Review and Measurements of Gustard DAC-X26

[Newk Yuler]

Considering the asynchronous nature of USB input, I see no reason for external reclockers of any kind -- the data is already reclocked by the USB receiver. That's been my experience with properly implemented modern DACs, both, in measurements and in listening tests.

What is the time period to which you refer "properly implemented modern DACs?" That's the reference I'm seeking. I appreciate your experience.

Amir measured Gustard X26 here with his AP, and I measured an older Gustard X20Pro (with much, much more modest equipment). The result puts jitter below -135dB mark with 0dBFS signal. This is far below audibility, even if you play music at the loudest setting that doesn't destroy your ears.

I understand what you're saying and again I appreciate your knowledge, but I believe Steve Nugent would say that jitter at that level isn't necessarily actual audible noise but garbage that would affect digital accuracy and cause audible issues, although he would say it like someone who better knows what he's talking about. I'm just a long-time disciple in his church. And you're a heretic! Grrr!

 

[pkane]

What is the time period to which you refer "properly implemented modern DACs?" That's the reference I'm seeking. I appreciate your experience.

I'd say the past 5 years. Probably more, but I wasn't using USB prior to that, and have only one slightly older USB DAC that isn't good at reclocking -- lots of jitter remains with four or five different reclockers I tried.

I understand what you're saying and again I appreciate your knowledge, but I believe Steve Nugent would say that jitter at that level isn't necessarily actual audible noise but garbage that would affect digital accuracy and cause audible issues, although he would say it like someone who better knows what he's talking about. I'm just a long-time disciple in his church. And you're a heretic! Grrr!

Jitter is not magic. If it affects audio then it must be present in the audio band at the output of the DAC. There are well known audible qualities to jitter. It has been studied, measured, listened to. Steve apparently isn't interested in anything that disagrees with his perception or the theories he constructed based on it. So there's very little left to discuss objectively.

 

[Krunok]

Considering the asynchronous nature of USB input, I see no reason for external reclockers of any kind -- the data is already reclocked by the USB receiver.

Exactly. Jitter doesn't exist in asynchronous communication, it is meaningless to speak about it in that scenario. USB receiver sends data to DAC chip via the clocking driven by high precision oscillators within device, usually 2 of them, one for 44.1 and multipliers and the other for 48khz and multipliers. Even in cheap Chinese DACs high precision oscilaltors that are covered wtih gold plated metal case to ensure temperature stability are used and jitter they produce is negligible and miles far from being an audible concern.

 

[wyup]

Exactly. Jitter doesn't exist in asynchronous communication, it is meaningless to speak about it in that scenario. USB receiver sends data to DAC chip via the clocking driven by high precision oscillators within device, usually 2 of them, one for 44.1 and multipliers and the other for 48khz and multipliers. Even in cheap Chinese DACs high precision oscilaltors that are covered wtih gold plated metal case to ensure temperature stability are used and jitter they produce is negligible and miles far from being an audible concern.

Asynchronous communication and reclocking doesn't eliminate a jittered signal. For example, a toslink 192KHz spdif optical signal is at the limits of toslink bandwidth, and I have seen an oscyllator image of plenty jitter signal waveform, and it is already difficult to the naked eye to discern 0s and 1s. If you feed this jittered signal to an asynchronous usb dac, it is going to have a hard time reclocking the signal.

 

[Krunok]

Asynchronous communication and reclocking doesn't eliminate a jittered signal. For example, a toslink 192KHz spdif optical signal is at the limits of toslink bandwidth, and I have seen an oscyllator image of plenty jitter signal waveform, and it is already difficult to the naked eye to discern 0s and 1s. If you feed this jittered signal to an asynchronous usb dac, it is going to have a hard time reclocking the signal.

Uhhh.. But you don't feed SPDIF signal to USB DAC via USB, do you?

USB DACs are typically connected directly to computer running software players which are reading the music files from storage or fetching stream from the Net, so no jitter there. As we said, USB communication is async so no jitter there as well. Once async packets reach USB receiver chip it buffers them and sends them synchronously to a DAC chip via internal link within the device but jitter there is negligible for the reasons I described.

 

[wyup]

I meant to an usb dac by optical toslink connection, not usb, sorry.
There is also jitter in usb. Computer clocks in usb signal is far from being perfectly synchronised and clean of interferences like ground-loop isolation and noise for audio, and that is the source of jitter even after asynchronous reclocking: distorted binary signal.
Usb data has error correction algorythms and restitution is different.

Yesterday I made a curious discovery. Same gear and same source files on same hard drive sounded different in two Windows installations on different hard-drives with same software. OS bus congestion makes a difference in audio jitter signal that goes to firewire dac.

 

[Krunok]

I meant to an usb dac by optical toslink connection, not usb.
There is also jitter in usb. Computer clocks in usb signal is far from being perfectly synchronised, and that is the source of jitter: mis-aligned binary signal.

You cannot speak of jitter when sending data asynchronously. If that is not clear to you than you don't understand the concept of jitter.

Jitter is variation in periodicity. Only synchronous communication (like the one used for SPDIF protocl) has periods at which data is sent. With asynchronous communication data is not sent by periods defined by some clock so speaking of jitter with asynchronous communication is meaningless by it's definiton.

 

[Sangram]

The nature of asynchronous depends on the driver that is sending the data, as well as the overall underlying hardware latency and timing. All are not built equally or written equally well.

If what you're saying was true, then every single DAC jitter measurement using (for example) the XMOS u8 chip would be exactly the same. The fact they aren't means there are more balls in play than it would first seem.

As another example, almost all pro audio interfaces have some sort of hardware mixer that is interfaced to the host computer using a hardware driver, and all of them use asynchronous transmission, even the interfaces form the early part of the century. Even they don't have identical (jitter) measurements, which again casts doubt on the theory that all async transmission is the same and is immune to jitter. That's totally untrue.

 

[wyup]

Well, I am not a expert on this, but asynchronous usb mode communication doesn't guarantee a perfect clocked signal afterwards. It is well agreed that USB is not a good medium for audio transmission, because it has noise from the computer and imperfect clock. Jitter comes from a clocked signal and USB is not free from jitter. Asynchronous mode means the DAC clock is the master signal, but the usb data comes clocked in some way, so there may be timing problems too.

I'm going to prove that jitter still exists in Aynchronous USB:

source 1: http://www.audiophilleo.com/home/definition/asynchronous USB

"Advantages of using asynchronous USB (together with the Audiophilleo's regenerative power supply) include:

• S/PDIF output quality is now completely independent of USB interface jitter. The particular computer used as a media server, or type of USB cable will no longer effect sound quality in a significant way. Be sure to try your existing USB cables, but differences between them and those supplied with each Audiophilleo are likely to be small.
• Fixed Frequency Clocks can be used because the master clock circuitry can now be in the audio device. Much lower jitter results, because it's not necessary to employ phase locked loops or similar approaches for syncing with "upstream" clocks."

source 2: https://www.cambridgeaudio.com/en/blog/our-guide-usb-audio-why-should-i-use-it

Asynchronous – this is the most complex to implement but it is a huge improvement on the other types. This is because it requests the data packets to be sent in time with its own clock’s timing, thus providing the lowest jitter and sounding by far the best.

 

[Krunok]

The nature of asynchronous depends on the driver that is sending the data, as well as the overall underlying hardware latency and timing. All are not built equally or written equally well.

If what you're saying was true, then every single DAC jitter measurement using (for example) the XMOS u8 chip would be exactly the same. The fact they aren't means there are more balls in play than it would first seem.

As another example, almost all pro audio interfaces have some sort of hardware mixer that is interfaced to the host computer using a hardware driver, and all of them use asynchronous transmission, even the interfaces form the early part of the century. Even they don't have identical (jitter) measurements, which again casts doubt on the theory that all async transmission is the same and is immune to jitter. That's totally untrue.

What I'm saying is not related (only) to USB communication but to ASYNCHRONOUS communication in general. As it is asynchronous there is no sync clock to follow thus no jitter as a measure of variation from that sync clock. With asynchornous communication data transfer is executed by the use of pulses sent when the previous operation is completed rather than at regular intervals.

For that reason jitter measurement with USB communication doesn't make any sense. You can measure jitter only with SYNCHRONOUS communication like SPDIF.

 

[Krunok]

Well, I am not a expert on this, but asynchronous usb mode communication doesn't guarantee a perfect clocked signal afterwards. It is well agreed that USB is not a good medium for audio transmission, because it has noise from the computer. Jitter comes from a clocked signal and USB is not free from jitter. Asynchronous mode means the DAC clock is the master signal, but the usb data comes clocked in some way, so there may be timing problems too.

 

[wyup]

I am educated enough to discuss this. You should discuss what I said and referenced above and read it well before posting jokes.
Other people is questioning your claim too. Referenced sources say jitter still exists in Asynchronous USB.

 

[Krunok]

I am educated enough to discuss this. You should discuss what I said and referenced above and read it well before posting jokes.
Other people is questioning your claim too.

Well, I told you everyhting about asynchronous communication (like USB) you need to know and you replied with this:
" It is well agreed that USB is not a good medium for audio transmission, because it has noise from the computer and imperfect clock."

As USB suffers neither from noise (no digital communication suffers from noise) nor it has imperfect clock (asynchronous communication is not driven by clock, if it was it would be called SYNCHRONOUS), this means I was not able to teach you anything so you either try to educate yourself or let someone who has better teaching abilities than me to try to teach you. But I really recommend you restrain yourself from posting until then.

 

[wyup]

Well, I told you everyhting about asynchronous communication (like USB) you need to know and you replied with this:
" It is well agreed that USB is not a good medium for audio transmission, because it has noise from the computer and imperfect clock."

As USB suffers neither from noise (no digital communication suffers from noise) nor it has imperfect clock (asynchronous communication is not driven by clock, if it was it would be called SYNCHRONOUS), this means I was not able to teach you anything so you either try to educate yourself or let someone who has better teaching abilities than me to try to teach you. But I really recommend you restrain yourself from posting until then.

You don't have to teach me anything. I am claiming referenced sources. Jitter still exists in Asynchronous.

asynchronous communication is not driven by clock, if it was it would be called SYNCHRONOUS

In Asynchronous mode the master clock circuitry can now be in the audio device. If you can't read it is not my problem.

As USB suffers neither from noise (no digital communication suffers from noise)

A pc usb connection is not digital, there is a medium carrying it and a powered connection, and noise from the computer behind it. It's commonly agreed. The signal is pulses on a analog voltage waveform. Educate yourself and don't waste my time.

 

[graz_lag]

... jitter still exists in Asynchronous USB.

Correct.
My 2012 Naim DAC-VI was declared by Naim at a total correlated jitter (random and sinusoidal) of 125psec, and not "zero"psec (0psec).
(Measured via the DAC’s USB input, sampling rate of 44.1kHz.)
The Naim DAC-VI used asynchronous mode USB data transfer.
Yes, Naim, not "My Sister Audio Device Industries" ...

 

[yummy]

if there's gonna be some tear down, just wonder if x26 using FPGA as spdif receiver, usually gear had ak4118 or similar chips

 

[maty]

Maybe it is a good idea to open a new thread about the jitter / noise / ... generated by the computer. As far as I know there are no measurements about it.

It is a problematic issue. I decided to verify it by myself and after optimizing my W10 Pro 64 bits to play multimedia the sound is much better than before. You also have to optimize the software players. And, if possible, use Kernel Streaming.

With Linux I have been unable to obtain such sound quality in my second audio system.

 

[Krunok]

Here's my last attempt to try to explain.

Block diagram of a typical USB DAC looks like this:



Green part, which covers asynchronous USB communication is green. Asynchronous in greek means "not guided by time" which is true as there is no clock at which rhytm data is sent, instead the receiver (USB receiver in DAC device) is asynchronously ("not in regular time intervals") asking for packets from sender (USB sender in PC). As there is no clock to which timing can deviate from there is absolutely no sense to talk about jitter in any kind of asynchronous communication. USB receiver in DAC (like XMOS X208 chip, for example) is writing received data into it's FIFO buffer. DAC chip is reading from that same buffer synchronously (with clocking provided by local oscillators in DAC device) so only in the red part of the data flow between USB receiver and DAC chip you have jitter, but that one is so small you can absoltely forget about it.

 

[pkane]

A pc usb connection is not digital, there is a medium carrying it and a powered connection, and noise from the computer behind it. It's commonly agreed. The signal is pulses on a analog voltage waveform. Educate yourself and don't waste my time.

Jitter generated by USB isochronous protocol is not the same as sample jitter. It can't be, since USB doesn't carry individual samples or the individual sample clock. USB carries 125μs packets, each containing a number of samples. Any jitter generated by the 125μs USB clock will be around the 8KHz mark and you'll see it on the jitter measurements with an AP or other devices. In fact, you can see in the X26 jitter measurement at around -145dB from 0dbFS J-test signal. Are you saying this is audible?

 

[Krunok]

Jitter generated by USB isochronous protocol is not the same as sample jitter. It can't be, since USB doesn't carry individual samples or the individual sample clock. USB carries 125μs packets, each containing a number of samples.

That is correct, but strictly speaking that is not jitter but "packet noise".