Is there a better interface other than USB right now?

[plankwalker]

I know the question of USB vs SPDIF has been talked about a lot. I have done research on google and the forums here.

 

It seems that in many cases Asynchronous USB is better than SPDIF on a PC, a Squeezbox? an Apple TV? Im not sure about Toslink on Macs. I just bought a new DacMagic 100 (might be trading for a Bifrost) but so far the USB sounds just a little better than than the RCA SPDIF out on my Soundblaster X-FI Extreme music when set to ASIO and Bit perfect. I may try a DIY Galvanic isolation as well.

 

I know there are a lot of adapters like the Audiophilleo, and the USB card used on the CAPS server.

 

So is USB pretty much the best connection at this point? Since the "books" call RCA SPDIF the best, are there any PCI interfaces that would best USB? Somehow I dont think a Julia@ or M-Audio or something is at this point. That is just my layman assumption.

 

Is there a way to build a better interface for a PC or MAC perhaps? Linear power supplies or something? The whole Jitter vs PC noise/interference thing is really confusing to me. I know clean power in only a part of it, but something that takes advantage of SPDIF

[Akapod]

I think most of the pro gear uses Firewire (if it doesn't use a dedicated PCI card).

[esldude]

Well of course the Asynch USB then goes into the dac via SPDIF. But assuming you mean USB to Dac rather than SPIDF from computer to DAC, I think in general you are correct that USB is better. Is there something better than USB, I don't know, maybe some I2S direct connection. USB is most definitely better than Toslink on a Mac or any of the Airports. Better than a Squeezebox by most accounts. Even a V-link seems to get you lots of benefit for not all that much expense.

[goldenpiggy]

As previous poster said, you've got Firewire (most pro audio interfaces and some high-end home stereo DACs), MADI from RME, AES/EBU, MOTU PCI cards, and now Thunderbolt interface (Universal Audio).

[crisnee]

I don't think USB is necessarily always better, but it is always better than direct optical or coax out from the motherboard of a Windows pc. There are some very well made sound cards that some people tout as being equal to or better than many USB implementations. I think I would have gone sound card in one pc, if I didn't have two mini pcs which don't fit pci sound cards.

 

-Chris

[The Computer Audiophile]

This is a popular question. No interface is better than all others all the time. The biggest factor is implementation by the DAC designer. There's no free lunch with any design. USB for example may benefit from isolation between the DAC and computer. This isolation can increase jitter is not done well. Come designers don't believe in isolation. S/PDIF RCA isn't a true 75 ohm termination whereas S/PDIF BNC is 75 ohm. Then there is S/PDIF optical TosLink that is electrically isolated but frequently has increased jitter and poorer performance.

 

Don't be fooled into thinking what professional audio engineers use is better. I know many who will only use AES/EBU and won't touch FireWire. Contrarily FireWire is popular among many other pros. What really should be done is ask people why they use a certain interface and listen for yourself. Was FireWire around first? Were the initial Adaptive USB implementations lackluster? Are good FireWire devices less expensive? There are tons of factors that go into these decisions.

[Miska]

Ethernet is quite nice, it allows access from multiple computers at once, allows use of copper or fiber, it is always isolated, etc.

 

It is also faster than USB or Firewire. Level of complexity depends on implementation. Downsides are similar to USB devices that don't use UAC.

[The Computer Audiophile]

Ethernet is quite nice, it allows access from multiple computers at once, allows use of copper or fiber, it is always isolated, etc.

 

It is also faster than USB or Firewire. Level of complexity depends on implementation. Downsides are similar to USB devices that don't use UAC.

 

I like Ethernet for the most part as well.

 

I don't believe it's always isolated -> http://www.computeraudiophile.com/threads/4265-Network-Isolation?p=51343&viewfull=1#post51343

[Miska]

I don't believe it's always isolated -> http://www.computeraudiophile.com/threads/4265-Network-Isolation?p=51343&viewfull=1#post51343

 

Of course if you use shielded cable, the shield is connected. Just like if you use earthed mains, there's connecting link through the earth pin. That is normal, but all the pairs are always isolated unless the hardware is incorrectly designed.

 

If you use normal UTP cable, it should be always isolated. And of course there's always possibility to use optical ethernet instead.

[obt_man]

I guess Thunderbolt may become the best new interface for digital music transmission, but obviously the technology in its inception and we will have to spend some time before we see any Thunderbolt DACs available in the market. However, given the speed of computer audio penetration it may happen quite soon.

[Miska]

I guess Thunderbolt may become the best new interface for digital music transmission, but obviously the technology in its inception and we will have to spend some time before we see any Thunderbolt DACs available in the market. However, given the speed of computer audio penetration it may happen quite soon.

 

Thunderbolt with optical link would be nice too. However it will also require device specific drivers which for some reason seem to scare audio equipment manufacturers...

[Julf]

Of course if you use shielded cable, the shield is connected.

 

Just as with balanced audio cables, the shield doesn't have to be connected at both ends - one end is enough.

 

Just like if you use earthed mains, there's connecting link through the earth pin.

 

But what does it connect? In well-designed (and balanced) equipment, the earth doesn't have to

be connected to anything signal-related.

[Miska]

Just as with balanced audio cables, the shield doesn't have to be connected at both ends - one end is enough.

 

But of course unless you make custom cables or hardware that depends on where the cable is connected. Some equipment uses complete plastic connectors (ADSL modems typically) where such cable ends up having at least one end of the shield unconnected. Also many wall sockets don't connect shield if such is available, so it's connected only at computer end.

 

But what does it connect? In well-designed (and balanced) equipment, the earth doesn't have to

be connected to anything signal-related.

 

It should be connected to the equipment chassis when it's used. With ethernet that's of course not signal related, because the wire pairs are differential. That's the case for example with desktop computers. And it is commonly subsequently connected to the unbalanced signal grounds (USB/Firewire cable GND/shield etc).

[Musiq]

This is a very interesting thread. In my opinion, as the site founder, I believe that everything rely in the quality of the respective Transport/Dac sections. In the overall, there are DACs that are ugly in USB but have very good S/PDIF performance, bad AES but good ST Fiber, etc.

 

Regarding all these real time protocol (AES, SPDIF, TOSLink, USB Audio Class), none of them are reliable to transmit data WITHOUT error. In all these implementations there are issues with timing and jitter. At protocol level, the best implementation I have seen that allows to have pretty reliable clock is I²S. It also allows to transport information on low cost media, which I know is not at cable manufacturers taste

[mav52]

I don't think USB is necessarily always better,

but it sure is convenient

[dallasjustice]

This one is rarely implemented but shows the most promise as far as bandwidth and electrical isolation.

[Julf]

Regarding all these real time protocol (AES, SPDIF, TOSLink, USB Audio Class), none of them are reliable to transmit data

 

One of the good things with the computer audio revolution is that it allows to move to proper buffered data transfer protocols, where jitter and data loss are not a problem.

[goldenpiggy]

Regarding all these real time protocol (AES, SPDIF, TOSLink, USB Audio Class), none of them are reliable to transmit data WITHOUT error. In all these implementations there are issues with timing and jitter. At protocol level, the best implementation I have seen that allows to have pretty reliable clock is I²S. It also allows to transport information on low cost media, which I know is not at cable manufacturers taste

 

Just that I2S was designed for internal (internal as in within one enclosure or on a PCB) data transfer where you don't go over long distances or have to interface two components. There may have been proprietary implementation on a very few DACs in the 90s where you could interface externally over I2S, but it never caught on.

 

AES/EBU is pretty reliable and is the de facto in pro audio and broadcast. I regularly send 8 channels (4 stereo) 24bit/96Khz over 300ft on one CAT5 and it sounds great; never seen an issue. Ultimately it is up to how good the receiver is able to recover the data and timing off the data.

[The Computer Audiophile]

Just that I2S was designed for internal (internal as in within one enclosure or on a PCB) data transfer where you don't go over long distances or have to interface two components. There may have been proprietary implementation on a very few DACs in the 90s where you could interface externally over I2S, but it never caught on.

 

AES/EBU is pretty reliable and is the de facto in pro audio and broadcast. I regularly send 8 channels (4 stereo) 24bit/96Khz over 300ft on one CAT5 and it sounds great; never seen an issue. Ultimately it is up to how good the receiver is able to recover the data and timing off the data.

 

I've heard similar statements about I²S. I believe it was designed to go distances measured in single digit centimeters.

 

I've had great luck with AES/EBU as well.

[plankwalker]

In the book "The Complete Guide to High End Audio" the author refers to the Lynx AES16 as an option for a good interface. Id like to know how it compares to devices like the Audiophilio

 

http://www.lynxstudio.com/product_detail.asp?i=13

[The Computer Audiophile]

In the book "The Complete Guide to High End Audio" the author refers to the Lynx AES16 as an option for a good interface. Id like to know how it compares to devices like the Audiophilio

 

http://www.lynxstudio.com/product_detail.asp?i=13

 

 

I have an AES16 and have used it extensively. It's a solid product. If one's DAC has word clock output they can send clock back to the AES16 for a nice sonic improvement as well. I don't have any experience with the Audiophilio. The new Berkeley Audio Design Alpha USB is quite a few steps beyond the Lynx AES16 and about $1,000 beyond it.

[esldude]

Have an Audiophilleo. It is very good. Have a friend with a BADA. Just haven't convinced anyone to do a side by side yet.

 

I also agree about AES/EBU. Seemed an improvement on my transport over any coax. Wish the Audiophilleo had that as an option.

[barrows]

The direct answer to the original question is YES, with the following caveat: as long as everything is implemented very well; with a well designed asynchronous USB receiver, connected to a well designed, low noise, USB transmitter (like the SOtM USB card). Ethernet can be very good as well, but is proving to be much harder for most manufacturers to implement perfectly.

 

As for I2S, yes, it is true that originally ordinary I2S was designed for signal transmission within a single component, and using ordinary I2S for distances over about 10 cm can be problematic, but, it is important to point out that there is an alternative version of I2S using balanced signal transmission which is very robust for transmission over longer distances between separate components. This balanced type of I2S transmission is used by PS Audio in the PerfectWave Transport and PerfectWave DAC combination, and is also offered as an output option for Sonore Servers and input on Sonore DACs. Additionally, Twisted Pear Audio offers a DIY version of balanced I2S for it's DACs and a single ended to balanced converter which can make any single ended I2S signal balanced for transmission over distance.

Using balanced I2S can be a much better alternative to SPDIF for digital transmission between components because it does not have the compromise of embedding the clock signals and data signals on the same wire (requiring the jitter prone de-embedding of those disparate signals).