USB vs. SPDIF/TOSLINK should there be a sonic difference?

This topic has been debated much elsewhere. But the critical question of whether there is any sonic difference between "competent" interfaces of each kind has not been settled in any rigorous manner, though anecdotes abound.

Since the system requires a sender, a medium, a receiver and then circuitry to decode/unpack/de-interlace/buffer the received bits before they are converted back into an analog waveform there are too many variables to give a unitary answer. To the best of my knowledge there have been no serious double blind tests testing this.

I have a device that has coax, optical and USB interfaces and I have done time-aligned rapid switching testing between the different interfaces using the same signal - however they were not done blind so the results are worthless

there simply is no cut and dry answer to this question, and given that so many companies do not thoroughly test or share test results of their own equipment, it will be nearly impossible to get a consistent and accurate answer that does not involve the phrase "to my ears". 

furthermore every company has their own design methods, their own standards, their own "tests" or lack thereof, and one companies USB implementation or jitter removal process may be quite different from another.

that said

companies like benchmark with their dac1 series which are extremely thoroughly tested  and very well reviewed, will indeed claim that you will not hear a difference between coax/usb/spdif and will claim equal "filtering" through all inputs. 

I think the best you can do is buy from a company that thoroughly tests, documents, and is either trusted in the industry and/or 3rd party tested/reviewed.

This may be a stupid question but is there any way to get a 24/192 signal out of my PC to a USB DAC short of using a USB-Toslink cable/converter box (like directly from optical out of my sound card)? Also, is there any difference between optical vs. Toslink next to Toslink vs. USB? Can USB sound better than both or vice versa?

There are some DACs that can do 192kHz over USB but the USB interface chips that allow it are usually expensive and hard to implement so DACs that use them are usually pretty expensive.  They also usually require proprietary drivers on most systems and those often cause trouble as well.

Regarding what sort connection is best it really depends on implementation.  USB is usually a bit more difficult to implement properly but a properly implemented USB connection isn't going to be audibly different from a properly implemented coax connection.

Does your sound card do 192kHz over optical?  I didn't think optical went over 96kHz.  Coax can, but your sound card has to support it.  If you just want to play 192kHz content and get it to your DAC1 you're best off adding a high quality sample rate converter to your music player software and using your DAC1's USB input.  192kHZ doesn't really provide any measurable advantage over 96kHz.  You DAC1 already uses asynchronous sample rate conversion to change everything to ~110kHz before feeding it to the DAC chip anyway.

Adding a USB to S/PDIF converter in the middle won't do much of anything useful.  Its just another possible source of jitter and distortion.  The DAC1 should probably reject whatever jitter gets thrown at it but even if the 192kHz sample rate did any good the DAC1 is just going to turn it in to ~110kHz anyway.

What about the theoretical capacities of the two interfaces? Is anyone theoretically better than the other? I know USB wasn't made with audio in mind and optical was so maybe that might affect the sound at least in theory?

Neither one really.  They're both so open ended that you can't say anything definitive.

Well, there are some theoretical differences. Like galvanic isolation, which you always get with optical but not always with usb interfaces. Also, s/pdif is a standard of sending data and USB is a connection standard, so they are a bit different things. Most DACs with USB inputs simply code the incoming data into s/pdif, then it runs to a reciever chip where it's coded to I2s, and then runs to a DAC chip. So the difference is where s/pdif coding happens, in your soundcard which is connected to a DAC, or inside the DAC.

If the USB reciever chip codes the signal directly into I2s then it's a different story, as you can see it's more complex than just having usb input in your DAC.

There are quite obviously differences between the two interfaces but the question is if or how they matter.  Both have different sets benefits and drawbacks but in this application neither contains either a disadvantage that can't be worked around or an advantage that the other can't catch up to in either theory or practice.

They require different implementations but in the end they can both achieve the same quality.

To me, personally, direct l2s conversion from USB looks better, at least in theory. Proper s/pdif connection is somewhat troublesome, tight impedance matching with coax and supposedly higher jitter with optical. Coding the clock and music into one signal is also supposedly a cause for jitter. Ofc modern reciever chips can clean up a lot of jitter if well implemented, so it's nothing that can't be handled with. But it's like polishing a turd if we can simply not have such turds in the first place.

USB chip that outputs l2s theoretically looks much simplier. As long as it happens in one box and we don't have to use cables I see no advantage of s/pdif over l2s, while the latter has less disadvantages that have to be worked around. Manufacturers never solve all the problems unless it's some kind of top range product, so sticking with less problematic solution is more likely to produce better results, and that's how these solutions may vary. At least I think so

It looks like I left out something fairly important that I meant to say in an earlier post.  I had started writing something longer but just trimmed it down when it wasn't going anywhere useful.  It was basically that either can be made good enough that you'll have to use an audio analyzer to tell them apart instead of just your ears.

If you're talking absolute performance then one probably does have an advantage over the other.  I haven't looked at the specs in detail enough to come to a conclusion on that and didn't think it was relevant in this context.

I've found this article, I think it explains things better than either of us:

http://www.audialonline.com/html/articles/spdif_or_usb/1.php

Quote:

Originally Posted by nick_charles 

This topic has been debated much elsewhere. But the critical question of whether there is any sonic difference between "competent" interfaces of each kind has not been settled in any rigorous manner, though anecdotes abound.

 

Since the system requires a sender, a medium, a receiver and then circuitry to decode/unpack/de-interlace/buffer the received bits before they are converted back into an analog waveform there are too many variables to give a unitary answer. To the best of my knowledge there have been no serious double blind tests testing this.

 

I have a device that has coax, optical and USB interfaces and I have done time-aligned rapid switching testing between the different interfaces using the same signal - however they were not done blind so the results are worthless

 

 

Great comment.

I use a Monitor 01  USB-to-SPDIF gizmo between my PC and my DAC, I can't say if the 'overall' sound is better, but using the USB input on the DAC straight from the PC I used to get a certain amount of random clicks, like maybe one every five or ten minutes. I don't get any clicks using the USB-SPDIF gizmo.   I don't know why this should be, but in my setup, I was getting clicking no matter which of my 5 or 6 DACs I used.  Irritating loud clicks.  Maybe some ground loop between the DAC and the PC that the USB was participating in? And the USB-SPDIF gizmo isolates this?  I dunno.  And, of course, this is some kind of "problem" not an overall sound quality thing.

People SAY that SPDIF sounds better than USB.  I haven't done any comparison listening, but I am skeptical.  I do know that when Stereophile tests DACs, they tend to test a little better on SPDIF inputs.  Using the USB inputs usually results in somewhat higher low-level nonlinearity. These are measurements, not listening tests. I don't know if these differences that Stereophile measures are audible, but the measured differences DO seem real.  Not all the DACs they test show this; some of the expensive ones they tested recently had nearly identical USB performance.

Also, these differences show up more for red-book digital (16/44.1) When using higher bit rate / sample depth material the differences lessen between USB and SPDIF.  Again, I'm not sure about audibility. 

(Many SACDs and high-rez files are MASTERED DIFFERENTLY from their red book counterparts, and so, YES, they sound different from the CD. But given the same file in red book and high rez versions it seems supportable to say that people can't hear the difference.)