Wavelength WaveLink Asynchronous USB to S/PDIF converter.

[Grahame]

Thanks for the clarification. I stand corrected.

[JayDee]

My biggest concern is that there is no way to audition this thing unless you wait for a meet, pay $250 if you don't like it, or plan a trip to Utah. Granted the same thing could be said about many products, usually there are measurements, comparisons between competitors by reviewers, large numbers of reviews, company reputations, and often internal pictures or detailed descriptions to help buyers decide. I think I've mentioned that ad nauseam on head-fi already. I'm sure this will change for the Diverter, but I would have expected the manufacturer and distributors to support the process by setting up more information and opportunities.

All of these concerns are acknowledged, and we're doing all we can to rectify them. I wish I could significantly advance the schedule and have everything in place right now, but it takes time, generally much more than I'd like. At least one formal review is pending, and others are scheduled. The Diverter has already been at some meets, and will be at more in the future. Obviously we'd like to have everyone who has the interest to have an opportunity to try one out, but please realize that each one of these is entirely designed, machined, assembled, and tested by me personally; it is simply not possible to send one out to everyone who might want to play with it casually.

[manaox2]

All of these concerns are acknowledged, and we're doing all we can to rectify them. I wish I could significantly advance the schedule and have everything in place right now, but it takes time, generally much more than I'd like. At least one formal review is pending, and others are scheduled. The Diverter has already been at some meets, and will be at more in the future. Obviously we'd like to have everyone who has the interest to have an opportunity to try one out, but please realize that each one of these is entirely designed, machined, assembled, and tested by me personally; it is simply not possible to send one out to everyone who might want to play with it casually.

Thanks for answering my concerns, that means a lot in my mind. BTW, I want to say that to me your not my opponent. I mean, I wouldn't refuse splitting a cab with you if the opportunity presented itself. I look forward to trying this and possibly being surprised.

[JayDee]

Isn't the whole point of this device to take USB audio and convert to S/PDIF to a DAC? Wouldn't jitter measurements as well as ensuring that the data is preserved bit-perfect be the meaningful measurements?

Yes, I agree. And that is what will be offered... when they are ready.

Of course you have no control of the quality of the S/PDIF interface on the DAC end , so I'm not sure how you would go about offering a meaningful measurement, particularly of jitter.

You're absolutely right about the receiving end having an influence on the jitter measurements. That is why measurements are done under repeatable conditions - in my case, using a 75 ohm cable coaxial, terminated into a SMT 75 ohm, .05% resistor. Most equipment is tested in this repeatable, controlled fashion, although the designer knows full well that the environment will unavoidably alter the performance. On the speaker side of the business, for instance, most speakers would be designed with anechoic or quasi-anechoic measurements, but the room environment in which they are used alters that sort of textbook response in a major way.

In the case of a SPDIF interface, I can show you what happens with it is improperly terminated, as I happen to have these measurements handy.

The first waveform shows the results from an improperly terminated interface (cable has an incorrect impedance). You can see a reflection (dip) riding on top of the square wave.

The second waveform shows the results with correct termination. Incidentally, this is SPDIF output from a current generation Diverter.

[digger945]

So is it fair to say that incorrect cable termination may lead to a change or shift in timing, and timing is the name of the jitter game?

[JayDee]

Thanks for answering my concerns, that means a lot in my mind. BTW, I want to say that to me your not my opponent. I mean, I wouldn't refuse splitting a cab with you if the opportunity presented itself. I look forward to trying this and possibly being surprised.

Sure, thanks for your kind words. FWIW, I expected opposition, and although the experience often feels like running the gauntlet, I welcome the opportunity to articulate my views on the forums. It has been an interesting experiment.

I think I've probably posted more than enough for today, so I'm going to sign off for a bit. I'll continue to monitor the thread and will post if there's something that merits comment.

Here's looking forward to that cab ride.

[JayDee]

So is it fair to say that incorrect cable termination may lead to a change or shift in timing, and timing is the name of the jitter game?

Yes - jitter has a several causes and influences, and one of them is interface jitter. This is the primary reason why digital cables have a sonic signature, though by a different mechanism than those used for analog connections. The parasitic effects are there in both cases, but any timing issues on the analog side would show up as phase shifts that are so small as to be undetectable except in extreme cases. With digital, the timing issues are much more pronounced and consequential.

It is unfortunate that the RCA connector has become the de-facto standard simply by winning a popularity contest. I yielded to that influence myself in the sense that the first two generations of Diverter used an RCA connector, but I knew from day one that I didn't like it from a technical point of view. The third gen uses a BNC, but still accommodates those who want to use RCA connections by supplying an adapter.

[morphsci]

I never claimed that it did. This was a difference test, not an absolute test, designed to show one thing, and one thing only: that the Diverter is, at least as far as the methods of this test allow, transparent in the way it handles digital data. The results of the test would be valid for comparison if and only if the testing device were calibrated to some kind of recognized standard, allowing anyone else with similarly calibrated equipment to repeat the test results with some other device. I wasn't attempting to be nearly that formal or comprehensive in this test.

Well that is all well and good but unfortunately it does not answer your potential customers questions. Why should I pay 4 to 10 times as much for this device? Is it better than the cheaper products? Since I cannot do a reasonable in home test (Unless I want to pay $260 for the privilege) then at least some information could be made available. Who cares if other manufacturers don't believe your measurements. The measurements provide refutable eveidence while assertions are just that. Oh, and I refuse to take the comments of anyone who uses "scooter" in their nomiker seriously. When we used to call someone scooter, it was not a term of endearment.

[grenert]

With digital, the timing issues are much more pronounced and consequential.

What is the timing issue in your waveform? The dip is a small amplitude effect which I can't imagine would have any effect on how the digital signal is interpreted on the receiving end. Is there some other problem?

[JayDee]

What is the timing issue in your waveform? The dip is a small amplitude effect which I can't imagine would have any effect on how the digital signal is interpreted on the receiving end. Is there some other problem?

It depends on where the reflection occurs. That waveform isn't a particularly ugly case; it's just one I happened to have immediately accessible that shows what a reflection looks like. My intent here was simple to illustrate a cable reflection, rather than to show what a really bad digital waveform looks like.

[screaming oranges]

My diverter order ships sometime mid-late next week. It will definitely be available for auditioning at the December meet. I will also take the Trends UD10.1 and the Musiland Monitor US01 for anyone interested in personally seeing (hearing) what the deal is with each of these.

[manaox2]

My diverter order ships sometime mid-late next week. It will definitely be available for auditioning at the December meet. I will also take the Trends UD10.1 and the Musiland Monitor US01 for anyone interested in personally seeing (hearing) what the deal is with each of these.

Your making me want to order a m2tech hiface just to compare.

[screaming oranges]

Your making me want to order a m2tech hiface just to compare.

Heh. I was thinking of ordering on emyself to do just that, but I need to divert funds elsewhere.

[Sherwood]

I need to divert

[screaming oranges]

By the time I realized my unintentional pun, the edit option had expired. I was hoping no one would catch it, but it looks like Big Brother is watching...

[luvdunhill]

what's generating a waveform with a 7ns rise time?

[morphsci]

what's generating a waveform with a 7ns rise time?

Hope?

[Pars]

John Swenson posted a really nice overview if you will of async USB on audioasylum which might be of interest to many of you:

Digital Drive - Some thoughts on async etc. - John Swenson - October 24, 2009 at 15:23:20

[luvdunhill]

Hope?

heh.

[cetoole]

John Swenson posted a really nice overview if you will of async USB on audioasylum which might be of interest to many of you:

Digital Drive - Some thoughts on async etc. - John Swenson - October 24, 2009 at 15:23:20

Nice, seems to be a great summary of the issues. It is somewhat surprising you dont see more slaved PCI transports, as it really doesnt take much to implement, and some even can be made to sync to an input signal, so would require no hacking.

[Grahame]

Interesting read as well. I wonder how long it will be before people want meta data embedded with the transmitted audio data

[Hopstretch]

Interesting read as well. I wonder how long it will be before people want meta data embedded with the transmitted audio data

About six minutes. Because, now that you've mentioned it, I want that.

[Dusty Chalk]

Without reading the article -- isn't that what CD-Text is?

Will read article.

[ar-t]

I bought a pre-production version of the Legato, which also uses Wavelength's Streamlength firmware. Differences between Legato and Wavelink:

- cost is $500 for production version, I got my preproduction version for $400. Difference is the preproduction version has a generic Hammond case... its pretty big... about 1.5 times the size of Wavelink.

- restricted to 44/16 (I guess GR did this to segment the market)

Someone just made this thread known to us.

Yes, it does 44/16 only, but "GR" had nothing to do with that. That was solely our choice. Our goal was to make the best converter possible, specifically for 44 kHz. The only way to do this was to have one and only one internal clock. For lowest possible jitter, the clock had to work at 11.2896 MHz. Working at twice that frequency, which would have allowed using additional sampling rates, would result in higher jitter.

A lot? Well, maybe not. Would it be audible? On some systems, yes. Since our intention was to get the best possible performance from Redbook CDs, it dictated an architecture that might seem overly restrictive to a fair segment of the market.

Several folks who have pusrchased one, having gotten by that restriction, have said "I have x hi-res files, and xxxx Redbook CDs. OK, I see your point."

For those of you still wondering if we are ever going to have front panels.......

Yes. The folks who were in charge have been sacked. I am in charge of that, now. (May be a mistake.) The raw panels have not been shipped yet. Something about "Uh, we had a lot of snow up here, and.........uh.........some orders have slipped." Once they arrive, add 1 week for the logo to be applied.

Pat

[digger945]

Fortunately j4cbo is going to do his own solution based on a less outdated chip...

Could we get some details?

Dan, it'll be interesting to see what j4cbo comes up with.

Me too.