High end streamers, waste of money or not, please?

[webster]

Hi Stack Audio. You say you believe “that a digital streamer set up does make a strong difference”. Since you actually make and sell a digital streamer, the Link II, can you actually back up this claim with any measurement at the output of any DAC that shows your streamer makes a measurable difference?

Or maybe if he's that confident of an improvement he'd be prepared to loan out a demo unit to any PFM'ers who are interested in testing - I'd certainly like to give it a go.

 

[StackAudio]

Or maybe if he's that confident of an improvement he'd be prepared to loan out a demo unit to any PFM'ers who are interested in testing - I'd certainly like to give it a go.

Hello webster.

The idea of circulating a demo unit to the users of PFM is something we would certainly be open to.

This is something new to us and would be really grateful if the wonderful users here could shed some further light on how it usually works (we just need a way of tracking the unit's movements)?

We are currently short on our demo units as they are with reviewers but we can consider this seriously upon their return.

Thank you for the input!

 

[webster]

This is something new to us and would be really grateful if the wonderful users here could shed some further light on how it usually works (we just need a way of tracking the units movements)?

That would be somebody other than me as although I've suggested it I've never actually participated before.
I'm sure there will be some interest along with expertise as to how it would work from other members though.

 

[StackAudio]

Hi Stack Audio. You say you believe “that a digital streamer set up does make a strong difference”. Since you actually make and sell a digital streamer, the Link II, can you actually back up this claim with any measurement at the output of any DAC that shows your streamer makes a measurable difference?

Hello AndyU!

We are currently working on a white paper with John W to show the measures and figures based on the network bridge's processes and outputs. However, this is still in the works as we want it to be ready for analysis from other audio engineers.

Having said this, we do have an image depicting the phase noise with and without the detox treatment on the original LINK here:

detox by Theo Stack, on Flickr

Sorry we can't provide more than this at the moment but we will be more than happy to share the document as and when it is finalised!

 

[Yomanze]

The question then is: what would it be in that digital noise that potentially impacts sonics? The digital signal itself looks the same, just surrounded by less noise, and both look like they'd be interpreted in the same way at the digital receiver end.

 

[AndyU]

Hello AndyU!

We are currently working on a white paper with John W to show the measures and figures based on the network bridge's processes and outputs. However, this is still in the works

Wouldn’t be the only thing that JohnW has “in the works” so I’d best not hold my breath!

 

[John Phillips]

The question then is: what would it be in that digital noise that potentially impacts sonics? The digital signal itself looks the same, just surrounded by less noise, and both look like they'd be interpreted in the same way at the digital receiver end.

Yes. This is exactly the question that seems always to be left unaddressed in promotion of digital-side upgrades that are supposed to improve the audio side. How does phase noise on the digital side of a DAC (and the other noise sources that also exist there) get onto the analogue side and at what level? And why can't the DAC at least in principle reduce the noise to an inaudible level?

It certainly seems possible for a DAC to stop the various noise elements of a standard Raspberry Pi USB output making it to audible levels on its analogue output. I was curious about this issue and I threw various digitally-simulated unwanted signals at my DAC via its asynchronous USB input from a R.Pi and not yet found anything audible even at maximum audio gain.

There may be other DACs that deal less than well with digital-side noise, of course. I would be interested in anything credible that is amenable to a reasonable test.

 

[beammeup]

Hello AndyU!

We are currently working on a white paper with John W to show the measures and figures based on the network bridge's processes and outputs. However, this is still in the works as we want it to be ready for analysis from other audio engineers.

Having said this, we do have an image depicting the phase noise with and without the detox treatment on the original LINK here:

detox by Theo Stack, on Flickr

Sorry we can't provide more than this at the moment but we will be more than happy to share the document as and when it is finalised!

Isn't your Stack Audio Link based on the Pro-Ject Streambox S2 Ultra?

I own this model and it would be good to put the Streambox up against your Link if you are thinking about allowing peeps to audition your box. I also have the sMS-200 Ultra for comparison.

 

[Tw99]

Yes, I am of the strong belief that a digital streamer set up does make a significant difference.

It is especially more noticeable when playing your favourite music from a program (and server) that is processor hungry (can add far more interference to the output than anyone would want!).

I'm of the strong belief, which is backed by understanding the technology and looking at relevant measurements, that a digital streamer doesn't make a significant difference. I.e. you can get the same results from a £50 Rpi as you can from a £2000 audiophile voodoo box.

Your comment about downsides of playing from a program that is "processor hungry" and can "add interference" is just technobabble. As long as there's enough processor power to get the bits out of the relevant interface, it doesn't matter what else the processor is doing. Furthermore, if the processor is doing less, that means it's spending more time hopping between power levels, internal clock speeds etc. It would be equally easy to make up a story about that being bad and resulting in "muddled sound" or some other bollocks.

 

[davidjt]

I was curious about this issue and I threw various digitally-simulated unwanted signals at my DAC via its asynchronous USB input from a R.Pi and not yet found anything audible even at maximum audio gain.

I'm not an expert, but as I understand the process it's not a question of removing audible noise (there won't be any) but of achieving a lower noise floor and thereby allowing improved clarity and precision and revealing low level detail. If even I can hear the difference a filter makes then surely anyone can. (Although I wouldn't expect a £X000 purpose-built media server to need one - I'm using a s/h HP laptop.)

 

[deebster]

Hello webster.

The idea of circulating a demo unit to the users of PFM is something we would certainly be open to.

This is something new to us and would be really grateful if the wonderful users here could shed some further light on how it usually works (we just need a way of tracking the unit's movements)?

Something like this in my experience: the owner starts a 'tour' or other suitably-named thread, fishies add their names to the list and it is then sent from one to the next by Royal Mail Special Delivery with thread posts or private messages to the owner saying who has it and where it is then off to.

Seemed to work very well on the one I joined in on.

 

[sq225917]

A digital signal is either perfectly recoverable or not, there's no level of quality about it. Bit perfect or not, there's no middle ground.

That said, depending on the quality of the connected device, noise on the digital lines, and any connected power rails, could find their way into the analogue domain through secondary effects.


The only valid test is to measure the analogue output of the dac with and without the DUT in the loop.

 

[gustav_errata]

How does phase noise on the digital side of a DAC (and the other noise sources that also exist there) get onto the analogue side and at what level? And why can't the DAC at least in principle reduce the noise to an inaudible level?

I'm not an expert but shouldn't it result in flipped bits? Isynchronous and asynchronous USB transfer both lack error correction, so those bits will be passed on to the DAC chip, which will then misencode that sample on the analogue end.

To which extent that is audible, I have no idea. As for how common it is, it should be higher with faster asynchronous transfer rates, but it's probably more dependent on the cable length.

 

[webster]

Something like this in my experience: the owner starts a 'tour' or other suitably-named thread, fishies add their names to the list and it is then sent from one to the next by Royal Mail Special Delivery with thread posts or private messages to the owner saying who has it and where it is then off to.

Seemed to work very well on the one I joined in on.

Presumably each 'fishie' coughs up his or her postage cost?

 

[Darth Vader]

I'm not an expert but shouldn't it result in flipped bits? Isynchronous and asynchronous USB transfer both lack error correction, so those bits will be passed on to the DAC chip, which will then misencode that sample on the analogue end.

To which extent that is audible, I have no idea. As for how common it is, it should be higher with faster asynchronous transfer rates, but it's probably more dependent on the cable length.

No. If interested look up PCM and NRZ to see how the data goes down the cable from music player to a DAC.

Cheers,

DV

 

[gustav_errata]

No. If interested look up PCM and NRZ to see how the data goes down the cable from music player to a DAC.

Thanks. I'm familiar with PCM but I wasn't familiar with NRZ.

According to Wikipedia's info on NRZI (used in USB), "The two-level NRZI signal distinguishes data bits by the presence or absence of a transition at a clock boundary.". So if there's phase noise that delays the detection of a transition at the clock boundary, how would that manifest?

 

[Olsen]

I have the jaundiced view...
Er... precisely because he only sold turntables and wanted as much of the pie as possible!

Actually if memory serve, I think he sold Isobariks too.
Until launch of LK1/2 mid 80'ies, Linn promoted Naim amplication.

But truth some manufacturers are doing their marketing as such, Harbeth another example, just speakers are their vote for most importance item in chain, some people are very sensitive to marketing, becoming many disciples out there.

 

[Tw99]

Thanks. I'm familiar with PCM but I wasn't familiar with NRZ.

According to Wikipedia's info on NRZI (used in USB), "The two-level NRZI signal distinguishes data bits by the presence or absence of a transition at a clock boundary.". So if there's phase noise that delays the detection of a transition at the clock boundary, how would that manifest?

If you have a decent clock recovery circuit on the receiver then it's totally irrelevant.

 

[EPear]

I'm of the strong belief, which is backed by understanding the technology and looking at relevant measurements, that a digital streamer doesn't make a significant difference. I.e. you can get the same results from a £50 Rpi as you can from a £2000 audiophile voodoo box.

Your comment about downsides of playing from a program that is "processor hungry" and can "add interference" is just technobabble. As long as there's enough processor power to get the bits out of the relevant interface, it doesn't matter what else the processor is doing. Furthermore, if the processor is doing less, that means it's spending more time hopping between power levels, internal clock speeds etc. It would be equally easy to make up a story about that being bad and resulting in "muddled sound" or some other bollocks.

well, I am of no strong belief whatsoever but have had several streamers and streaming solutions, from cheap to voodoo boxes as you call them, they sound different so I ended up with the best sounding ones. so good luck with whatever measurements and reading that makes you happy

 

[beammeup]

Hello AndyU!

We are currently working on a white paper with John W to show the measures and figures based on the network bridge's processes and outputs. However, this is still in the works as we want it to be ready for analysis from other audio engineers.

Sorry we can't provide more than this at the moment but we will be more than happy to share the document as and when it is finalised!

...also just looking at your website - I found a reference on there for an alternative to BubbleUPnP found here: https://stackaudio.co.uk/softwareplatform/ - mconnect player - you mentioned one of its key features: 'Internet Radio' - whereabouts do you do that on mconnect? Or is the info you provided wrong? I've got the paid for version.