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

and a DC version... Just say in..
Shouldnt the thread now be called 'The Yggdrasil A2 owners guide to spending money on distortion adding box's'...
A new thread one called "Noobs easy ways of saving money for the Mofet retirement fund.."

 

I got a Firestone Audio RedKey USB -> SPDIF converter dongle from Massdrop yesterday and listened a bit today. It was something like $20 or $25.

Here's what it looks like: http://imgur.com/a/e33Ij

It draws more power than the CCK on my iPhone 6+ can supply, but it works just fine from my 5k iMac.

Head to head against the Raspberry Pi2 streamer with Pi2Media SPD2 hat, it's... competitive, I guess. I haven't had a chance to listen as closely as I'd like, but I can say it's not embarrassed by the 5 or 6x-more expensive streamer. Different use cases between a network streamer and USB -> SPDIF converter, I know... but still, it seems to do pretty well in comparison. Could be a nice travel-friendly alternative to bigger SPDIF converters unless you're looking for that last 2% improvement, in which case give me some time to listen more closely and report back.

I don't have a more fancy USB -> SPDIF converter to compare against. And no, I'm not going to see if my Wyrd makes it better. I gots better things to do. :->

Only drawback so far as that the RedKey has this garish white LED behind a plastic panel that makes it look kind of like a high school science project.

 

By trade I'm an IT Project Manager (by training I'm an Anthropologist/Sociologist, but IT pays a good deal better ). Lately I've been reading about the Theory of Constraints. Basically, it holds that any system will be held back by a small number of constraints. The important thing in the Theory of Constraints is that at any one time, a single constraint is the one which will hold back the system. Releasing other constraints will have no impact on the system because they aren't the constraint (think bottleneck) limiting the system.

The implication is that you waste your time (and money) if you're not addressing the current constraint. After you elevate (clear) the system constraint, then you move on to the next limiting constraint. Etc.

In audio terms, you might get far better return on investment elsewhere in the audio chain unless you're already at end state for most of your audio chain

Rob

 

Funny, I just read through The Phoenix Project. Interesting analogy to audio.

 

Yet another good/bad(eye of the holder) analogy from madaboutaudio....

Soundsystem performance constraints are like car performance constraints?

Music Source Quality > Transport/Data Link > DAC > Amp > Tranducers
Fuel Quality > Fuel Pump/Air Filter Flow > Engine > Drivetrain > Tyres

 

Can anybody comment on the Bryston BUC-1 as an interface for removing jitter and converting USB to AES? I'm curious how it compares to the Mutec and the Wyrd (I realize that Wyrd is just USB).

I have my eye on the Yggdrasil and would like a USB decrapifier to go along with it.

 

I tried searching this thread, but could not find anything.

Any opinion on using a cheaper PCIe sound card to get SPDIF coax output from a computer and bypassing USB? Something like an ASUS Xonar card or something even cheaper just to get coax out.

I know a lot of people are talking about the pro audio PCIe cards with AES outputs.

What's the difference?

 

The Yggdrasil's USB input, which is bus-powered, still does noticeably improve with the iFi decrapifier trio (Mercury + iUSB3.0 + Gemini). I didn't bother using a USB-to-AES solution for Yggdrasil either.

Without resorting to the even crappier optical, I cannot avoid using USB altogether since my computer is an iMac... or so I thought until recently, when I realized there is an Ethernet-to-AES converter by Focusrite. Not sure if it's worth the high price tho.

 

If you just want toslink/spdif coax, you can try this:
https://www.amazon.com/ASUS-Sound-Essence-STX-II/dp/B00ONSBF4K

TCXO clocks:
http://gecid.com/data/mmedia/201504060800-37326/img/05_asus_essence_stx_ii_7_1.jpg

 

I have that STX II card.
I just tried it using the Moby-->Vali 2 --> Utopia:
There wasn't a huge difference between USB direct vs. Xonar coax. I will listen more, but I would say the coax was a little more detailed and brighter. Maybe more plankton.

Then moving on to Gungnir Multibit --> Rag --> Utopia:
The coax was slightly brighter, more detail, more plankton. But not a huge difference. I thought I could tell the difference in a blind test.
I had my daughter switch randomly and I failed the test miserably. So, maybe it's all in my imagination or my ears aren't good enough.

 

In my (admittedly foolish) attempt to find better USB audio performance, I came across this pretty interesting product:

https://www.amazon.com/Sonnet-Allegro-Pro-PCIe-card/dp/B00XPUHO10



The PCB board seem well designed and high quality. And Heat sinks on USB interface card! Maybe it might not improve usb audio but it should be faster for file copy between two usb harddrives. I'm gonna order one for my PC not for audio improvements but for data transfer improvements.

Also there's something else:
I guess equipment is used by people like @atomicbob :
https://www.amazon.com/Industrial-Grade-USB-Hub-4-Port-Isolation/dp/B00IPAHWYO
https://www.black-box.de/en-de/i/13513/Industrial-Grade-USB-Hub-with-Isolation

Intona looks like a plastic toy compare to this.

 

Rednet requires the use of licensed software (Audinate Dante Virtual Soundcard) to interface with the computer but if you held the permanent license for the eventual owner it looks like people could use the 14 day free trial for the loaner period. It's also decidedly not plug and play like most audio products are so that might turn some people off.

The thing that lured me to changstar and eventually here was your straight shooting about dacs in the infamous hf thread, it's such a refreshing change compared to the echo chamber you usually see in threads over there.

 

I agree, and the RedNets are not for the tech challenged.
And, at least for us in this hobby, this is bleeding edge territory to be sure.
Both the s/w and h/w are going to be going thru revisions as we (the hobby side) become a larger voice that both Focusrite and Audinate are going to deal with.

And of course there will be other gear that will come along which might make implementation easier and more plug-n-play.
But it will take some time to get to that level of sophistication.

But having said all of that, on suitably resolving systems once AOIP is heard for what it can do, USB in any guise or form is recognized for what it is, and what it does and doesn't do.

And many of us early adopters are still figuring out what some of the key aspects are in order to dial in all of the pertinent details to arrive at 'tweako' results.
And it seems that digital is just as susceptible to 'external influences' as analog can be, and perhaps even more so.
And if your aim is to get the most out of the potential that digital has to offer, well at this point the jury is still out as to what the parts list looks like, although it is forming up fairly quickly.

And to be sure the demands in terms of gear and resources just in order to get these RedNet boxes up and running will also act as a limiter in terms of who can actually provide the necessary 'environment' in order to be able to even consider trying AOIP in the first place.

Such is the nature of bleeding edge explorations.

JJ

 

OK, it's stupid question/theory time. Bear with me here, I'm trying to decrapify my understanding.

I get the essential nature of USB Audio streaming. There's, no error correction, no retry and it runs across a suboptimal bus that's often crappily implemented.

In contrast say we have AES, which is balanced SPDIF, which is apparently better than USB Audio.

So I build a device that takes this shitty USB Audio stream, decrapifies it and outputs a shiny new AES stream.

OR

Alternately I have my PCI(e) based card that does 'native' AES without the USB crap in the way.

Now my USB Decrapifier is almost as good, but not quite, because USB is not entirely decrapifiable. That Lynx card still has the edge. But sorry for me, I have a laptop with no PCIe slot, so I spend a small fortune on decrapification.

--- Given the above ---

USB IS good for file transfer. You can retry all day long and error correct and all that other good shit. And looky, all my music is just a bunch of files...say...if the issue is USB Audio Streaming, and the fact that all I have are USB ports, why don't I build a device that DOESN'T use USB Audio, but gets the bits via good old reliable file transfer via USB and generates a 'native' AES stream like the bloody Lynx card?

For that matter, how does a good DAP work? I imagine it's not shunting a USB Audio stream around internally. Something picks up a file and generates an I2S stream that gets fed to the DAC.

What am I missing? There must be a reason this device doesn't exist?

 

I have tried using portable DAP, tablets and phones as transport, all of them didn't sound as good as Pc USB.​

 

Yup, they're doing USB Audio streaming, and likely a noisier job of it. I'm talking about just a DAP, accessing files and feeding the bits to the internal DAC.

 

let me inject some hard data into this thread.

I performed a null test sending data out through the USB XMOS driver to a Gustard U12 DDC to SPDIF, and looped back to an RME Fireface UC SPDIF input then back through the RME USB driver. The DAW is able to send and receive data simultaneously from two different audio drivers. A start and stop pulse was inserted into the wavefile for data alignment. Data was 44 KHz 24 bit stereo channel decorrelated pink noise. Time aligned received samples were subtracted from samples sent. Results: not one single sample level out of place. Not one single bit error in 67,737,552 bits sent and received.

Please note that correct sample levels only implies all data sent was received correctly. It does not describe the timing of data presentation to a DAC (jitter). That is another story as is separation of digital from analog power / ground; noise imparted on the analog signal by the digital signal in a DAC, which may be factors in why some digital audio paths to a DAC sound different than others.

Here is the short version of USB -> SPDIF -> loopback -> SPDIF -> USB data stream accuracy.
Track 1 is source data sent
Track 2 is loopback received data
Track 3 is track 2 subtracted from track 1, the difference



Here is a very high y axis zoom on track 3, the difference. Not one sample difference.

 

That's a great test! So it says bits are just bits, which is fantastic, we can eliminate one type of variable. Assuming some real issues are lurking in the mists of confirmation (and other) bias present in the USB neurosis crowd, what's left is jitter and noise. Which are USB related, just not affecting bit accuracy. Certainly my Mojo behaves differently with the power cable plugged in - I can hear tiny artifacts in the white noise.

So it's a matter of galvanic isolation and reclocking, and we should be pretty much toe to toe with SPDIF. Some argue asynch USB with an isolator/reclocker is the best available right now - maybe they're right. And of course we need to take into account any coloration that the decrapifier introduces might be mistaken for an 'improvement" as well. The AOIP stuff isolates well of course, and explains why a Rednet & Mutec for reclocking cocktail seems to be the apex of the transport chain at the moment.

Do you mind sharing what computer you used, usb hub etc. in the test?