[btbluesky]

Quote:

Originally Posted by DaveBSC /img/forum/go_quote.gif I don't understand using USB in a high end system at all. USB is not only bit rate limited but seems to suffer serious flaws as a digital transmission format. Particularly interesting was a set of tests showing the difference between a laptop connected with USB running on its batteries and connected to its power supply. When switching to wall power, the amount of jitter absolutely exploded, though even battery powered USB was still a jittery mess. S/Pdif and AES/EBU both have their problems, but USB seems to be far worse than both.

Since founder of stereophile just pasted away, I'll just use their links...
Ayre usb dac
Bel Canto USB transport
wavelength
I'm just including 1 magazine, and I'm pretty sure each of the reviewer have 30 or 50k+ downstream equips. You want more google enjoythemusic or other highend site for more models/reviews.
So what is your source for the "jitter mess". FYI, wavelength/ayre/dCS, their implementation is technical impossible to get jitter at the USB transport process (async USB), because the clock Master is at the DAC, not the computer.

I have a Havana w/ an old PCM27xx usb receiver chip, but butter smooth sound. Bel Canto coming in soon coz I'm planning on getting the monarchy.

 

[Bemopti123]

To future proof your purchase, as well as buy into a well known and storied past, I think the Resolution Audio CANTATA series, while not yet widely reviewed, it holds the best promise.

It is the best compromise of a machine that can serve as a stand alone digital front end, as well as a DAC for any future applications you might want. BTW, the designer has invented a new way of communicating from PC that uses Ethernet and then, adapts it to USB?

 

[DaveBSC]

Quote:

Originally Posted by btbluesky /img/forum/go_quote.gif So what is your source for the "jitter mess". FYI, wavelength/ayre/dCS, their implementation is technical impossible to get jitter at the USB transport process (async USB), because the clock Master is at the DAC, not the computer.

Laptop on battery power, jitter at around 150ps.

Laptop on wall power. Jitter at around 2800ps. I haven't heard any of the true asynch USB DACs like the Ayre, so I don't know how they compare to S/Pdif. The Bel Canto is not asynch though, it's the same setup as any of the other DACs out there with a 24/96 USB input - mediocre.

 

[DanD]

DaveBSC, may I ask which DAC is that measurement for ?

 

[c3p0]

I had the Ayre and Chord QBD76 for a few days, and the chord just sounds so much more transparent and natural. Music comes alive and instruments and details sound so much more real.

I decided to keep the Chord.

 

[DaveBSC]

Quote:

Originally Posted by DanD /img/forum/go_quote.gif DaveBSC, may I ask which DAC is that measurement for ?

It's data provided by Cambridge in their response to the Stereophile review of the DAC Magic. Obviously the Cambridge's 16/48 max USB input isn't as good as the 24/96 inputs on more expensive DACs, and this problem may not be an issue if a true asynch USB DAC can slave the computer's clock. I don't know.

The point is though that jitter from the source via USB is a significant problem, one that almost all of the USB DACs like the Bel Canto converter can't solve.

 

[btbluesky]

Quote:

Originally Posted by DaveBSC /img/forum/go_quote.gif It's data provided by Cambridge in their response to the Stereophile review of the DAC Magic. Obviously the Cambridge's 16/48 max USB input isn't as good as the 24/96 inputs on more expensive DACs, and this problem may not be an issue if a true asynch USB DAC can slave the computer's clock. I don't know. The point is though that jitter from the source via USB is a significant problem, one that almost all of the USB DACs like the Bel Canto converter can't solve.

Thus I said your point is flawed.
That Cambridge is using the lowest chip PCM27XX, plus its a real lowend DAC. And a laptop??

1. I'm talking about all silence PC with a overbuilt power supply. BTW, computer power supplies are some of the MOST high spec "transformer" because they're supplying juice for sensitive components. Get a high spec one, and all these graphs would be diff.Just a laptop w/ a poor equip doing bad can prove only that setup is bad. Real logical flaw/over generalization.

2. Not all implementation are alike, transport and DAC are so implementation-dependent. There are 3 ways of doing USB, PCM27xx/29XX 16/44. The TI chip for 24/96, and async. Among these 3, there are so many implementation. PLUS modern DAC has jitter rejection in SPDIF, read the last page of the bel canto USB review, using benchmark and Bel canto DAC, the measured charts are identical because the newer DAC can easier resolve those jitters.

3. Async is probably the best, a rather holy grail of USB. But I got the Bel canto w/ TI chip 24/96 coz its the currently best method that let me try out the other spdif DAC. Another even better one would be empirical offramp going into northstar using I2C. But price is high.

4. goto HDtracks, try some 24/96 samples.The ease of use pay itself in no time.

5. unless u do all vinyl, computer is the ONLY way to go moving forward. USB 3 is coming with 10x more bandwidth + backward compatible. It'd be even better.

 

[DaveBSC]

Quote:

Originally Posted by btbluesky /img/forum/go_quote.gif 5. unless u do all vinyl, computer is the ONLY way to go moving forward. USB 3 is coming with 10x more bandwidth + backward compatible. It'd be even better.

I agree with you on this point. The PS Audio PWT's ability to read 24/176 HRx files right from the DVD is rather silly. Why would you want to do that when a music server would provide an entire collection of HRx recordings and whatever else at the touch of a button?

I do not agree though that a Win7 or OSX machine running off an Enermax or PC Power and Cooling or whatever "high quality" PSU you want to use, via USB can be called high-end. At least not without an asynch USB DAC. Haven't heard those, don't know. What I have heard though of USB DACs running the PCM270x, PCM290x or the Centrance/adaptive 24/96 chip has either been significantly worse or somewhat worse than S/Pdif, with the best being about equal with optical digital, another format that shouldn't be used unless there is no other choice.

In my opinion, the best way to go would be a dedicated music server running dedicated software with SSD drives and a real power supply, connected to a DAC via I2S. Next best would be S/Pdif or possibly Firewire, supposedly that actually works very well, with the last choices being optical digital or USB.

 

[Bimber]

I think Berkeley Audio Design DAC one of the best

Berkeley Audio Design Product

 

[BrianS]

hmm, i think the emm dac-2 sounds pretty good using asynch USB with RSA alethias digital master USB cable, 256k mp3/apple lossless, and cannot tell difference with battery powered vs. plugged in laptop or using different power schemes in vista. i have not compared much, but if i recall correctly it sounds a little better than my upgraded cdsa (not using sacd of course but have limited selection here)

btw im using foobar with asio drivers

 

[SoupRKnowva]

i would say EMM dac-2, lavry gold, or MSB DacIV, though i havent heard any of them

 

[sxr71]

Quote:

Originally Posted by DaveBSC /img/forum/go_quote.gif I don't understand using USB in a high end system at all. USB is not only bit rate limited but seems to suffer serious flaws as a digital transmission format. Particularly interesting was a set of tests showing the difference between a laptop connected with USB running on its batteries and connected to its power supply. When switching to wall power, the amount of jitter absolutely exploded, though even battery powered USB was still a jittery mess. S/Pdif and AES/EBU both have their problems, but USB seems to be far worse than both.

Totally agree. Also what kind of galvanic isolation is employed? Maybe you don't hear faint traces of HDD access noises and such on the $2000 USB DACs but still USB isn't so great for other reasons as you mentioned. Also I can't for the life of me fathom what a $1000 software program is supposed to be doing to help matters.

 

[sxr71]

Quote:

Originally Posted by btbluesky /img/forum/go_quote.gif Thus I said your point is flawed. That Cambridge is using the lowest chip PCM27XX, plus its a real lowend DAC. And a laptop?? 1. I'm talking about all silence PC with a overbuilt power supply. BTW, computer power supplies are some of the MOST high spec "transformer" because they're supplying juice for sensitive components. Get a high spec one, and all these graphs would be diff.Just a laptop w/ a poor equip doing bad can prove only that setup is bad. Real logical flaw/over generalization. 2. Not all implementation are alike, transport and DAC are so implementation-dependent. There are 3 ways of doing USB, PCM27xx/29XX 16/44. The TI chip for 24/96, and async. Among these 3, there are so many implementation. PLUS modern DAC has jitter rejection in SPDIF, read the last page of the bel canto USB review, using benchmark and Bel canto DAC, the measured charts are identical because the newer DAC can easier resolve those jitters. 3. Async is probably the best, a rather holy grail of USB. But I got the Bel canto w/ TI chip 24/96 coz its the currently best method that let me try out the other spdif DAC. Another even better one would be empirical offramp going into northstar using I2C. But price is high. 4. goto HDtracks, try some 24/96 samples.The ease of use pay itself in no time. 5. unless u do all vinyl, computer is the ONLY way to go moving forward. USB 3 is coming with 10x more bandwidth + backward compatible. It'd be even better.

I don't disagree with the convenience of computer audio and I have embraced "solutions" like the Squeezebox. It just makes life easier and the music more enjoyable when you don't have to clean your (vinyl) discs and deal with the alignment of your cartridge or even to get up to change the CD/SACD. However computers also have S/pdif outputs in both optical and co-axial. Many of them have firewire also. But the electrical outputs come from some heavily RFI filled boxes. There is almost no filtering on the connections from a computer and I think its audible. Maybe I've been playing with some low end implementations or suffer from a ground loop but it does not engender confidence in a computer source. The Squeezebox may come with its own set of issues but I think its cleaner than a computer as a source and maybe that's just in my head.

Is there an effective way to decouple the noisy PC environment from our DACs? Another thing that does not engender confidence is a claim that a $1000 software program is helping matters. I thought that as long as you bypass Kmixer in Windows you get a bit perfect output. Isn't there a way cheap or free way to bypass OS resampling on a Mac?

 

[grokit]

When speaking of USB audio processing, we really need to separate it out further as to whether or not it's transmitting via "adaptive" USB mode, which is also considered "synchronous", and uses the computer's own clock as the master timer-keeper. This method of audio data transmission is the easiest to implement by far, as it requires no computer programming skills. Unfortunately, the computer cannot maintain perfect timing of the data sent via USB this way, so the receiving chip on the DAC adapts to this drifting signal by re-adjusting its own frequency every millisecond, which tends to cause high levels of jitter.

The other method of USB audio data transmission involves "asynchronous" USB (which is not to be confused with asynchronous samplerate conversion), which uses a clock in or near the DAC, and allows it to drive the DAC directly, thereby not relying on an unstable computer clock. It is called asynchronous because the DAC's master clock is controlled by a high-precision fixed-frequency clock. This clock controls the datastream from the computer to the DAC's buffer. This method requires proprietary programming skills; you need to write a dedicated driver for the cheaper asynchronous DACs, and/or be capable of more advanced firmware programming for the higher fidelity versions. Obviously this can expose the end user to driver, platform, and firmware issues.

 

[grokit]

Quote:

Originally Posted by sxr71 /img/forum/go_quote.gif Is there an effective way to decouple the noisy PC environment from our DACs? Another thing that does not engender confidence is a claim that a $1000 software program is helping matters. I thought that as long as you bypass Kmixer in Windows you get a bit perfect output. Isn't there a way cheap or free way to bypass OS resampling on a Mac?

On a Mac, there is really nothing to bypass compared to what Microsoft puts in your way with Windows (although it's getting better with Windows 7), and I just use a main USB port (not from a hub) for audio output, as it sounds even better to me than the built-in optical out. On a PC, I need an optical out, either built-in or from a sound card to bypass all of the internal noise and the default OS's latency, largely negating the need for an ASIO bypass protocol. IMHO, or course