Jump to content
IGNORED

Async USB everywhere!


Recommended Posts

Everywhere I look 'async' USB dacs seem to be popping up! As far as I was aware Gordon worked on the principle and has now licensed Ayre, DCS have their own version, but today I read that the 'beresford' dac ( a reworked dac from TEK in Taiwan ) is also 'async' USB.

I freely admit I simply do not know enough about the subject, can someone explain 'async' USB to me, are there different implementations or only one true path?

Thanks in advance,Keith.

 

Link to comment
Share on other sites

Peter Hi, thanks for the link, I have read Chris' summation before, here is the quote,

'The news from Stan Beresford is that the Caiman uses asynchronous USB:'

 

"But as far as clocks are concerned: I prefer to trust the 12MHz PLL clock that operates the timing in my USB circuit just millimeters away from the USB chip, instead of a clock somewhere embedded in a heavily populated PC motherboard. IRQ and NMI delay issues would quickly ruin any accuracy in such a setup, producing a healthy dose of jitter. And yes, I was once a PC engineer, designing PC accessories for my former employer. So I have hands on experience of clock delays in a PC setup."

 

With all due respect to Chris is his article definitive, because I also read this the other day, another 'async 'USB' dac, this is the new Westlake design for Audiolab.

 

'Async USB? Licensed from Gordon Rankin?'

 

 

'If you're clever enough you don't need a license from Gordon Rankin or his code or have to call it "Asynchronous USB"

 

USB is indeed "async" and clocked from the MDAC - own implementation'

and this quote, when I asked whether the op was sure the new dac was async.

Quote,

'Keith, I was under the impression that there isn't any discrete time signal/word clock in your typical PCM data-stream. Rather that the word clock is recovered from the actual data blocks themselves, ie 10 blocks equals 10 time units, ( a gross simplification).

 

Isn't there only a discrete time signal when some form of clock-lock is used and a separate time signal recovered from clocking the bits in and out and sent over it's own link, or multiplexed with the music data stream in some proprietary signal format.

 

Isochronous/adaptive USB is source lead, it's pushed from source to DAC as the source sees fit, whereas Asynchronous is 'dragged' from the source, at a specific rate defined by the DAC.

 

Async USB just locks the source to the DACS clock, just like some DACs have done for years, In fact John Westlake was one of those designers who made a consumer DAC combo that utilised a DAC generated clock-lock.

 

I've no doubt it suits purveyors of expensively marketed and priced dacs to poo-poo others endeavours in the same field, Async is only a control set for clocking the source to the dac, it's nothing more than a device driver. Shiny cases and lack of understanding might make some dacs worth $20k to some parties, not me.'

 

Just wondered if async had been implemented under a different name in the past, thanks as always in advance,Keith.

 

 

 

 

 

 

Link to comment
Share on other sites

Firstly, if a PCM2902 is being used, this chip is adaptive - no ifs, not buts, no maybes.

Secondly, 12MHz is actually the wrong frequency for a DAC.

 

I suspect he's using an asynchronous rate converter, which isn't the same thing, and causes it's own set of problems, and if this is the case, Benchmark & Bel Canto already do this,

 

your friendly neighbourhood idiot

 

Link to comment
Share on other sites

"I've no doubt it suits purveyors of expensively marketed and priced dacs to poo-poo others endeavours in the same field, Async is only a control set for clocking the source to the dac, it's nothing more than a device driver. Shiny cases and lack of understanding might make some dacs worth $20k to some parties, not me.'"

 

Funny, I feel the same way about those Cessaro speakers.

 

Wavelength Silver Crimson/Denominator USB DAC, Levinson 32/33H, Synergistic Research Cables and AC cables, Shunyata Hydra V-Ray II with King Cobra CX cable, Wilson Sasha WP speakers with Wilson Watch Dog Sub. Basis Debut V Vacuum turntable/ Grahm Phantom/Koetsu Jade Platinum. MacBook Pro 17\" 2.3GHz Quad Core i7, 8GB RAM, Pure Music, Decibel, Fidelia, AudioQuest Diamond USB Cable.

Link to comment
Share on other sites

You should stick with what you have first hand experience with. Again, your speakers look they they have toilet bowels attched to them. Is that witty enough for you?

 

 

 

Wavelength Silver Crimson/Denominator USB DAC, Levinson 32/33H, Synergistic Research Cables and AC cables, Shunyata Hydra V-Ray II with King Cobra CX cable, Wilson Sasha WP speakers with Wilson Watch Dog Sub. Basis Debut V Vacuum turntable/ Grahm Phantom/Koetsu Jade Platinum. MacBook Pro 17\" 2.3GHz Quad Core i7, 8GB RAM, Pure Music, Decibel, Fidelia, AudioQuest Diamond USB Cable.

Link to comment
Share on other sites

that many manufacturers are keen to jump on the async bandwagon - thanks largely to Gordon's efforts, awareness has risen dramatically. Async USB ( if done properly ) can in theory completely eliminate source-induced jitter - as can properly done Ethernet and firewire. This doesn't mean that these devices must have low jitter, but one of the major contributors to jitter has been eliminated by using this type of mode.

 

I think that the major problem is that we have something that is technically a good thing, but there's so much hyperbole and technobabble spouted that at times it's hard to see the wood for the trees, so it's easy for manufacturers to make vague statements like this. Fortunately there is a tool available on the Mac that basically says what mode the DAC is - can't remember what it's called, no doubt someone will know ( Chris? )

 

Looking at the Caiman specs on the website ( http://www.homehifi.co.uk/products/Caiman.html ), it says quite specifically it uses a PCM 2902, which is an adaptive chip, but to be fair no async claims are made there. The DAC needs to be designed from the outset to use Async USB, and it's no coincidence all the well-known async solutions ( Ayre, dCS, Wavelength ) use a TAS1020B...

 

your friendly neighbourhood idiot

 

 

Link to comment
Share on other sites

I'm not that upset Keith. Believe me.

 

You seem to know what is and is not Async USB by your previously made statements.

 

So just give us all a break with your "Who Me?"

 

Wavelength Silver Crimson/Denominator USB DAC, Levinson 32/33H, Synergistic Research Cables and AC cables, Shunyata Hydra V-Ray II with King Cobra CX cable, Wilson Sasha WP speakers with Wilson Watch Dog Sub. Basis Debut V Vacuum turntable/ Grahm Phantom/Koetsu Jade Platinum. MacBook Pro 17\" 2.3GHz Quad Core i7, 8GB RAM, Pure Music, Decibel, Fidelia, AudioQuest Diamond USB Cable.

Link to comment
Share on other sites

You give me far more credit than I deserve, I understand the principle but have no idea how it is implemented or indeed whether another designer tackled the problem before Gordon, and simply called it something else, how to DCS implement their async USB.

I completely understand if you do not like my bowels.

Keith.

 

Link to comment
Share on other sites

Lars, firstly I don't know Keith, or his audio company, but I do know that he seems to be a polite kind of a chap, who is clear in his position as a retailer and has always seemed like a nice bloke, so calm down, OK?

 

Secondly, there seems to be so much misinformation that I thought I'd try and explain ( again ) why the term "asynchronous" is used, and what it means - this is true for properly done ethernet, firewire, USB, but all examples given are for USB ( as that is what this thread is about )

So, in the first case - adaptive, or possibly "synchronous": The audio is synchronous with the carrier clock. USB works on 1kHz ticks - a tick happens every millisecond, and for 44.1k audio, 9 ticks each have 44 samples, and the 10th has 45 samples in it, so over a second, or 1000 ticks, we have transferred 900*44+100*45=44100 samples. For the DAC to work at this rate, it has to adjust itself so it "consumes" this amount every second, which it can only do by aligning itself to the "tick" from the USB, which is derived from the PC. If it disagrees ( i.e. it uses a different clock ), then the DAC will either consume too much, or too little data, and will over or underflow it's buffers. Now the issue is that the 1ms tick is generated by the PC, and the specs are pretty horrendous for jitter and accuracy.

So, the second case works because the audio is no longer synchronous with the the 1ms tick. The DAC basically tells the PC how much stuff it wants to keep it's buffer from overflowing or emptying, so, if the DAC clock is slightly faster than the USB tick, it will ask to PC to send 44101 samples in the next 1000 ticks. The important thing here is that the DAC no longer has to track the tick - when it happens is largely irrelevant, and the accuracy is determined by the clock in the DAC, not the accuracy of the tick.

 

your friendly neighbourhood idiot

 

 

Link to comment
Share on other sites

OK, it looks like I took the "Quotes" as his statements if I'm interpreting things correctly now. I take it back. Your speakers are works of art Keith :)

 

Wavelength Silver Crimson/Denominator USB DAC, Levinson 32/33H, Synergistic Research Cables and AC cables, Shunyata Hydra V-Ray II with King Cobra CX cable, Wilson Sasha WP speakers with Wilson Watch Dog Sub. Basis Debut V Vacuum turntable/ Grahm Phantom/Koetsu Jade Platinum. MacBook Pro 17\" 2.3GHz Quad Core i7, 8GB RAM, Pure Music, Decibel, Fidelia, AudioQuest Diamond USB Cable.

Link to comment
Share on other sites

you have some ugly speakers ;) Seriously, good to see things being patched up - I think CA stands alone in the general friendly, open nature of it's forums.

 

The key thing that the famous three™ do is that they comply with the USB standard for async, so don't need any drivers - this may or may not seem like a big deal, but given how relatively small a lot of high-end companies are, ongoing support must be a concern...

 

I think that the thing is that anything that uses a PCM2x0x chip for USB can't be that fantastic in jitter terms.

But... many companies use a TAS1020B with the Centrance code, so just picking the USB chip isn't the way to go either. As always, do your research before parting with your cash, and again, there's no guarantee that even a super-duper async DAC done just as it should be will suit your tastes, but I wouldn't be happy giving my money to someone who was being somewhat economical with the truth - is there a link to the Caiman quotes?

 

your friendly neighbourhood idiot

 

Link to comment
Share on other sites

Any suggestion that my DAC is a reworked version of another company's DAC is of course character assassination by Keith. I have been designing products for quite a few companies for nearly two decades now. That includes designs for TEC. I decided to sell the DAC under my own name after I got made redundant from my previous job, and did not have to worry about being dismissed by my former boss for breaching the terms of my employment.

I have made the point about my DAC design many times over many years, and Keith is well aware of that. However, he persists in peddling a load of lies because he is upset that he was rebuked for approaching TEC with the intention of importing my products behind my back into the UK.

 

Stan

 

And by the way, I give a money back guarantee if one is not satisfied with my DAC. I won several magazine awards, and have several thousand users world wide who are quite happy with my product. I have nothing to be ashamed of and can count my DAC range within the top 20 of recognized brand names amongst DACs on a world wide basis. And my DACs are not built in China, unlike most other DACs in my price range.

 

Link to comment
Share on other sites

Could people consider that in increases an price has higher expectation?

 

Comparing a Beresford product at circa $US250 versus Wavelength and Ayre and (other recognised async USB DAC that I do not know) based on outputs when there is at least 2 or more differences in orice from the Beresford is not a good way of measuring value for money.

 

Listening is.

 

Keep on Upgrading!!!

Link to comment
Share on other sites

except a big thank you to I_S!

 

NOW I get it! :) I know you've written about this again and again, but this time, for dim me anyway, I think it got through! Thank you, an excellent thread.

 

ps - I'm now going to encode all of my mp3's at 10k so my ticks will be able to easily keep up with my tocks. Genius! :)

 

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share



×
×
  • Create New...