Case and Power
My preferred PSU is a high-speed ultra low noise regulator based on a design by Walt Jung. Again, this is a partial DIY kit that is provided by Peranders Sjöström and known as the JSR-03 regulator. The JSR-03 is used as a pre-regulator, outputting around 9V to the I2S and WM8741 modules. Although a 12V relay is present on the Doede PCB to select from PCM2707 or TDA1543 output, I found a 9V input perfectly capable of making the relay switch. The USB to I2S DAC has an onboard 6V regulator circuit that requires around a 2V margin to regulate effectively, so I'm within its bounds.
 A Maplin KR56L Aluminum case provides ample room to house all the PCBs, with enough room left to add modules at a later date if required. |
To provide the raw or pre-regulated DC, I use either a 12V Wal-Wart PSU or an 1800mAh 12V Li-On battery pack. Total current draw of all combined parts is in the region of 500mA, so the battery pack allows for around three hours of listening before needing a recharge. An alternative would be to buy a high capacity sealed lead acid battery, though I decided against that mainly due to the additional shelf room it would require. I only use the battery packs for critical listening and component comparisons. For normal everyday use, the battery pack is simply removed and input switched over to the Wal-Wart PSU. Subjectively, I find the battery input to provide cleaner, tighter bass notes, more perceived air throughout the mid-band and high frequency range, and better stereo imaging. It's not a smack in the face difference compared to the Wal-Wart PSU, as pre-regulation by the JSR-03 offers superlative noise rejection throughout the audible frequency range. Still, the difference is enough for me to accept the hassle of changing out and charging batteries for that last ounce of playback clarity.
If you're not interested in going to the nth degree of audio resolution, the Wal-Wart solution is probably the best way to go. You'll also find some discrete regulator solutions offered in package deals with these DAC kits on the DDDAC and Twisted Pear pages. Either of these options should lead to a very satisfactory performance, although the Jung based regulators have superlative paper specifications.
Facebook
Twitter
RSS
114 Comments
View All Comments
mindless1 - Monday, December 1, 2008 - link
The fact is, even audio streams that measure the same can and do still sound different. The problem is the resolution of measurement and the misconceived notion that the brain interprets sound at a fixed interval as measured.There's a lot of snake oil in the audiophile world, but there's also a lot of what you'd like to call "magic" only be cause you don't accept it as non-magic.
Any decent DAC would in idealized theory be as good, but in practice a different IC topology may lend itself better to certain inherant localized noise frequencies and cutoffs, be better mated to the circuit it's dropped into, have drifts from thermal changes, etc. If they were all the same why would there be so many? I will agree that which modern DAC is used in a reasonably good design matters less than what follows after it in the chain but the best way to minimize any potential for degradation is to start out with what is most likely to minimize it in every way possible then following this concept the entire time, waiting and seeing if the end result is audibly different rather than downplaying them all without knowing the additive result yet.
CSMR - Monday, December 1, 2008 - link
Yes if you are maximizing quality you will choose the best of all components. But it's more sensible to care about cost and time too, so you have to prioritize. The value of the research about the unimportance of DACs (at at beyond a certain level) is that you can stop worrying about this part of the chain and spend you time/money where it is important.JonnyDough - Tuesday, December 2, 2008 - link
Exactly my point. Why spend more on one component if it's going to be "bottlenecked" by another one. Spending $2000 on a nice amp is crazy if your other components are crap. I think in the end it's like all other techs. You want decent stuff for a reasonable price unless you have more money than brains. While I scoff at people willing to spend more than few thousand on a sound system, without them we wouldn't have gotten to where we are today - with good possibilities on the market. As long as you like what you hear, who cares if it's perfect? At one point is something "good enough?" I mean honestly, there are people starving in the world, friends dying of cancer, etc...and we want to worry about whether or not something is inaudibly "perfect." Blow your money on something that MATTERS, you can't listen to music every hour of everyday unless your job is singing.CSMR - Tuesday, December 2, 2008 - link
I meant you can spend time/money on speakers/room acoustics/dsp but food for starving people is admittedly a better use of time and money.JonnyDough - Wednesday, December 3, 2008 - link
That's actually a matter of perspective. For example, you can donate food to a starving kid in some third world country and next thing you know that child has three kids and no way to feed them - and you've just marginally made the problem worse, not better.I think money is better spent on education, which leads to fewer babies. Many countries do not have quality farmland that can support the population on their own (the U.S. supports a large percentage of the world's population). Then there's natural habitat destruction, pollution from oil for having to ship food to them, etc. You get the picture.
JonnyDough - Monday, December 1, 2008 - link
While I was generally agreeing with CMSR above, I guess what everyone ends up saying is that hearing is subjective, which is something audiophiles are always agreeing upon yet they still love to argue over things.Does a vinyl record produce better sound than a CD?
It depends on what an individual values most.
A CD may produce less static noise, but have a "duller" sound, seemingly less highs and lows. It all depends on what one appreciates with their own ears.
While there are measurements that can be taken in a closed chamber, one sound system might actually sound better in a specific home or room than another for whatever reason. Even so, the average person can rarely tell a difference these days between the moderately priced components.
The really silly thing is that people will spend $200 on a sound card and then use cheap plastic speakers.
Rajinder Gill - Monday, December 1, 2008 - link
Lol, it's ok, everyone in entitled to have their say.It's the first piece so bear with us while we try to cater to a wider set of ideals.
With regards to the obj/subj stuff, it's an argument that'll never be solved.
later
CSMR - Monday, December 1, 2008 - link
Sorry if I came on too strongly; I was just expecting that Anandtech would have not necessarily an expert take but at least a more technical take than you get in audiophile communities.I would do some more fundamental thinking about what are the key factors in computer audio playback.
As a start you could rank in terms of what is most critical:
computer performance, computer quietness, software (drivers, dsp), analog line out, volume control, amplification, cables, speakers, room acoustics.
I won't give you my list but you should think about it and it is possible to compare many of these things in a very quantitative way, and some are more important than others by several orders of magnitude.
AnnihilatorX - Monday, December 1, 2008 - link
A dumb question about correction.The theory behind it is simple enough, but why do we need dedicated hardware to do the correction? Why can't we use simple graphical equalizers? I understand graphical equalizers are quite discrete in the range of selectable frequencies. How much difference does it make?
Rajinder Gill - Monday, December 1, 2008 - link
DRC - ease of use and range of control, plus taking care of things in the digital realm rather than in the analogue.You could fudge some control with a multi band digital equaliser and the means to measure the response, but it'd be damn tedious and utimately limited.