DACS: USB VS spdif

corElement

Well-Known Member
Joined
Dec 27, 2009
Messages
2,565
Points
113
Location
gurgaon / Delhi / NCR
I'm confused about something,

Suppose we have an external DAC which takes usb and spdif,
the spdif input comes from a soundcard in the form of toslink or d-coax...

Would this signal not be influenced by the sound card as vs direct USB data stream?

Now I'm not talking about 24/96 24/192

I'm talking about the quality of sound itself.

For example I have an onboard soundcard and xonar soundcard both have similar specs digital out but both digital outs sound heaven and earth apart.

Would the external dac also not be affected by this difference in sound as against a data stream like USB without a soundcard in between influencing the sound?

Also isn't spdif a lossy transport?
 
For whatever the reasons, there is a difference in how my DAC sounds with the two inputs (USB/Coaxial Digital).

The USB input sounds "hard", for lack of a better way to describe it. The USB input also throws a discernibly narrower sound stage. I prefer to feed the DAC with the coaxial digital output of the Xonar STX by a large margin. My DAC does not have an async USB input, BTW.

But I think we shouldn't generalise on this, as different DACs (are anecdotally supposed to) have different optimal sounding inputs.
 
But I think we shouldn't generalise on this, as different DACs (are anecdotally supposed to) have different optimal sounding inputs.

Absolutely, but in your case the spdif output is coming from a burr brown dac on the xonar which in a way has already been operated on.

The question comes into it's own when you use a poor quality spdif output to feed the dac vs the dacs consistant USB SQ.

The topic perplexes me a little lol.
 
For example I have an onboard soundcard and xonar soundcard both have similar specs digital out but both digital outs sound heaven and earth apart.

Can you explain?
How can digital out sound heaven and earth apart?

Did you check if the output is 24bit vs 16 bit between the two? Also is the sample frequency same?
If the bit rate and sample frequency are same, there shouldn't be any differences?
 
...
How can digital out sound heaven and earth apart?
...

It shouldn't, but in my experience, it does.

I base this on my experience with the digital out of my WDTV (1st Gen) and the digital out of the Xonar STX. Both were fed into the same DAC (first a Beresford Caiman, and later a Rega DAC). Everything else (interconnects, amplifier, speaker wire, speaker, listening position and room) were the same. But the presentation of the music was quite different.

I don't have the technical knowledge to talk about this, but I can speculate on one aspect: Won't there be a difference in the digital out if the way in which a digital stream is handled (before it is sent out to the DAC) is different? For example, a Juli@ has dedicated clocks (crystals? duh!) for the 44.1kHz family (and multiples) and for the 48kHz family (and multiples). I'm pretty sure an onboard setup will not have such precise/dedicated components, and both "families" will be handled by a generic component that uses firmware/software based processes to process both inputs before sending them to the DAC.

Correct me if I'm wrong here, my understanding could be very wrong and I'd love to learn more about this too!
 
Can you explain?
If the bit rate and sample frequency are same, there shouldn't be any differences?

As per my layman's understanding, lot depends on the clock. If a wave took 10 milliseconds to complete during recording, it should be completed in exactly 10 milliseconds during playback. Else the wave will get compressed or get expanded and the sound will never sound the same. Unfortunately, in real life, two clocks are never the same. Good soundcards have better clocks I presume. This deviation from the actual periodicity of the original wave is called jitter.
 
we did a onboard coax out vs. Juli@ coax out to the DacMagic and found no noticable difference.
Blasto - we should try it will the M1 sometime :) probably that may show some difference.
 
@Hydra/All

The clock difference comes into picture if the DAC processes absolute realtime data.
Most DACs as I understand come with some "buffer" to store the bitstream for a second or so and uses its own internal clock over buffered stream.

So the clock differences should not matter with high end DACs.

-All the above from the net.
 
In his review of the M1 DAC, Michael Lavorgna notes a difference between the USB and Toslink and he seems to prefer the former:

"My preferences in terms of listening to the M1 DAC were very nearly a toss-up between USB and Toslink the latter having a slightly more appealing edge while the former added a bit more weight and resonance. I bet some would describe the Toslink input as being more accurate. I wouldn't necessarily since the presentation that grabs you most is by my definition the most accurate...

Through the USB input as well as the Wavelength Proton, Ms. Hinrich's keyboard was more of-a-piece. I mention this albeit subtle difference because it highlights and perhaps exaggerates the character of the M1's Toslink input in that it tends to emphasize incisiveness which can impart a spotlit quality to some upper frequency information and a leaner feel to music overall."


Musical Fidelity M1 DAC (the newer one) | AudioStream
 
In his review of the M1 DAC, Michael Lavorgna notes a difference between the USB and Toslink and he seems to prefer the former:

"My preferences in terms of listening to the M1 DAC were very nearly a toss-up between USB and Toslink the latter having a slightly more appealing edge while the former added a bit more weight and resonance. I bet some would describe the Toslink input as being more accurate. I wouldn't necessarily since the presentation that grabs you most is by my definition the most accurate...

Through the USB input as well as the Wavelength Proton, Ms. Hinrich's keyboard was more of-a-piece. I mention this albeit subtle difference because it highlights and perhaps exaggerates the character of the M1's Toslink input in that it tends to emphasize incisiveness which can impart a spotlit quality to some upper frequency information and a leaner feel to music overall."


Musical Fidelity M1 DAC (the newer one) | AudioStream


^ this review again touches upon the gap I'm trying to highlight.
H'es playing it through a device which already HAS a dac, so it's TWO dacs working on the signal instead of one.


Let me put it another way.

Take a crappy (onboard dac ) soundcard with toslink
Take a goood (xonar dac) soundcard with toslink
Take USB from 1 pc
Take USB from 1 laptop

Connect all 4 to a DAC with 4 inputs respectively (hypothetical scenario)

In theory both toslinks should sound the same but reality is they wont and dont.
The only way they'd sound the same is if they were both the exact same hardware which they are not.
They are using dacs inside the soundcard so basically its soundcard dac -> toslink -> external dac
vs
Two USB inputs from two different pc's directly hookedup to the external dac, aka 1 single dac.

Won't the signal be cleaner if theres just 1 DAC working on it (USB+DAC) instead of two (soundcard/CDP+DAC) ?
 
Last edited:
Me also learning, only, about these technologies, but I think I see a basic error there:
They are using dacs inside the soundcard so basically its soundcard dac -> toslink -> external dac

Translate DAC: Digital to Analogue Converter.

No, your data is not being Converted to Analogue before being passed to the Toslink (or any other digital) output. The fact that it is being processed by the card does not mean that it is passing through the DAC chip. My theory: it only does that when you want analogue out.

All digital is data, and all data is just 1s and 0s --- but not all circuits are equal. The soundcard is converting the data from it's native, computer-motherboard state (which I suppose is a parallel bus) to a serial output of a different protocol. The justification for a more expensive soundcard over a cheap, built-in, MB sound section, is that it will do this more reliably. I can accept this, but I can't speak from experience, as these days I use only analogue-out from my PC.
 
Very very confusing!!
One thing digital can do properly is lossless transmission through cables.
Why is even that not happening right?

Is data through HDMI cable different too then?

Isn't USB mode like USB file copy?
 
@BLASTO, the issue of contention is about the point of origin, and not about what happens between the point of origin and the receiving end.

We're wondering whether the digital stream that originates from a sound card is superior to the digital stream that originates from the onboard SPDIF header on a motherboard, and whether the digital output through the USB protocol is of lesser quality than a digital stream from the sound card or motherboard. And so about whether the device/component that "prepares" the digital stream (before it is transmitted) matters.
 
Last edited:
I agree with Thad. In the case of the Toslinks, the 0s and 1s aren't being converted by soundcard's DAC into an analogue signal before being sent to an outboard DAC. I imagine it would work like a CDP connected to an external DAC - the 0s and 1s are read off the disc and sent through S/PDIF to the external DAC, bypassing the DAC within the CD player.

I am sure there must be some differences in the way different soundcards read the data off the hard drive before sending them out of the Toslink. Which may explain the difference in quality...
 
From what I've learnt..the most "untouched" signal..the "shortest path"..the "least number of junctions" between the origin and destination is the cleanest signal of them all,

However things are never that simple, in case of an onboard soundcard you have a huge number of issues originating out of how the soundcard handles the signal.

For example my onboard soundcard, when it's not in use, the signal drops and starts to create low level distortion and when something is played again theres a high pitched sound before audio plays again. This is manipulation of the signal. This manipulation of the signal is what bothers me in the head. The fact that it's being manipulated at ALL is whats bothering me. There is no way the signal is not being manipulated in a soundcard regardless of toslink coax or analog.

When I compare this to USB, that entire manipulation is missing, its

motherboard > Dac > Amp

Vs

Motherboard > Soundcard dac > External dac > Amp

If you ignore soundstage and gain boundaries and focus on 100% pure signal detail, have you guys founds USB to be more detailed or spdif?

Also, isnt SPDIF a lossy medium of transport or is that limited only to codec? Do you transmit the spdif as PCM or Dolby or what? (I understand this dependant on what soundcard you use but if you have this setup please tell us)
 
besides 1s and 0s, the bits need to arrive just in time as music is nothing if timing is incorrect. Computer's CPU load from 100s of applications running may prevent it from being able to send the bits as and when music listener is expecting them.
 
Just imagine your computer's motherboard as a busy main road with turnings. Take first left for the soundcard in a PCI slot, take the right turn for USB, second left for hdmi. The CPU is the traffic policeman, although certain components have the right to do stuff by themselves (bus-mastering). It is also an eight-lane highway (parallel) and, to take some of those turns, the bits have to line up and be sent, one by one (serial) up or down a narrow path.

The traffic policeman is a bitch. He is quite likely to walk out in front of some audio data bits and hold them up while some video data bits take priority. This kind of problem can get so bad that the machine simply will not play audio properly, but mostly, it doesn't. The audio equipment is designed to work in and with this very audio-unfriendly environment. The essential reason for this is that audio is, necessarily, a real-time experience, but the PC design does not recognise it as such. Why? I have no idea*, I guess there is a historical answer: I only know that I suffered by it.

in case of an onboard soundcard you have a huge number of issues originating out of how the soundcard handles the signal.
Not really, no more than the whole problem just mentioned. Sound cards have been around a while. The stuff that is built-in, these days, is heaps better than the add-in cards we used to pay extra money for years ago (Soundblaster? Exactly.)
For example my onboard soundcard, when it's not in use, the signal drops and starts to create low level distortion and when something is played again theres a high pitched sound before audio plays again. This is manipulation of the signal.
I think this is not manipulation --- but a fault, either in hardware or software. I've used a few soundcards, including built-in; I used to use digital-out; I never came across anything like that.

Also, isnt SPDIF a lossy medium of transport ...?

No, the concept of "lossy" does not apply to the digital signal at this stage. If the original file was compressed, lossy or lossless, it will have been uncompressed by your PC's CPU. There are exceptions, such as transmitting compressed data over wired/wireless net to a Squeezebox, but that data is still very much in the computer domain, not the audio domain.

Your operating system, quite apart from the traffic junction scenario I just mentioned, may well be manipulating your data, and that manipulation may be unwelcome. If you want the same bits that were read from the hard disk to be delivered to your sound card, then, for starters, you need to get Microsoft's fingers out of the pie, which is why people use and recommend alternative drivers.

There are also manipulations that may happen either in software or on your soundcard that you want and need to happen, for instance conversion of sample rates where necessary. Not all manipulation is evil.

motherboard > Dac > Amp

Vs

Motherboard > Soundcard dac > External dac > Amp

As mentioned before, please leave that "dac" out: it is not happening.

Actually, both paths are the same, and amount to

Motherboard -->Thing that converts to audio digital format --> Dac --> Amplifier.

USB is section of circuitry on your motherboard, so (whether built- or plugged-in) is your sound card. USB is not more direct. USB may be a Universal Serial Bus, but it knows when it has an audio device attached, and uses particular protocols to talk to it. That protocol is published; google will help you find it; it goes beyond my understanding (or need to understand).

USB is also, potentially, a very congested corner of your motherboard. It may be handling your printer, scanner, external discs, sound device, and more, all at once. Just because you're busy working doesn't mean you don't want to listen to music. With all that sharing going on, it can matter which USB socket you plug your sound device into.

(E&OE. Corrections and disagreements welcome!)


*Yes I have :eek:. Simply, the PC does nothing in "real time." It does a bit of this, then a bit of that, then a bit of something else, then back to this again. It communicates with hardware through interrupts. Audio interrupts may not have the highest priority. If they have to wait long enough, the music stops.
 
Last edited:
MacBook Pro> DacMagic Plus> AES Pre> AES Amp> MA Gold Speakers (using USB)

MacBook Pro> DacMagic (Optical to SPDIF conversion)> DACMagic Plus> AES Pre> AES Amp> MA Gold Speakers (using SPDIF)

All MIT cables

In my system both (USB and SPDIF) sound the same. I think it all depends on how they are implemented.

FYI: The Well-Tempered Computer
 
Just imagine your computer's motherboard as a busy main road with turnings. Take first left for the soundcard in a PCI slot, take the right turn for USB, second left for hdmi. The CPU is the traffic policeman, although certain components have the right to do stuff by themselves (bus-mastering). It is also an eight-lane highway (parallel) and, to take some of those turns, the bits have to line up and be sent, one by one (serial) up or down a narrow path.

This is a very helpful analogy :D

The traffic policeman is a bitch.

:eek:hyeah:

Not really, no more than the whole problem just mentioned. Sound cards have been around a while. The stuff that is built-in, these days, is heaps better than the add-in cards we used to pay extra money for years ago (Soundblaster? Exactly.)

I think this is not manipulation --- but a fault, either in hardware or software. I've used a few soundcards, including built-in; I used to use digital-out; I never came across anything like that.

True.. but it doesn't change the fact that it is, still happening. The only solution I can think of is that there is no way to avoid it other than getting a proper sound card where the sound is treated properly (I don't use the onboard one anymore, this is just hypothetical talk) which again is more money spent.


No, the concept of "lossy" does not apply to the digital signal at this stage. If the original file was compressed, lossy or lossless, it will have been uncompressed by your PC's CPU. There are exceptions, such as transmitting compressed data over wired/wireless net to a Squeezebox, but that data is still very much in the computer domain, not the audio domain.

I see, that is good to know because I was always confused about this.

Your operating system, quite apart from the traffic junction scenario I just mentioned, may well be manipulating your data, and that manipulation may be unwelcome. If you want the same bits that were read from the hard disk to be delivered to your sound card, then, for starters, you need to get Microsoft's fingers out of the pie, which is why people use and recommend alternative drivers.

What would you recommend for Xonar soundcards?

There are also manipulations that may happen either in software or on your soundcard that you want and need to happen, for instance conversion of sample rates where necessary. Not all manipulation is evil.

Quite an interesting bit of info there. I was running at 192KHz and never really tried lowering it for files which are around the 44-48khz mark. Thought I'd try out your logic and just lowered it..strangely vocals came up front and percussion got suppressed a little. Very interesting. Now I don't know which sounds better haha.
 
My media player will do sample rate conversion, but I too noticed some difference when actually setting the drivers (and thus the interface) to the same sample rate as the music file. Not extensively tested though.
What would you recommend for Xonar soundcards?
I give my self an entirely other heap of problems by using a firewire interface with Linux. With Windows, I never went further than XP, and just used foobar or VLC without anything other than a bit of cutting back on Win-Unnecessaries, so can't comment, but there is lots here on ASIO and, err, brain cell failure, wasapi? (yes, but I nearly typed wasabi!)
This is a very helpful analogy
Cheers :) It's my way of understanding stuff.
FYI: The Well-Tempered Computer
A vital source :)
 
Check out our special offers on Stereo Package & Bundles for all budget types.
Back
Top