|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
192.253.240.236
In Reply to: RE: SPDIF vs. XLR posted by jllaudio on August 30, 2016 at 20:21:27
Hi,
Unless the DAC rejects jitter very well (which means it has some form of variation on the genesis digital lense build in), all long cables will degrade signal quality and add jitter.
Given that where there are plugs and sockets mismatches are almost unavoidable, the cable system with the greatest mismatch at the plugs will usually do worst, presuming all other parts of the system (Source, actual cable, DAC's input) are correctly engineered (which in some cases is a big presumption).
If you compare the Canare "75 Ohm" RCA Plug and generic 75 Ohm BNC Plug, you will find one having around 55 Ohm CI and the other around 65 Ohm CI. Generic RCA Plugs come in at around 45 Ohm CI, as they were designed for 50 Ohm systems.
For AES-EBU the nominal impedance is 110 Ohm but XLR Plugs are more like 300 Ohm. Any of these is sufficient to create reflections, but XLR will be worst and BNC will be least bad.
The Cable also comes into this, especially at 5m length and 192kHz sample rate, where delays in the cable become material.
The precise nature of the SPDIF receiver is also important, some of these chips are designed for AES-EBU signal (up to ten time the signal level of SPDIF) and barely lock with SPDIF signals and as a result have much higher jitter than needed to be in practice and with SPDIF signals.
Ultimately non of these connections (RCA, BNC and XLR) are ideal and I recommend keeping SPDIF/AES-EBU connections as short as possible (preferred << 12").
My advise would be to extend the USB Cable run by using any of the various "USB Repeater" Gizmo's as extender. At a recent show I demonstrated this with a total of 12 generic 1.5m USB Cables daisychained using B->A adaptors and four iFi iPurifier 2.
Not only was it an > 18m run of USB, it used very generic cables and many adaptors and created a seriously bad condition. Even removing only one of the iPurifier 2 from the chain dropped the connection.
In practice I cannot say that I heard much if any difference with this massive kludge of USB Cabling (I cannot recommend using it like that of course) over just using 1pcs of the Cables, no adaptors and a single iPurifier 2 at the end of the 1.5m cable.
Incidentally, no need to use iFi's product, I am sure the alternatives out there can do the same job just as adequate. Put at the end the particular device you like the sound best with your DAC or USB-> SPD converter.
Ciao T
At 20 bits, you are on the verge of dynamic range covering fly-farts-at-20-feet to untolerable pain. Really, what more could we need?
Follow Ups:
Thorsten wrote:
"For AES-EBU the nominal impedance is 110 Ohm but XLR Plugs are more like 300 Ohm. Any of these is sufficient to create reflections, but XLR will be worst and BNC will be least bad."
The 110 Ohm characteristic impedance of an AES/EBU digital cable is actually such a wide-range specification that it almost seems arbitrary. The actual specification falls somewhere between a generous 110 ohm ± 20% (88-132 ohm) range. I've built several DIY AES/EBU digital cables albeit short-length 0.5m cables, with the only stipulation being a simple twisted pair with a very tight twist ratio along with cable resonance damping, and find the AES/EBU digital interface to be subjectively superior to any S/PDIF cable interface, regardless of lacking any hands-on measurements, IME.
A more meaningful caveat of an AES/EBU cable interface is the notion that a balanced interface is less sensitive if not insensitive to cable and connector quality vs. a single-ended rca interconnect cable interface. A typical pro audio balanced cable and pro audio XLR connector can present just as lackluster performance as a mediocre single-ended rca cable tends to present. It took a number of DIY experiments to discover the importance of using a true high-performance conductor and XLR connector in order to achieve a more than satisfactory presentation via an AES/EBU interface, IME.
Hi,
I suspect your DAC uses a Cirrus Logic or AKM SPDIF receiver. These work poorly with SPDIF signals as they are designed for direct interfacing to AES/EBU which has at least ten times the common SPDIF signal levels. Unless suitable active signal conditioning is added SPDIF with these devices is much worse than it has to be.
And yet are the most common SPDIF devices and usually using the pretty terrible datasheet application too, all the way down to the inappropriate (slow slope, low bandwidth) transformer models recommended in the app note.
These receivers practically need AES/EBU signal levels to trigger cleanly.
Also, 0.5m is still pretty short, depending upon your source's rise time and sample rate this may be short enough to avoid any of the transmission line issues that cause reflections and signal distortion.
Try it with 5m runs.
I find once the DAC is equipped with a decent internal clock and a memory buffer there is so little difference between SPDIF, AES/EBU and USB as long as they are all galvanically isolated, that I'd be hard pressed to say I hear any.
Ciao T
At 20 bits, you are on the verge of dynamic range covering fly-farts-at-20-feet to untolerable pain. Really, what more could we need?
The Sabre DAC chip features a built-in S/PDIF decoder with inherently low jitter. It's actually a primary reason why I chose a Sabre DAC, since I don't use the DAC for a typical USB application.
See link:
Hi,
> The Sabre DAC chip features a built-in S/PDIF decoder with
> inherently low jitter.
Nope, it doesn't.
It features an SPDIF decoder with absolutely no jitter rejection whatsoever (meaning it's output has nanoseconds of jitter) and follows it with an asynchronous sample rate converter that attempts to clean that mess up.
That is not quite the same as "low inherent jitter". Heck, it's a whole other ketle of fish.
Many of those who have experimented with the Sabre find it performs better if the ASRC is placed in bypass mode by making the Sabre's clock linked to the source, even if measured jitter increases, but especially if jitter levels can be kept as low as if employing the ASRC.
But what matters in the end is that you enjoy the music.
Ciao T
At 20 bits, you are on the verge of dynamic range covering fly-farts-at-20-feet to untolerable pain. Really, what more could we need?
Fair enough. I misinterpreted the information. My impression was that the built-in S/PDIF decoder of the DAC chip helps to provide beneficial jitter reduction. Nonetheless, the Sabre DAC design is said to involve effective jitter reduction circuitry/methods.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: