FORGETTING the Digital to Analog conversion part, what is BEST Digital source?

[asdf1000]

23 hours ago, barrows said:

suspect Beer may actually have some settings in his computer which are compromising his USB performance, this happens pretty easily, especially on Windows.  

 

Hi barrows

 

Can you share some high level tips, for us Windows users, using USB?

 

If I set my Windows playback as the laptop speakers, but in Roon (for example only) I use the ASIO driver for my USB DAC then is that sufficient?

 

This way the default Windows sound engine is not crossing paths with Roon's ASIO selection?

 

 

[asdf1000]

7 hours ago, barrows said:

i see you are using ROON, OK.  ROON has a lot of oversampling options, you might want to play with those some.

 

Hi barrows. Sorry I don't use Roon's up-sampling, I let my DirectStream up-sample to 20x DSD rate.

 

I thought you were hinting at optimizing Windows settings for USB audio playback earlier.

 

[asdf1000]

26 minutes ago, lucretius said:

 

20x -- that would be like DSD1284 

 

 

Hehe yeh it's something like that. I think some of the Chord DAC's FPGA go to even higher sample rates (2048 x original sample rate)

[asdf1000]

21 hours ago, barrows said:

Hi Mark,

 

Here are some things to ponder:

 

1.  The quality and implementation of the Ethernet Renderer has a huge impact on sound quality.  How and why this is is beyond the scope of forum posts, but very many details of the design and implementation of an Ethernet Renderer have implications for ultimate sound quality.  Most DAC manufacturers who add an onboard Renderer are looking for a very cost effective solution, but a cost effective solution is not the one which will sound best.  

 

2.  An Ethernet Renderer is essentially a small computer, with all the noise implications which that has, a high speed, fairly powerful processor, etc.  Putting this is in the same chassis with sensitive clock circuits, and analog electronics is a very challenging thing to do.  The noise from the Ethernet Renderer can easily couple into parts of the circuit where it should not be.  Many of the advances in current DAC design has to do with reducing processor noise in proximity to the DAC, clock, and analog circuitry: putting an Ethernet Renderer inside a DAC chassis is in direct opposition in trying to reduce noise inside the DAC.

 

I have heard only one DAC with built in Ethernet which really sounds great: the top level Linn Klimax DS, these go for over $20K or so.  If you look at the pictures of the internal construction of these you can see the lengths to which Linn goes to isolate (both physically and electrically) the Ethernet Renderer portion from the DAC and analog section of this component:  The entire chassis is milled from block aluminum, and thick walls shield the DAC/analog section from the input and power supply section.  On the Linn, Ethernet is the only input, so they can concentrate all of their engineering and budget on getting this right, and still the DAC is still quite expensive.

 

BTW folks: if anyone wants to compare the PS Audio DirectStream DAC via the onboard Ethernet "Bridge" versus using an external Ethernet Renderer and USB, be sure to remove the "Bridge" when trying the external Renderers as this will remove it as a noise source.  If the Bridge is still in the DAC it will be powered up and still adding its own noise to the internals of the DAC.

 

A lot of built in Ethernet renderers have sample rate restrictions as well, most I have heard of will still only do up to 24/192 and DSD 128, there are good external Renderers available that have no such limitations and can do PCM up to 32/768 and DSD 512.  This is very important for those who may be interested in oversampling in software. 

 

 

A really great post. 

 

I've observed the exact same things. I don't think we are (yet) at the point where seeing an ethernet input on a DAC automatically means that is the best sounding input. It may but it may not. You still need to listen.

 

The DirectStream DAC is a great example I found. Of course many others say the Bridge II sounds best and they are not wrong. There's no wrong answer.

 

Obviously this is just talking about SQ (for all the reasons discussed by barrows) and putting aside the huge convenience an ethernet input DAC brings in terms of the spaghetti solution.

 

But this is changing as mentioned above - more attention is being paid to the ethernet physical interface. The next year and two will be interesting to see developments with ethernet interfaces.

 

 

[asdf1000]

15 minutes ago, lmitche said:

Barrows, yes the microrendu and ultrarendus sound very good, but so does a custom configured directly connected upsampling machine like the SGM.

 

But do you really want a powerful upsampling machine so close to your DAC (and other RF/EMI sensitive HiFi gear)?

 

Isn't that the main advantage of low powered, low noise endpoints like the rendu's?

 

They aren't doing any decoding (with Roon or Audirvana anyway) close to the DAC - they just play PCM and DSD with the decoding done further upstream (in a different room ideally)?

 

[asdf1000]

3 minutes ago, beerandmusic said:

"ideally", native dsd streamed to a streamer/dac with no need for interface or cables between the 2.

 

Wireless is not necessarily the automatic ideal - again RF/EMI may (or may not) be a factor around your DAC and other sensitive gear.

 

Also, I'm talking about a transmitter in/on your DAC - not the WiFi router in another room.

 

[asdf1000]

3 minutes ago, beerandmusic said:

agree there...i would do wired enet...i never suggested wireless.

 

Whoops, sorry I mis-read. You want a streamer and DAC combo. My bad

[asdf1000]

22 minutes ago, beerandmusic said:

 

Ok...sounds great....license the card to HOLO SPRING so they can sport an enet port!

Get same developers to create a module to support ALEXA (wink)

 

Ayre Acoustics and PS Audio (and another major DAC manufacturer whose name escapes me) use these ethernet to I2S cards...

 

http://www.conversdigital.com/kor/product/product01.php

 

As barrows pointed out earlier (and I agreed), just seeing an ethernet port on a DAC doesn't automatically make it the best sounding input. It may be the most convenient for sure, but you still need to listen to hear if it's the best sounding input, for that DAC.

 

There are many Ayre Acoustics and DirectStream DAC owners that think the USB input sounds better than their DAC's ethernet input... and many who think vice versa of course.

 

I think the ethernet physical interface will be improving quite a lot over the next year and two.

 

[asdf1000]

15 minutes ago, barrows said:

Yes, the Convers board is an option available now at a very low price, unfortunately performance of it is not SOTA.

 

The board itself is one thing but it's implementation is another thing (as you pointed out in one of your earlier posts).

 

In the case of the DirectStream:

 

"The USB draws a constant amount of current + some current that depends on the sampling rate.

The Bridge is much more complicated in that there are many kinds of inputs possible with differing compute loads, think of it like a small PC.

On the other hand the DS’s power supply is really 4 separate regulators on two separate transformer windings, the bridge uses a separate regulator from the analog board…"

http://www.psaudio.com/forum/directstream-all-about-it/bridge-ii-power-consumption/#p78284

 

Best to listen and leave it to your ears to determine which is your favourite input.

 

 

[asdf1000]

1 minute ago, beerandmusic said:

but i just found doing native dsd over dlna sounded better.........

 

Yup, you listened and compared both inputs and preferred one. That's the way to do it.

 

4 minutes ago, beerandmusic said:

omparing native DSD to sony bluray player via DLNA vs same file sent out my pc usb to any usb dac (including the likes of a schiit multibit bifrost

 

I didn't think schiit DACs supported DSD?

 

[asdf1000]

4 hours ago, marce said:

how do you mean improving?

 

High sample rate support - there aren’t that many DACs with ethernet input that support DSD256 and PCM384kHz (or higher) via ethernet and do it well.

 

This will improve quickly. One example is the Zman board which has other features, in addition to DSD256. There would be others and custom solutions being worked on too.

 

[asdf1000]

23 minutes ago, Ralf11 said:

I've always wondered why optical isn't used more...

 

Actually if you ask the designers of the DirectStream DAC and Chord Dave / Hugo etc, TOSlink is their favourite input for music content <192kHz and DSD64 (via DoP)

 

TOSlink is effectively immune to RFI, is perfectly isolated from the source and they say their DAC's (above) effectively eliminate jitter.

 

TOSlink obviously doesn't support >192kHz and DSD128 and over.

 

Fiber input (over the network) DACs would be cool. Playback Designs have had a fiber input for a long while.

[asdf1000]

42 minutes ago, Ralf11 said:

I want better optical than TOSlink

 

You can have that right now.

 

Buy a USD6500 Playback Designs Merlot DAC and their USD2500 USB-XIII Audio Interface. The USB-XIII supports DSD256 over the network.

 

Or their USD6500 Sonoma Syrah music server if you want to skip USB and just have server fiber output -> DAC's fiber input

 

[asdf1000]

1 hour ago, sandyk said:

 

 A mediocre clock performs much better when  supplied by far better isolated power supplies, which Gryphon Kalliope  DAC appears to have in abundance.

 

An underwhelming review, where it needed a USB Regen to really shine !

 https://www.stereo.net.au/forums/topic/117214-mola-mola-dac/

 

 

 

A great read, even if it's just one (or a couple) person's subjective observations/experiences. As barrows mentioned, the comment by the OP about the DirectStream DAC's USB input being galvanically isolated is incorrect.

 

The Merlot DAC that's mentioned in that thread, is something special indeed (I've heard it directly driving my HD800S headphones).

 

I have a DirectStream DAC and Hugo2 and have heard the Chord Dave - all claim that inputs sound the same but many owners of those DACs have shared preferences of certain inputs over others.

 

The Merlot DAC was the first time I really heard all DAC inputs sounding the same, and just like the OP of that thread mentioned, the Regen didn't make an obvious difference to my ears, with that DAC. The Merlot's USB input is not isolated either. I'm just one data point of course, other may have difference exerperiences.

 

Apologies to the OP - I've gone off topic.

[asdf1000]

On 18/01/2018 at 10:37 AM, barrows said:

Anyone who has compared the sound quality of the microRendu to the (rare) 1.4 version has heard this difference as well, it is well documented.

 

Hi barrows. What did you mean by "rare" here. The v1.4 upgrade is still available, no?

[asdf1000]

2 hours ago, *progear said:

And this is why: a couple of months ago a Dante module for the Lynx Hilo arrived at the studio. The Hilo so far had been fed via USB, not so much because we believed in USB superiority over Thunderbolt, but because we wanted the DAC to work with any computer customers could possibly bring along. With the Dante module installed, the Hilo sounded an awful lot better. This made us even more interested in Dante, so we ordered some inexpensive boards from Swiss Micromedia

 

Hi thanks for sharing some info from the Pro side.

 

Is there any technical reason why you didn't just add the Lynx LT-DANTE LSlot to the Hilo? And use the Hilo's existing outputs? 

 

https://www.lynxstudio.com/products/lslot/lynx-lt-dante/

 

Or just cost reasons?

 

And how exactly did you integrate the Micromedia boards with the Hilo?

 

Cheers!

 

 

[asdf1000]

2 minutes ago, *progear said:

As I said above, these things are cheap and they blow the best USB converters I know out of the water.

 

Thanks, yes that was noted.

 

Which of the better USB converters have you used/tried, if you don't mind me asking?

[asdf1000]

9 hours ago, *progear said:

Native audio formats with dante boards are TDM and I2S. In two-channel mode the dante device speaks native I2S; if built into the dac or linked via slot card there will be no conversions whatsoever. The shorter the signal path for unbalanced I2S, the better the sound. My theory.

 

Yep there's no such thing as a free lunch and adding more stuff in the chain may have pro's but also con's (trade-offs). The audible effects of any trade-offs will vary with every system of course.

 

So ethernet input DACs (including networked amps and networked active speakers) have great potential and there are nice ones already out there and it's safe to assume numbers will grow and over time USB will become a legacy input.

 

But it was discussed earlier in this thread that ethernet-to-I2S interfaces can be tricky to implement well inside a DAC (power supply noise isolation, phase noise/jitter, RF isolation) - especially with high sample rates like many around here use (DSD256 and DSD512). These Dante ethernet-to-I2S interfaces only have to deal with up to PCM96kHz I believe?

 

I was quite close to going the Dante route last year but held back after hearing about Merging's Zman board. Still waiting to hear if/how that develops.

 

I've heard great things about the Lynx Hilo as a Dante networked DAC.