USB Nervosa Thread Decrapifiers, pro interfaces, and bears oh my

I don't think it's perfectly clear cut that AES is superior to SPDIF - SPDIF using coax has the advantage that it's immune to skin effect, AES has the advantage of much higher voltages, differential signal. From what limited understanding I have gathered, the on-paper jitter of USB can be below that of SPDIF/AES. Most of the spdif receivers I have come across specify 50ps. The best of USB I have found specifies 10ps. Of course this is just one metric on paper. USB apparently has noise issues AES does not have to deal with. This could be part of the issue with not being designed for audio.

I would also suggest that the specific design of the DAC (and computer) at hand will have quite an influence. Some DAC are more sensitive to jitter. Some are more sensitive to noise. Some DAC's have good SPDIF inputs. Some don't. Some DAC's reclock every digital input. Some don't. Some DAC's have good USB inputs. Some don't. Some computers have good USB output. Some don't. Lastly not all ways of getting AES out of a computer are equal. AES cards must draw power from motherboard power rails, so hopefully they have very good PSRR on the onboard regulators. External AES interface have advantage of possible better power supply but must deal with another inteface to get the musics out of the computer.

I think this is a big source of confusion, and I'm not sure how well anyone will ever prgoress towards resolving or forming a good understanding which can be shared and agreed by others. We all have to rationalise based on limited information. I am likely biased towards USB given my experience. Others are probably biased the other way. A last even more problematic aspect is that at any time aspects of performance may be visible or invisible depending on system resolution, or even good or bad depending on system resolution, synergy etc. I have completely reversed my opinion on a number of components over time as my system evolves...

All part of the fun but sometimes it does my head in.

 

I can state with certainty, that in my system AES3 is superior to SPDIF and by an ever widening margin, and both of them leave my 'tweaked' implementation of USB far behind in terms of SQ.
This isn't based upon specs or theory or mulling over the various technical aspects, but by listening, to music, music I am very familiar with.

And after delving into a bit of the history and origins of the generation of the specs and the intended design of these 2 designed for digital audio formats, it has become clear(er) to me that SPDIF is aimed at the consumer level of operation and performance whereas AES3 is the full meal deal.

And granted my experience is based upon using a RedNet box as the source of these 2 formats and not using a computer as the connection to the DAC.
In fact I have gone one step farther in completely electrically isolating the computer from the generation of the digital audio stream that feeds my dac.
So this conclusion is 'biased' in that it is based upon a single mode of operation, even so the differences are immediately apparent and obvious, even on cheap $40 speakers, let alone tweaked HD800's

But since the RedNet box is a ProAudio device it doesn't focus upon the consumer level of operation, nor the ease of use expectations inherent with this level of performance, but it does provide SPDIF connectivity and at this same level of ProAudio execution.

And further, the use of SE vs Balanced for these 2 formats along with their requisite connectors, for me just further cements the perspective of 'doing it right' vs. 'doing just enough to get the job done' (which usually translates to being cheaper, a major design factor in consumer level gear).

JJ

 

Would you expound on this please, very interested!

 

The RedNet ethernet boxes from Focusrite allows a ProAudio multi channel distribution system to be created using only etherenet cable instead of snakes and the like.

'We' have begun using them to replace all of the methods of sending digital audio to our dacs.
USB, SPDIF, (coax and optical) these data streams, when sent from a computer, tend to get 'polluted' by all the noise and jitter that resides in those electrical environments.

These ethernet to SPDIF, AES3, (our primary digital audio formats) converter boxes are vastly more capable than our mere 2 channel needs, but the SQ is also vastly superior to all of our previous experiments using USB, or SPDIF, or TOSLINK etc.

And one of the advantages is the use of ethernet ports, cables etc, which are galvanicly isolated, well at least the signals are.
What was suggested, and I and others have implemented, was to use even further isolation by employing an ethernet to optical to ethernet 'bridge'.

This in effect completely electrically separates one end of the ethernet feed, the noisy computer/player/streamer, from the other end, in this case the RedNet box.
And by using an optical cable in the middle, ALL electrical interconnections are severed.
This complete galvanic isolation is a step beyond just relying upon the transformer coupled isolation employed in the ethernet design as there are no wires tying one end to the other.

JJ

 

Thanks JJ, I appreciate the information greatly! What are you using currently for the optical bridge? I'll read up on it...

-Ron

 

here is the link to the devices
http://www.tp-link.us/products/details/cat-4792_MC210CS.html
and amazon's info page
https://www.amazon.com/gp/product/B003CFATKQ/ref=od_aui_detailpages00?ie=UTF8&psc=1

The whole setup (2 transceivers and the optical cable) runs under $150.

Do read up on these systems and how to properly implement them.
In short you need to match up the ethernet cards capability with the transceivers.

JJ

 

I too used to think that using AES was better than using USB directly, as I used to use an external USB/AES converter, I actually went through quite a few of them with the last one being the Reflink which, a short time ago, was second only to the Berkely BADA USB as far as USB/AES converters went. Actually I think Berkely also preferred AES, and similarly Wireworld suggested I use their AES cable over SPDIF. That was quite an investment and I have recovered a small fraction reselling those products. I have yet to sell the Reflink, but for reference it bested Audiophilleo AP2 with the purepower upgrade, again something from 3+ years ago many have probably forgotten or never heard of. Take this as some indication of how fast things move in digital audio, and especially digital inputs. But yeah the Reflink is IME very good D/D converter in it's own right, and at the time of purchase only behind a more expensive converter.

In any case, even with very good USB/AES converters and AES cabling, the USB32 USB/I2S module in my DAC, which BTW had issues with IMD, was much better than the Reflink with Wireworld Platinum AES cable (according to others also better than BADA USB). This with even mediocre computer sources such as my laptop. OK that was has been my subjective experience. The difficulty from there is to attempt to develop theories of causal relationships. Similarly I have tried using USB/AES with a Metrum Hex DAC comparing to various other inputs such as optical and USB (Amanero). In that case again USB presented a more full bodied, detailed and nuanced result.

Optical on the Hex was quite unexpected result, very lean and "detailed" in macros sense but lacking in body and no more resolving than USB. At varying levels of resolution and system synergy that leanness may be welcomed as a step forward. Mostly I find aggressive treble to be recording (and transducer, and software, and amplifer etc etc) dependent - could well be that the aggressiveness is already there to begin with, and is only being revealed for what it is. Also Hex had this quality of purity of tone in the treble, and may be unperturbed by aggressive treble regardless of it's origin or cause.

Could be that Rednet is a different beast to USB/AES, but until someone wants to post me one to demo I can't know... That would also be something I would not be willing to pay for given my preference for inside-DAC and not external D/D boxes. Maybe someone local will buy one...

In any case, I think I could safely speculate that where all things equal good quality direct USB outperforms even very expensive USB/AES *where there is no special reclocking after the digital inputs inside the DAC. Could be that external USB/AES interfaces combines the worst attributes of both, but I would tend to doubt that based on my limited experience and understanding. I would speculate further than other interconnect means such as LVDSI2S are less restrictive and may in some cases allow some of the design freedoms of external interfaces such as more room for power supplies and noise isolation to offset the fact that jitter is only ever going to increase after a connection like USB or AES67 (without reclocking), and might allow for more flexibility in connection, for example a high end SD card transport using FPGA, or a compatible CD player, although those all seem to use SPDIF or AES or usually both. I am also wary of too many instances of reclocking, I have played around with that and it can be really nice to use but is usually always a tradeoff - just matter of if the net result is subjectively positive in a given system.

"Hurdles that need to be overcome" I see with AES67 are that if the computer only has one NIC, that same I/O controller will need to handle both the internet as well as the streaming of audio. I imagine it would be a bit like throwing your mouse and keyboard on the same USB controller as your DAC, ie likely to create potential dropouts and challenges to prevent buffering issues. With a computer with two NIC's or say the dedicated Rednet PCIE card I would not expect that kind of problem, but this would be something not a lot of audiophiles etc would be comfortable with. Luckily my server has two NIC's and spare PCIE slots but the case only allows half height cards. A lot of NIC and USB cards are half height thankfully, but it seems most of the stuff targeted at pro audio assumes use of normal size computer cases that allow full height PCIE eg what you expect with DAW. On my workstation that would be a different story, but my workstation is not as high performance for audio as my dedicated DIY music server which is admittedly heavily geared towards USB.

@johnjen you are correct that comparing audio interfaces in terms of measurements is very problematic, eg it seems that firmware and software play a big part, but this does not always show up in specs or measurements, depending how they are performed. I refer to the specs to attempt to rationalise what I hear. My intuition is that at least someone must understand whats going on eg the people that develop the firmware...

 

What we have been experiencing is that by not using USB, at all, tends to leap frog beyond even the best that USB can offer in any and all possible configurations we have tried.
And some have gone WAY deep into a wide range of possibilities, and in multiple combinations, trying to see 'how high is up', WRT USB.

However these many tweaks do significantly improve the SQ that USB can offer.
In my case even my 2-Wyrd setup was a significant step up, and for some this is more than enough.
But not for all.

And really what we are talking about here is a wire, a way of sending digital signals from one box to another.
And right now the 'best' wire costs $1K+.
I suspect that price will drop, but when and by how much is an unknown.

And this is a changing landscape, for sure, as these experiments are ongoing and further refinements are still being explored and tested.
And the entrance into this whole arena entails a steep learning curve and is costly, so for many it is 'wise' to wait and see.
And this RedNet system does make demands upon the upstream h/w, let alone if you want to tweak it even further.

But for those who are willing and able to pursue exploring this $1K wire, the results are most gratifying.

As for the variability of s/w, I have lived thru multiple updates in both f/w and s/w on my PSA PWD.
It can be frustrating as the SQ may improve but functionality drops off, or visa versa.

IOW there are no free lunches…


JJ

 

I think the price of external converters complete with power supplies, case etc is fairly constant, and likely there is some overhead costs for developing new transmission protocols. I think once it gets to the stage of AES67 receiver cards which outputs I2S and can be installed inside the DAC (cheapest & simplest), or in an external box sending LVDSI2S or AES to the DAC that would open up more possibilities, or even when we start to see more typical sources supporting dual NIC and more options for PCIE NIC specialised for this application in half height form factors that can be integrated into custom made music servers. I seem to recall the developer of Regen was working on some sort of NIC/RJ45 based interconnect method as well. The hope with AES67 is that there will be more possibility for interoperability and standardisation to avoid the mess we have with for example LVDSI2S, as used by PSA in the PWD and their transport (now also adopted by AudioGD).

You are right definitely no free lunches. For exmple with PSA LVDSI2S it is designed for very short distances, eg between dedicated hifi devices and may well not be the optimal solution for getting music directly out of computers. I recall there is HDMI connector based I2S PCIE card but it uses TTL signal as they think that is better...

In fact that is the main danger that is faced by whichever standard will replace USB - should learn from the mess surrounding I2S interconnect standards. Nothing short of total end to end compatibility is needed. Any sort of incompatibility or limited compatibility will surely discourage adoption from manufacturers, and it will likely end up just like LVDSI2S.

 

I bought a Yggdrasil DAC a few weeks ago. As an experiment I connected a $30 Sony Blu ray via a 20 year old $30 Wireworld digital coax. I fully expected that the "your gear sucks ballz" light on the Yggdrasil would come on but it did not. I thought it sounded better than USB from a Mac Mini with a Regen in the chain playing .wav files in Audirvana +. The improvement was not night and day but still pretty noticeable.

Most of my CDs are in boxes in the garage. I don't plan on spinning discs again to get this improvement. I'm also not going to spend ~$1500 on a chain of gizmos to take USB performance up a few notches. There is a lot of discovery and change going on in this area right now and hopefully in a year there will be a single box solution that "fixes" USB for an affordable amount. I'm going to wait this out and put my money towards a new amp or headphones in the mean time.

 

Hasn't anyone made a single box ethernet solution yet? Something that plugs directly into the ethernet jack of your computer and pumps out an analog signal.

 

That is the Rednet people have been crazy about.

 

Wow, what a busy thread this has been.

Everyone needs a reference, I think or you are constantly chasing changing things. Marv has a great chart of seven axis on amps, and the criteria mostly applies or could be adapted. Improving dynamics may cause soundstaging or stage depth to vary, clearer mids thins bass, etc. A Berkeley Alpha USB converter could have a different pattern from the Mutec on any given DAC/Amp/Reproduction chain. If the DAR chain is not revealing, it won't come through, so will seem like nothing is happening.

Also, we all agree newer dacs usually have better USB, but this doesn't mean the USB is always better, or worst for all dacs. Its probably not your imagination, given the high sales of USB->SPDIF converters, good find on the Toshiba player!

Schiit uses C-Media CM6632 for the Yggdrasil, and C-Media CM6631A for the Gungnir Multibit, I think due to limited availability at the time. And I still see people say the Yggdrasil is better SPDIF/AES. Plus Baldr just prefers AES.

Yeah USB sounds good, but SPDIF can be better, i totally concur.

Thats a lot to spend for a wire, but if it brings you home, I may disagree (and I don't) but i totally support your right to decide for yourself its worth it to you. I hope that we get to a better place, but that's not always a given. SPDIF and USB have its sound quality merits, and I actually like USB with the Wyred4Sound on the Gungnir Multibit.

Try borrowing or stealing another Digital cable, and see if it changes the sound. It may or may not with your setup.

The problem with spinners is variation from model to model, and unit to unit, even. If you find a good one, keep it. And they will break down over time.

 

Pardon my ignorance, but isn't it Spdif or Aes out, not analog?

 

Model # of that Sony or you didn't hear it

 

Anyone used this lil Wyred 4 Sound USB reclocker. Looks like an inductory price of just $179. Could this and the Mutec 1.2 = a Mutec 3+USB?

 

Yep. Derp.

 

I did, sounds better than the Regen or any of the Mutecs on one set up I had. Almost as good as the Sonore Microrendu, but these are on a speaker setup.

 

It's less than a year old but I think it's the BDPS1500. So ok, I guess it was more like $40-$50.
Anyway, it was the basic model without all the built in apps and wifi stuff.

 

Uh, no, AES, like SPDIF is a digital audio format, not analog.

That's why they use 110Ω/75Ω cables due to the high frequency harmonics of the square wave signals being sent.

And while Rednet does make an ethernet to analog box as well we are mostly only using the SPDIF and AES outputs from the RN-3 and the D16 boxes.


JJ