Wavelength Wavelink USB to SPDIF Converter

[barrows]

Note. I have found that the Wavelink seems somewhat sensitive to vibration. I think this falls into the "everything matters" category. I did a little quick damping to the inside of the box, and on the pcb itself, and used some aftermarket footers (Boston Audio "Tuneblocks"), and noticed a small level of improvement (yes, enough to notice). I suspect mass loading/damping the top of the box might be beneficial as well, but I need to find something nice to do this with, perhaps a nice slab of graphite. It also seems that the Wavelink is improving over time (as Gordon suggested it might), but as I find it difficult to go back in time to make a direct comparison, I cannot be sure of this.

One thing is for sure, my system is sounding fantastic, and I am listening to a lot of music. Thanks Gordon.

 

[mabe]

 

Hi Barrows,

 

first, thanks for sharing your comprehensive listening experience with the Wavelink.

 

I’m just wondering why Gordon doesn’t provide any further information re this converter. There are neither product/concept descriptions nor necessary specs public available yet.

Have you got any technical descriptions or product specification along with you Wavelink delivery? Just to bring some light in this darkness.

 

As you already mentioned, specific listening tests on specific HiFi-systems aren’t always transferable to other configurations and are sometimes of limited value. Therefore – at least for technical oriented potential customers – some common product information could be very beneficial.

 

Thanks,

 

Manfred

 

 

[barrows]

What kind of specs are you looking for? Most here are pretty familiar with how Gordon Rankin's "Streamlength" asynchronous USB code works. The only actual technical spec that would be relevant to sonic performance for a product like this would be jitter, and I suspect there is not a number published for this because there is no standard accepted measuring technique that would allow such a number to be comparable with jitter figures produced by other companies. Publishing a numerical value for jitter would really only confuse the issue.

I do suggest that those interested in a product listen to it, but I do not believe a product like this, which operates entirely in the digital realm (while I believe many things about audio are subjective, this is a product which is not really subject to that kind of evaluation-ie it either gets it "right" or "wrong"), is going to produce different results in different systems. All that matters for sonic perfromance in a USB-SPDIF converter (assuming it is not making gross errors, like dropping samples) is jitter and how it isolates the DAC from the computer. It is very clear to me that the Wavelink offers very low jitter performance, both from an understanding of how it works, and how it sounds. The SPDIF output is transformer coupled to provide the required isolation.

 

[aljordan]

dallasjustice wrote:

 

I am using an Antelope DA re-clocker between my Lynx card and NAD M2 which gives me galvanic isolation and allows me to use a high end cable like the MIT MA-X, which does make a huge difference on its own.

 

If you don't mind, could you check the Wavelink against the AES 16 without the reclocker in the mix and let me know your impressions?

 

Thanks,

Alan

 

[dallasjustice]

I didn't want to make that comparison primarily because I then would be comparing my Cardas Lynx card AES cable with the MIT MA-X SPDIF cable. The MIT Oracle MA-X DOES make a huge difference. I could try it later, but I am not really interested. If I had to guess, I would say the Lynx and Wavelink would be more equal without the reclocker. Antelope Isochrone DA really doesn't change things that much though. It really just provides some isolation between the computer and the DAC and lets me use a high-end AES cable.

 

I am sure this stuff is somewhat system dependent, so I don't want to make any generalization about the Wavelink. The Wavelink produces a nice sound, but it doesn't beat the Lynx card in my system.

 

[mabe]

Barrows,

 

 

'What kind of specs are you looking for? Most here are pretty familiar with how Gordon Rankin's "Streamlength" asynchronous USB code works.'

 

Most here are also pretty familiar with Refrigerators, but all Refrigerators I’m aware of are specified.

 

Many here are also pretty happy with CEntrance code based adaptive USB/SPDIF converters as well….

 

Ok, but, from the technical point of view I would expect some sonic and measurable advantages if the implementation of an asynch. USB/SPDIF converter is done right. But, how to figure out without any comprehensiv concept/product description nor specification … so at the end it can easily result ‘in good faith’.

 

Re: jitter – I really don’t expect a raw jitter number as it provides no valuable information for us at all. In our (audio) applications we learned/experienced that random phase noise (LF) generated i.e. from the oscillator (clock) and data-correlated noise/jitter may influence the sonic quality - depending on…you name it.

I.e.: by selecting/using a decent crystal oscillator the phase noise number (at defined bandwidth) could yield in a valid number we can believe in. – And we could compare it to other products – even w/o a standard accepted measuring method.

 

'It is very clear to me that the Wavelink offers very low jitter performance, both from an understanding of how it works, and how it sounds.'

 

Questions again:

- so how does it really work ?

- what does ‘very clear’ mean to you? ..based on...

 

Perhaps resulting in: …..

 

 

Specifications:

 

see part 2

 

 

 

[mabe]

 

Specifications: (example)

 

Sampling rates: 44.1, 48, 88.2, 96, 176.4 and 192KHz

 

Fixed frequency clocks: yes/no

 

Frequency accuracy: < xy ppm, < yz ppm, typical

 

Output: Transformer coupled, BNC connector, 75 ohms impedance

 

Output return loss:

0.1 MHz-100 MHz, < xy dB.

1 MHz-10 MHz, < yz dB

 

Output jitter: < x pSec (Note a) phase noise/jitter x Hz – 1 kHz

< y pSec 1kHz – 100kHz

< z pSec data – correlated noise/jitter (measured with …) or some equivalent qualitative information

 

Transfer type: USB, asynchronous

 

Driver: Thesycon USB Class 2 Audio driver

 

Power consumption: < y W

 

Voltage requirements: 120/240 VAC, 50/60 Hz (Via user-supplied IEC power cord.)

 

Weight:

 

Physical dimensions: X L x Y W x Z H.

 

Supplied accesories: BNC-BNC cable, BNC-RCA adaptor

Special damping tools to keep the crystal vibration low

 

Note a) measured with ….under …conditions

 

 

At the end of the day, I fully agree with you, the listening test at the specific HIFI environment will finally proof the performance and final decision criteria.

Prerequisite : availability of converter of choice.

Otherwise: good faith or comprehensive, comparable descriptions/specifications and last but not least some experience.

 

Manfred

 

 

[barrows]

I'll share what I know. Clear to me, because I know what to listen for in terms of jitter-I used to work for a high end audio company, and participated in the development of low jitter tranport and DAC products. I am used to comparing the sonic performance of components with lesser or greater measured jitter levels, and hence the general, audible effects of jitter.

 

Specifications:

 

Sampling rates: 44.1, 48, 88.2, 96, 176.4 and 192KHz:

 

Yes, it handles all these rates

 

Fixed frequency clocks: yes/no:

 

Yes, two fixed frequency oscillators

 

Output: Transformer coupled, BNC connector, 75 ohms impedance:

 

Yes

 

Transfer type: USB, asynchronous

 

Yes

 

Driver: Thesycon USB Class 2 Audio driver

 

Only for Windows, works with native drivers (USB 2) with Mac OSX 10.6.4

 

Voltage requirements: 120/240 VAC, 50/60 Hz (Via user-supplied IEC power cord.)

 

none, USB bus powered

 

Weight:

 

Light

 

Physical dimensions: X L x Y W x Z H:

 

14 cm x 11 cm x 6 cm, exclusive of connectors

 

 

Supplied accesories:

 

Wireworld Ultraviolet USB cable, Wireworld Ultraviolet BNC SPDIF cable, 1 BNC-RCA adapter

 

 

[mabe]

many thanks Barrows, may be Gordon can help with some missing informations.

 

[Wavelength]

Gang,

 

Really listening is so much better than specs. Again the problem here is what the receiver thinks!

 

Here are a few things... First return loss is a great spec if you graph it. But it is really useless to post numbers like this:

 

Output return loss:

0.1 MHz-100 MHz, < xy dB.

1 MHz-10 MHz, < yz dB

 

The return loss was measured by my good friend Pat at ART. He has my network analyzer as he knows how to use it better than me. The impedance was verified and return loss minimized.

 

Output jitter: < x pSec (Note a) phase noise/jitter x Hz – 1 kHz

< y pSec 1kHz – 100kHz

< z pSec data – correlated noise/jitter (measured with …) or some equivalent qualitative information

 

Jitter... for SPDIF there is actually two places you can measure this. If I set the Wavecrest for edge to edge I can get a pretty good idea as to the jitter. Look guys anyone talking about jitter in the 1KHz region or above knows nothing about jitter. Between 0-10Hz is were it makes audio go south. That is why looking at the spectrum noise of regulation with a really good spectrum analyzer is so important when design this stuff. Companies that make SPDIF receivers use 1KHz, that is because it looks like crap below that.

 

With the WaveCrest I usually get about 5-6ps max depending on how the thing acts that day. It's not really good down that low as it so close to it's minimum (400-800fS).

 

The other place to test this is at the receiver side and I can do that with the Prism dScope and it has been tested with the Ap as well. For some reason the Prism always seems to be 2x that of the Ap??? got me..

 

Prism is usually at 300-400ps and Ap was tested at 120-280ps depending on sample rate. EYE pattern tests for both the Prism and Ap look really good.

 

Frequency accuracy: < xy ppm, < yz ppm, typical

 

This is usually about 5.2ppm for the 22.5792 and about 5.8ppm for the 24.576. As you get higher in frequency so does the accuracy.

 

Power consumption is like 160ma off the VBUS line. There are a number of regulators for the processor, usb phy and SPDIF output and reclocker.

 

Really still this is all numbers! You are better off trying any product with your DAC as it could screw these numbers to all hell and very easily much up your sound to no return.

 

Thanks

Gordon

 

 

 

[ar-t]

Someone has recently asked us, a lot of questions, about this thread. I do not have time to read this thread, in detail, but let me address one issue.

 

I measure stuff like return loss and phase noise, because that is what I used to do, for a living. So, it is what I am most comfortable doing. We spec it, because we can quantify it, and are confident, in our performance. Simple as that.

 

Does it prove our stuff is better than the next guy's stuff? No, of course not. Just shows where our head is at, when we design and build stuff.

 

Likewise, when one of our buddies ask us, to quantify their gear, we do it. I do not disclose what we measure, for whom, and what the results are. That is for the benefit of our buddies. If they care to share it, that is their privilege.

 

Having said all of that, specs can be very misleading. Especially when it comes to a complex issue, like jitter. That subject is beyond the scope of this thread. Frankly, jitter specs, on digital audio stuff, are pretty much fluff. Even the ones we post. The devil is in the details. Sometimes, we like to tweak the devil. Take all of them with a grain of salt.

 

[dointhefish]

Got my Wavelink today. It sits between my Windows7 server running J. River and my Audio Research DAC7 with some Audioquest Cinnamon USB cable and AQ VDM-3 S/PDIF BNC-to-BNC. It replaces my Halide Bridge which I thought was awesome until I heard this. The Bridge renders lots of nice detail through the mids and highs but doesn't quite get the low end to my liking. Not so the Wavelink. It's a full spectrum experience. Lots of detail and impact all over. Shows just how great a piece of gear my AR DAC7 is. Voices have an in-the-room quality they didn't have before. Drum beats have more impact and cymbals crash beautifully. ECM recordings sound their absolute best with instrument separation more distinct than ever. Horns and saxes have more shimmer. Weather Report's Heavy Weather album sounds more 3 dimensional than I remember hearing before.

 

I started this journey of jitter reduction with the HiFace which sounded good but then upgraded to the EVO which sounded better. Then came the Halide Bridge which raised the bar even further. I thought about getting the Alpha USB when it becomes available but the Wavelength appears to have a portability advantage so I'll be taking this to work to see how it plays w/ my Lavry DA11.

 

Give its price to performance I think I've probably hit my limit ($900!) but the performance is right where I want it so I think I can relax with this for a while.

 

Andrew

 

Windows7->AQ Cinnamon USB->WaveLink->AQ VDM-3->AR DAC7->AQ Colorado XLR->Krell KAV-400i->Elac FS 607 X-JET

 

[firedog]

Just for accuracy, is 24/192 capable. The previous post is about the older model Off Ramp.

 

[57gold]

I have a very different system: Mac Mini w/SSD and 8GB RAM > WaveLink > BNC into Wadia 861i > McCormack DNA-500 > Aerial 10Ts.

 

Tried Halide Bridge for a week. Did a nice job, plenty of detail, but light in mid and bass body. Got WaveLink when I still had Bridge and the WL filled in the body and made the stand up bass swing, the tom toms and bass drum thump and the voices fill in with chests and warmth. Not knocking the Bridge, as would have kept it if I did not also have chance to audition the WL.

 

WL responds nicely to cable upgrades, the AQ Coffee with battery dealio, other USB cables like the WW Silver and Black Cat BNC in my system bumped clarity and musicality from stock cables. PM 1.8 (now that I have it set up right) is fantastic as is Audirvana with Integer Mode. WL makes these incremental software upgrades easy to hear and appreciate...and they are like what you would expect from major hardware updgrades, like going to the $30K DAC stack from the $5K one.

 

Was thinking about upgrading my almost 10 year old CD player/DAC, but WL and computer audio with memory play, HOG mode, 2X upsampling, integer mode...have raised the bar to such a high musical level (like way beyond what I was expecting when I went to a server based system) that I'm not convinced that I'll get better music than I'm getting with the 861's 24-bit Burr-Brown 1704 DAC chips. Wadia and others are still using them in DACs...I guess because they make music sound like music.

 

Will have to upgrade for higher rez stuff beyond 24/96. But remakes of Stone stuff that I have already bought 2 or 3 times that was recorded 40+ years ago is of no interest and most of the hi rez catalogue is "audiophile" music, stuff that I'm done buying to listen to once and put in the pile of a dozen or so like recordings that I don't care to listen to. Can happily wait with the WL bridging the tech gap.

 

Gordon made a great product for those with good DACs and CD players with BNC/SPDIF inputs...you may never feel the urge to spin a disc again, I don't.

 

It's that freakin' good!!!