Topic: Learn me: Asynchronous and USB DACs (Read 2696 times)

Mr.Sneis · « **on:** May 02, 2013, 06:03:34 PM »

Before making any further of an ass of myself I just wanted to make sure I understand these terms proper and in the simplest form possible. I get the impression that many a folk gloss over the details either because they don't really know or because they do know but take for granted that blokes like me may be misinformed. I've found a crap ton of info explaining these things by "big" names but not so much in neanderthal English; please help me to know if I am on the right track!

Isochronous - The USB mode through which audio data is transferred... seems to apply for both Asynch and Adaptive USB.

Asynchronous USB - it is a means to separate the clock of the DAC from the clock of the computer. In this case we want for the DAC to be controlling everything and not the computer. At first the usb/spdif converters were introduced to provide this ability for DACs but these days we get DACs that have this included.

Adaptive USB - In this case the DAC is a slave to the computer. The computer's usb could be erratic creating jitter in its output, the DAC has to pick up the pieces and make sense of it.

Asynchronous Sample Rate Conversion - A jitter reduction method used in Adaptive USB. This has actually been around for quite some time before Asynchronous USB was commercialized to the masses.

Recently I was confused by this being marketed along with the Parasound ZDAC: "Asynchronous sample rate conversion for all inputs including the USB input." This is quite a cleverly worded selling point because it appealed to a dummy like me on the surface while in fact saying there's nothing new going on to the well informed.

Here it seems the reviewer has it right: "Handling the USB input in Isochronous Adaptive mode is the TI TAS1020B USB receiver and the Zdac re-clocks and upsamples all incoming data to 24-bit/422kHz via the Analog Devices AD1895 sample rate converter before handing off the upsampled signal to the Analog Devices AD1853 DAC"
http://www.audiostream.com/content/parasound-zdac

Here I think the reviewer may be confused: "The Zdac is a compact but full featured unit with plenty of connectivity. Inputs include coaxial and optical SPDIF as well as asynchronous USB."
http://www.innerfidelity.com/content/parasound-zdac

And here unfortunately shows that I am probably not the only one taking others' word for it: http://www.head-fi.org/t/655104/parasound-zdac-reviews

FWIW consensus seems to be that Adaptive or Asynchronous DACs can both sound good or like crap so don't necessarily trust in the jargon only.

Anaxilus. · « **Reply #1 on:** May 02, 2013, 06:53:27 PM »

This adds even more fun: http://centrance.com/products/dacport/

Jitter Management - Debunking the Asynchronous Myth
"Some manufacturers may lead you to believe that Asynchronous USB transfers are superior to Adaptive USB transfers. This no more true than saying that you "must" hold the fork in your left hand. If you know what you are doing, you will feed yourself with either hand.

The USB argument comes down to jitter management and goes as follows: In Asynchronous mode the device is the clock master. In Adaptive mode, the computer is the clock master. Either way works fine, if correct design principles are followed. Here is the tricky part that often gets omitted: No matter which side is the source of the clock (PC or DAC), the two devices are still connected by the USB cable and the digital data on that USB cable is always irregular because the computer is involved. Computers do many things at once and end up sending data over USB in irregular intervals, no matter who is the clock master on the bus. This irregularity causes jitter. So, there is no jitter-free solution, just like there is no dust-free house. Irregularity always creeps in and needs to be actively managed.

Here is where the Asynchronous vs. Adaptive argument breaks down: In either of the two clocking schemes, jitter is present during transmission. It's inevitable and also ok, if it is properly cleaned up prior to the D/A conversion, where it matters most. Neither clocking scheme is superior and both are capable of performing well if you know how to reassemble the bits prior to the DAC. Now, how do you actually do that? There are many ways, the oldest and simplest of which is buffering. Irregular data comes in, regular data goes out. The most important part is to make sure that samples leaving the buffer on the way to the DAC are clocked accurately. DACport employs JitterGuard™, a proprietary two-stage clock management system that does just that - cleans up the jitter on the USB bus so that samples are virtually jitter-free at the D/A conversion point. The result is a natural response, crisp definition, extra resolution and wide soundstage."

I believe the new HifiM8 will be Asynchronous. $:-\$

DaveBSC · « **Reply #2 on:** May 02, 2013, 07:03:53 PM »

Quote from: Mr.Sneis on May 02, 2013, 06:03:34 PM

Before making any further of an ass of myself I just wanted to make sure I understand these terms proper and in the simplest form possible. I get the impression that many a folk gloss over the details either because they don't really know or because they do know but take for granted that blokes like me may be misinformed. I've found a crap ton of info explaining these things by "big" names but not so much in neanderthal English; please help me to know if I am on the right track!

Isochronous - The USB mode through which audio data is transferred... seems to apply for both Asynch and Adaptive USB.

Asynchronous USB - it is a means to separate the clock of the DAC from the clock of the computer. In this case we want for the DAC to be controlling everything and not the computer. At first the usb/spdif converters were introduced to provide this ability for DACs but these days we get DACs that have this included.

Adaptive USB - In this case the DAC is a slave to the computer. The computer's usb could be erratic creating jitter in its output, the DAC has to pick up the pieces and make sense of it.

Asynchronous Sample Rate Conversion - A jitter reduction method used in Adaptive USB. This has actually been around for quite some time before Asynchronous USB was commercialized to the masses.

Recently I was confused by this being marketed along with the Parasound ZDAC: "Asynchronous sample rate conversion for all inputs including the USB input." This is quite a cleverly worded selling point because it appealed to a dummy like me on the surface while in fact saying there's nothing new going on to the well informed.

Here it seems the reviewer has it right: "Handling the USB input in Isochronous Adaptive mode is the TI TAS1020B USB receiver and the Zdac re-clocks and upsamples all incoming data to 24-bit/422kHz via the Analog Devices AD1895 sample rate converter before handing off the upsampled signal to the Analog Devices AD1853 DAC"
http://www.audiostream.com/content/parasound-zdac

Here I think the reviewer may be confused: "The Zdac is a compact but full featured unit with plenty of connectivity. Inputs include coaxial and optical SPDIF as well as asynchronous USB."
http://www.innerfidelity.com/content/parasound-zdac

And here unfortunately shows that I am probably not the only one taking others' word for it: http://www.head-fi.org/t/655104/parasound-zdac-reviews

FWIW consensus seems to be that Adaptive or Asynchronous DACs can both sound good or like crap so don't necessarily trust in the jargon only.

You're mostly right, but ASRC has nothing to do with USB. ASRC means exactly what it says: the incoming sample rate, say 16/44.1 is upconverted, usually to 24/192. This is an easily accomplished but generally crappy method of jitter control. 192 isn't a direct multiple of 44.1, and so you introduce as many problems as you solve when you do the conversion.

DaveBSC · « **Reply #3 on:** May 02, 2013, 07:09:17 PM »

Quote from: Analixus on May 02, 2013, 06:53:27 PM

This adds even more fun: http://centrance.com/products/dacport/

Jitter Management - Debunking the Asynchronous Myth
"Some manufacturers may lead you to believe that Asynchronous USB transfers are superior to Adaptive USB transfers. This no more true than saying that you "must" hold the fork in your left hand. If you know what you are doing, you will feed yourself with either hand.
The USB argument comes down to jitter management and goes as follows: In Asynchronous mode the device is the clock master. In Adaptive mode, the computer is the clock master. Either way works fine, if correct design principles are followed. Here is the tricky part that often gets omitted: No matter which side is the source of the clock (PC or DAC), the two devices are still connected by the USB cable and the digital data on that USB cable is always irregular because the computer is involved. Computers do many things at once and end up sending data over USB in irregular intervals, no matter who is the clock master on the bus. This irregularity causes jitter. So, there is no jitter-free solution, just like there is no dust-free house. Irregularity always creeps in and needs to be actively managed.

Here is where the Asynchronous vs. Adaptive argument breaks down: In either of the two clocking schemes, jitter is present during transmission. It's inevitable and also ok, if it is properly cleaned up prior to the D/A conversion, where it matters most. Neither clocking scheme is superior and both are capable of performing well if you know how to reassemble the bits prior to the DAC. Now, how do you actually do that? There are many ways, the oldest and simplest of which is buffering. Irregular data comes in, regular data goes out. The most important part is to make sure that samples leaving the buffer on the way to the DAC are clocked accurately. DACport employs JitterGuard™, a proprietary two-stage clock management system that does just that - cleans up the jitter on the USB bus so that samples are virtually jitter-free at the D/A conversion point. The result is a natural response, crisp definition, extra resolution and wide soundstage."

I believe the new HifiM8 will be Asynchronous. $:-\$

Says the company trying to sell you an adaptive mode USB DAC. You can implement adaptive mode well, as Centrance does, and you can certainly implement asynchronous extremely poorly (M2Tech Hiface). That does not mean they are just two sides of the same coin.

The Empirical Off-Ramp 3 used the TAS1020 with Centrance firmware. The OR5 uses asynchronous based on the Hiface OEM module. The OR5 MURDERS the OR3. IMO, properly implemented asynchronous cannot be matched by adaptive mode.

Mr.Sneis · « **Reply #4 on:** May 02, 2013, 07:24:05 PM »

Quote from: DaveBSC on May 02, 2013, 07:03:53 PM

You're mostly right, but ASRC has nothing to do with USB. ASRC means exactly what it says: the incoming sample rate, say 16/44.1 is upconverted, usually to 24/192. This is an easily accomplished but generally crappy method of jitter control. 192 isn't a direct multiple of 44.1, and so you introduce as many problems as you solve when you do the conversion.

I see, it seems you can even add asrc on top of having an asynchronous usb interface. Albiet at the expense of sound quality.

http://www.whatsbestforum.com/showthread.php?7420-Asynchronous-Sample-Rate-conversion-%28ASRC%29-DACs&p=124905&viewfull=1#post124905

Does this mean that if a dac is an upsampling dac (non-defeatable) then it is using asrc?

Another thing I've picked up on, feels like the audiophile market is wrought with manufacturers pushing asynchronous usb but not so much in the pro market - at least that I can tell.

DaveBSC · « **Reply #5 on:** May 02, 2013, 09:15:51 PM »

Yep, sample rate conversion and the USB interface are two completely separate things, and they are also different from oversampling, which is a hardware level multiplier of the of the sample rate. The PCM1704 for example usually runs with 8X OS.

You can have either asynchronous or synchronous US. With synchronous upconversion, 44.1 would be upconverted to 176.4 instead of 192. I think the Byston DACs use synchronous upsampling.

The pro guys are generally a lot slower to introduce new technology, and many are designed to use external clock generators anyway.

dsavitsk · « **Reply #6 on:** May 02, 2013, 10:30:52 PM »

I think you basically have it correct. Asynchronous may be theoretically better than adaptive, but good adaptive is definitely better than bad asynch. Both actually work pretty well. And jitter can be cleaned so a jitter reduction circuit may well be enough.

Additionally, ASRC chips do not actually need to upsample -- you can use the master clock for the ASRC. You will get the same sample rate out, and it turns out, generally lower jitter. How much lower probably depends on the input as the ASRC will have its own jitter, so it may make a really clean signal worse, but they will definitely make a bad signal cleaner. It has become fashionable among the cool kids to badmouth these chips, but I have found that they work pretty well.

More importantly, the analog stage, and the IV are both more important than any of this

I'd take a good IV with a PCM2707 as USB receiver over the best Asynch USB with an opamp IV any day.

DaveBSC · « **Reply #7 on:** May 02, 2013, 11:05:33 PM »

Quote from: dsavitsk on May 02, 2013, 10:30:52 PM

More importantly, the analog stage, and the IV are both more important than any of this I'd take a good IV with a PCM2707 as USB receiver over the best Asynch USB with an opamp IV any day.

Yes and no. The input stage should not be discounted as meaningless, or something that can be made irrelevant with enough jitter reduction techniques. I mean just listen to the same DAC and switch between linear and minimum phase filters. Same hardware. This stuff matters. The PCM2707 is a pile of garbage, period.

ultrabike · « **Reply #8 on:** May 02, 2013, 11:21:22 PM »

I don't think Doug is discounting the input stage at all. He is just giving more weight to the analog stage.

I have not worked with USB interfaces, but from the little I know and understand, either way jitter will couple through the USB interface. Therefore I tend to agree with Doug that the implementation a jitter reduction block can make or brake things with either Asynchronous or Adaptive approaches.

firev1 · « **Reply #9 on:** May 03, 2013, 12:25:07 AM »

Diyaudio has a very nice thread about ASRC down here, albeit complex to understand, http://www.diyaudio.com/forums/digital-source/28814-asynchronous-sample-rate-conversion.html#post332675.

I feel that the async usb operation results in the final implementation doing less or nothing at all to the signal at the output, is that right?

Whereas ASRC has a bunch of issues and from that article above, can only approximate the operations(as the filters and operations are actually in the analog domain) used on the digital signal which is where the fidelity issues come from?

Author Topic: Learn me: Asynchronous and USB DACs (Read 2696 times)

Mr.Sneis

Learn me: Asynchronous and USB DACs

Anaxilus.

Re: Learn me: Asynchronous and USB DACs

DaveBSC

Re: Learn me: Asynchronous and USB DACs

DaveBSC

Re: Learn me: Asynchronous and USB DACs

Mr.Sneis

Re: Learn me: Asynchronous and USB DACs

DaveBSC

Re: Learn me: Asynchronous and USB DACs

dsavitsk

Re: Learn me: Asynchronous and USB DACs

DaveBSC

Re: Learn me: Asynchronous and USB DACs

ultrabike

Re: Learn me: Asynchronous and USB DACs

firev1

Re: Learn me: Asynchronous and USB DACs

(adsbygoogle = window.adsbygoogle || []).push({});

Author Topic: Learn me: Asynchronous and USB DACs (Read 2696 times)