Digital Sources computer versus Streaming

[idiot_savant]

I like your thinking - I reckon you're angling for a sidekick role, but I am, umm, a man without a past, or an albanian goat herder or something

 

and,

 

I can do whatever an idiot can

 

your friendly neighbourhood idiot

 

[idiot_savant]

I'll elaborate tomorrow on how there may be a difference between the "interface free methods" - which I suspect will be very low, YMMV,

 

 

your friendly neighbourhood idiot

PS I hate "Your Milage May Vary", but I've used it myself. I've even posted it twice

 

[idiot_savant]

my purpose for posting here is that:

 

USB audio can be great or rubbish

Firewaire auio can be great or rubbish

Ethernet audio can be great or rubbish

SPDIF can be great or rubbih.

2 bits of string will paper cups will be fun

 

your friendly neighbourhood idiot

 

[cfmsp]

 

IS,

 

thanks for all you do here, in the way of posting.

 

Your explanations are quite clear, and many times leave no obvious follow-up questions, at least from me.

 

Please take any 'silence' as more of a ringing endorsement as to the quality of your outpourings, more so than lack of interest. I'm always quite happily surprised to see another of your posts, and I read them all.

 

Indeed, I'm quite interested in another topic, if it's one you have knowledge of, I'd be very interested in your response. I'll post a new thread, but just in case your 'monitoring of the skies' doesn't catch it, the topic will be avoidance of self inflicted pain of the RF / EMI /poor AC grounding variety.

 

thanks again for your efforts here. they are much appreciated.

 

Clay

 

 

 

[idiot_savant]

thanks for your kind words,

 

As it happens, I'm afraid the mighty booze have overcome me this particular session,

 

back for more another day,

 

your friendly neighbourhood idiot

 

[icebreaker]

"Please take any 'silence' as more of a ringing endorsement as to the quality of your outpourings, more so than lack of interest. I'm always quite happily surprised to see another of your posts, and I read them all."

... with Clay's remarks. But being far less of an expert than Clay on these matters I find your posts give me enough to contemplate. But there are always more questions... aren't there? ;>)

 

[idiot_savant]

As I have stated before, there are methods of having the interface completely remove itself from having an impact on the jitter, by simple virtue of having the critical clock in the system ( the one at the DAC ) being in charge of the flow of data into itself.

As mentioned earlier, every clock has intrinsic jitter, and the amount and spectrum of this in the DAC ( it being the master ) is pretty much down to how good the implementation of it is - so, even if your interface is jitter-free, if you muck up the critical clock inside your DAC, you will still have poor overall jitter performance. This can be seen in the evolution of one-box CD players - here, there is no reason for the DAC to do anything else but be the master clock, yet jitter performance between CD players varies ( some due to poor design decisions, like having the DAC not be the master inside the box ). In the case of the CD player, the design of the power supplies is absolutely crucial as you have a motor drawing variable amounts of current at different times.

 

Now, in our USB,Firewire or network DAC, we don't have to deal with electric motors, but we do have a problem. Due to the very nature of the fact that the DAC is the master, and is changing the data flow to keep it's buffer happy, the physical interface between PC/Mac/Hub and the DAC is asynchronous. This means that somewhere near the DAC is another clock that is nothing whatever to do with the audio. Now the problem is that separate asycnchronous clocks "beat" against each other, which can cause effects at the analogue out of the DAC.

So how do you fix this? Well, the answer is back to plain old good engineering - somebody taking care with the isolation and the shielding - hence you may find some differences between implementations, which isn't as simple as it may appear,

 

There are no examples I can think of where say 2 DACs are identical except DAC A uses interface X and DAC B uses interface Y - there a large combination of things, that in isolation are small differences, but that sum up to give quite different outputs.

 

However, I reckon a well engineered DAC will sound better than a poorly engineered one,

 

your friendly neighbourhood idiot

 

[Purite Audio]

IS, what should we look for in the real world, obviously price alone is no recommendation, what can we look for in the specs , given that manufacturers aren't always forthcoming?

Just listen? Keith.

 

[idiot_savant]

Is that they are so often wrong!

No other industry would get away with it - if you bought a new car for it's performance, and say it was quoted to do 0-60 in 5 seconds, yet took 10, you'd be upset!

Yet hifi manufacturers regularly just quote the theoretical performance of a component in isolation - or just make them up ( 1pS jitter! ).

 

At the end of the day, you listen to the stuff. I personally as an ethos like to look at measurements in magazines to see if a manufacturer lives up to their claims, and the ones that do would be a good place to start, as it is very often difficult to listen to the exact combination you would like, and you have to start somewhere! Also, bear in mind that your personal taste is extremely important - if you like the sound of valves, for example, then a solid state component, no matter how well engineered, may not be to your taste, and vice versa.

 

your friendly neighbourhood idiot

 

 

[Purite Audio]

IS, just harking back to dac chips/implementations in your experience do sigma deltas, multi- bit, NOS have any form of family character ? Or is the sound of each particular dac mainly due to its design and implementation, ie power supply analogue output etc.

 

[idiot_savant]

With most DACs, the key thing is the "weakest link" - the output will tend to gain the character of the thing with e.g. the most distortion. The main odd one out here is the NOS DAC, which will nearly always sound "different" due to it's extreme dependency on the source material and the partnering amp/speakers, together with it's HF rolloff.

 

This isn't to say it doesn't matter - but it will only matter if the output stages are of sufficient transparency, and matched to that DAC ( e.g. the output stages frequency response requirements will be different between DAC architectures ),

 

your friendly neighbourhood idiot

 

 

[cfmsp]

IS says:

 

"The main odd one out here is the NOS DAC, which will nearly always sound "different" due to it's extreme dependency on the source material and the partnering amp/speakers, together with it's HF rolloff."

 

The singular characteristic most often used to describe NOS is 'analog sounding'.

 

I wonder if this is due mostly to the HF rolloff, or what you describe above as 'extreme dependency' on the partnering amp/speakers, which might likely also be analog-like for those that seek out NOS DACs as the cure for socalled 'digitalitis'.

 

 

"Extreme dependency on the source material" is a good thing, yes?

Sounds like transparency to me, unlike the 'false' transparency referred to in countless audio reviews as 'ruthlessly revealing of the other components in the chain' (aka tendency towards 'distortion' in the upper midrange).

 

clay

 

 

[idiot_savant]

The HF rolloff is relatively subtle - starts at about 10k, and is a few dB down at 20k, so may well have some effect on the perceived sound. I think it's liable to do with the whole DAC that uses a NOS topology - these will tend to use a valve output stage, which may offer some cues that remind people of vinyl/analog. I don't believe that the lack of filtering is a good thing - a NOS DAC will cause very strange mixing effects on amps with less than ideal intermodulation performance, for instance.

 

The delta-sigma/multibit/NOS thread covers my views on the need for a filter ( i.e. I think you do ),

 

 

EDIT:

the "extreme source dependency" I was referring to is a good thing in one way

Unfortunately, in another it means that source material with a lot of HF energy will produce more mixing in areas that are close to the audio band...

 

your friendly neighbourhood idiot

 

 

 

 

[cfmsp]

 

"my views on the need for a filter ( i.e. I think you do )"

 

yes, I recall that. Indeed, filtering seems to be the new fertile ground for product differentiation in the DAC marketplace.

 

thanks as always for sharing your thoughts. I'd rather read one of yours than two of most anyone else (and any amount of my own prattlings).

 

Now if we could only coax CG back. He hasn't posted since the 'bye bye' remark that I can remember.

 

clay