will have nice clean power but may have clock jitter problems.

Don't fool yourself. Internal and external DACs can have plenty of 
jitter. External DACs that derive their clock directly from a USB source 
are pretty suspect in my mind, but they could have substantial PLL 
technology embedded in them that could help this.

External DACs that use an external low-jitter clock source, and then a 
PC sound card that will take it's clock from that same sound source, is 
likely to be a better system, all other things kept equal.

Whether we *hear* the jitter is a different (and interesting!) discussion.

> Aren't external DACs usually purchased as an upgrade to the internal DAC
> of a CD/MD/whatever? I would expect the sound quality to be noticeably
> superior or there'd be no point in buying it.

I don't know what the CD/MD/whatever means, but yes, extenal units are 
often purchased as 'upgrades'. Most systems offer good quality internal 
DACs. My Pro Tools system is like that. None the less people report 
better sound using (for instance, not a recommendation) the Cranesong 
HEDD 192 in bypass mode and separate word clock generator.

Two problems:

1) You'll drop about $5K for the opportunity to try it out.
2) Who expects anyone who just dropped $5K to tell you it *didn't* make 
a difference!!

One big problem with pushing toward high-end audio is that if you have a 
reasonably good system then you *cannot* hear the benifits of a great 
system by just trying it out for a few minutes. It's very interesting 
that when you sit in your studio using a great system and you substitute 
a reasonable component, like a mid-priced DAC, into it you often hear 
the differences right away. In short term listening tests human ears 
seem to be more able to pick out problems than hear improvements. My 
(small) concern about Daniel's proposed system is that once mastered 
he'll send his stuff to the people and maybe not get the response that 
he was hoping for.

But again, there is no reason that he cannot start this way and improve 
the system over time. We all do it. What he's talking of doing is very 
sensible given a small budget. We're just pointing out some of the issue 
that might arise.

> Of course, that's what salesmen call "telling the customer what he wants
> to hear", so it not guaranteed to be true :-)

Damn straight! ;-)

More information about the Linux-audio-user mailing list