dac and streamer or seperates

@russbutton

This is where it doesn’t make sense to me. If a DAC receives the exact same digital data from two different sources, then why would it then result in a different analog signal?

You might get a better understanding by reading about what one DAC manufacturer did to try to improve the analog output. This is a website for the distributor for Holo DACs. You want to scroll all the way down, past the prices and read the section titled: Some noted features:

DAC Design Used to Improve Digital to Analog Conversion

Perhaps I’m overly generalizing, but it is also very important for the source to have a very stable clock. This is where the better Aurender and other high end brands excel. For those that can afford it, very accurate external clocks and both source and DAC clock inputs theoretically provide the best environment for digital to analog conversion. I2s communication protocol provides another theoretically great way to put a system together. If your source has a very good clock, I2s can pass the clock to the DAC and provide much the same benefit as the external clock, but unfortunately, many I2s implementations don’t allow for a clock signal to pass on I2s. PS Audio would be a glaring example of this. Paul (PS Audio Owner, engineer and chief marketeer) relentlessly pushes I2s, but doesn’t design equipment that allows for the clock to pass from source to DAC.

An option is a used Bricasti M3 with built-in renderer. The brand’s products are more obsolete resistant than others. It will require a separate server to run Roon in order to stream. That means there is the cost of Roon plus an inexpensive server, like $1000+. The advantage is the savings on the streamer. The built in renderer sounds good. I hate to be one those buy-what-I-bought, but I do think the built in renderer offers a really good value and certainly doesn’t prevent future upgrades.

I have tried my iMac/audirvana as a server. Even through the network input of my M3, it sounds awful.

@vonhelmholtz  I read through the link you sent.   It speaks about clocking at both the source and at the target.  I have to say I'm mystified as to why clocking at the source is an issue.  Data being transmitted from one device to another is broken up into packets which are reassembled at the target.  That's how I can copy a data file from a PC in California and send it 2500 miles to a machine I have in Hawaii.  Those data packets have passed through more than a dozen routers, each router having physical connectors, wire and fiber cables of unknown nature, as well as the data bits being converted from electrical signals to light, etc.  And yet in audio, people seem to obsess over a single 3 foot length of USB cable.

When copying a data file from California to Hawaii, the copied file is precisely the same as the source file.  Not almost the same.  Exactly the same.  I don't think it's unreasonable to assume that the copy of a file from a PC to a DAC cannot also be equally accurate.  I still fail to see how the data received by the DAC would be different from a streamer than it would from a PC.

The first program I ever wrote was in assembly language on an IBM 1401, which had 16k of actual physical core memory.  The IBM 1401 is proudly on display at the Computer History Museum in Sunnyvale.    You could ***SEE*** the core memory in a cage!  There were these little wire grids with ferrite donuts that were the actual bits of system memory.  One of the things I learned was that there are really only three things a computer can do, and that everything else we see a computer do are elaborate combinations of these three operations.

1.  You can copy bits from one place in memory to another.
2.  You can compare bits in one place in memory to another.
3.  You can do a numerical add of the bits in one memory location to the bits in another location and then write them to a memory location.

That's IT.  EVERYTHING a computer does is combinations of those three operations.  And that's the beauty of digital data.   Copying data is done perfectly or it's not done.  The data transmitted to a DAC, which also a little computer, is precisely the same whether it comes from a PC or a streamer (which is also a computer).  When data is not accurately transmitted, it is corrupt and useless.

I can understand that clocking the data for conversion within the DAC is an important design element, but the data it is working on will be identical coming over the wire regardless of whether it comes from a PC or a streamer.

@mgrif104 @vonhelmholtz   Of course all of this discussion is rendered irrelevant with a system like the Dutch & Dutch 8c when you connect to it over a local IP network as a Roon target.  Everything is fully contained in the D&D 8c - DSP, DAC, active crossover, digital power amp...  There are no USB cables, no interconnects, no speaker cables, no cable lifters, no balanced or unbalanced connectors, no source clocks, no...  did I leave anything out?

I think the D&D 8c is about $14k now, but consider what you can spend on all those components as separate items, $14k isn't a bad deal.

@russbutton 

I’m not familiar with the D&D unit you’re referring to. But, it does seem like such a unit attempts to be a one box solution that cuts out the variables. The only thing I think you’re missing is that each component can be optimized for audio quality and there are necessary limitations to any one box solution whether that be in design topology, quality of parts, power supply(ies) and isolation. But, $14k all in is an attractive price point relative to what I have invested in my digital system.

If I read correctly, you’re still understandably skeptical of the source’s value in the equation and that because of your location, are unlikely to experience worthwhile alternatives. To that I say you’re fortunate as you won’t fall down the proverbial rabbit hole.

You’ve obviously spent more effort and $ on your vinyl rig than on your digital system - which I get. I used to have a vinyl rig and miss it at times. But it doesn’t take a lot of money to make notable improvements in even modest digital rigs.

As I experimented, I began to hear, and understand the benefit of optimizing all the disparate digital gear in my system. It is more than just 1s and 0s. As already discussed - files are being transferred bit perfect. But, the noise that’s carried along the same wire along with the slope of the voltage changes which the DAC reads as a 1 or 0 are critically important components and are plainly audible in direct comparison. The human ear can detect mere femto seconds of jitter. It is very well documented (in listening tests and actual measurements) the impact of incorrect impedance of a digital cable and the length of the cable.

The OP started this thread wondering about the merits of separate components and it has been a good discussion and one I hope helpful to you - a skeptic.  Being a skeptic is to be encouraged in audio as there are plenty of ways to separate people from their money. But, please keep an open mind and seek direct experience to weigh against the theoretical.

Best,