[Jimster480]

Google and Wikipedia are cool sources, but you should look for all the definitions. For instance "analog signal".

Electrical signals are analog. Your USB is transmitted (most of the time) over electrical wires. The representation of the data is digital, but the transmission is good old analog.
Same goes for optical transmissions. Unless you go down to the quantum theory, light is a wave, as well. It sure does behave like one for the considered dimensions.

Sorry but its just not the case.
Digital signal is not analog signal.
Power is not analog.
 
Square "waves" are not analog, its 1/0 there are no "waves" they are boxes.

 

[Jimster480]

 
They are nice search Engines/ Sources but I have spend my time behind a bench and I have yet to see the use either of those sources to produce anything that support that support this, however  If we are going to get down to the physics level that's not where I was approaching this from and I get the strong feeling that's where you are based on your  "light is a wave" comment. The Non Physics differentiation between Analog vs.Digital signals is one is a continuous waveform of infinite values while the latter is discrete and binary values. Experience and google and wiki say so . In the non physics world there is a clear differentiation in the signal type and characteristics and that is where I am approaching this from.  I think we are looking into different windows of the same house. If you have the sources that do go down to the "physics level" that will not make my eyes water I really would like to read it as I am approaching this from a pure electronics stand point of what you see on a scope and encoding methods etc (the dining room window, essentially finished the product) and you are approaching this from a stand point of a signal as it travels on the wire (the Kitchen Window, where the the raw Ingredients live).

Exactly, these other guys don't konw what they are saying

 

[cishida]

Sorry but its just not the case.
Digital signal is not analog signal.
Power is not analog.

Square "waves" are not analog, its 1/0 there are no "waves" they are boxes.

Both of you guys have a point - neither is completely wrong.

At a high level / consumer level digital and analog signals are "completely" different.

But from a computer or electrical engineering perspective - once the digital signal hits the transmission medium (air, copper, optical fiber, etc) it really has entered what we would call the analog domain where you're dealing with distortion, interference, modulation, attenuation, etc.

Yes, Ethernet and LTE are "digital" in that they exist to transmit 1s and 0s, but I can guarantee that things get plenty "analog" messy when your 1s and 0s interfere with other signals using the same media. The physical layer of digital transmission protocols mostly deal with converting between the potentially very messy analog domain and the pretty bits a computer can understand. use your cell phone or laptop in a crowded area it has to deal with plenty of signals that are not a pretty square waveform.

engineers / scientists designing digital transmission protocols cannot think of a digital signals as perfect square waveforms because they often are not.

Here's an analogy: in theory punch cards store perfect digital info (punched or not punched) but in the real world someone has to decide what it means when a chad is still hanging on by a corner.

 

[Jimster480]

Both of you guys have a point - neither is completely wrong.

At a high level / consumer level digital and analog signals are "completely" different.

But from a computer or electrical engineering perspective - once the digital signal hits the transmission medium (air, copper, optical fiber, etc) it really has entered what we would call the analog domain where you're dealing with distortion, interference, modulation, attenuation, etc.

Yes, Ethernet and LTE are "digital" in that they exist to transmit 1s and 0s, but I can guarantee that things get plenty "analog" messy when your 1s and 0s interfere with other signals using the same media. The physical layer of digital transmission protocols mostly deal with converting between the potentially very messy analog domain and the pretty bits a computer can understand. use your cell phone or laptop in a crowded area it has to deal with plenty of signals that are not a pretty square waveform.

engineers / scientists designing digital transmission protocols cannot think of a digital signals as perfect square waveforms because they often are not.

Here's an analogy: in theory punch cards store perfect digital info (punched or not punched) but in the real world someone has to decide what it means when a chad is still hanging on by a corner.

Yes I know what you mean, because there can still be power distortion.
But digital signals do not have interference in the same way as an analog signal.
There is no "static" frames are either transmitted or not transmitted.

There aren't clean 1's and 0's or "dirty" 1's and 0's.
Either you are getting the proper signal or you aren't.

 

[winders]

  There aren't clean 1's and 0's or "dirty" 1's and 0's.
Either you are getting the proper signal or you aren't.

 
This is what you are missing....there are dirty 1's and 0's!!

 

[cishida]

 
Yes I know what you mean, because there can still be power distortion.
But digital signals do not have interference in the same way as an analog signal.
There is no "static" frames are either transmitted or not transmitted.

There aren't clean 1's and 0's or "dirty" 1's and 0's.
Either you are getting the proper signal or you aren't.

Yes, but in digital audio timing is also important - is not quite as black and white as either getting a "proper" signal or not. You can be somewhere in between.
 
1. Variations in the timing of those 1s and 0s can be audible (jitter).
2. If an error occurs - say the signal for a set of 1s and 0s are unreadable -- there is no time to ask for re-transmission. The receiver just has to play whatever it gets and move on. These errors can also be audible.

 

[Jimster480]

 
Yes, but in digital audio timing is also important - is not quite as black and white as either getting a "proper" signal or not. You can be somewhere in between.
 
1. Variations in the timing of those 1s and 0s can be audible (jitter).
2. If an error occurs - say the signal for a set of 1s and 0s are unreadable -- there is no time to ask for re-transmission. The receiver just has to play whatever it gets and move on. These errors can also be audible.

Yes exactly,
But this is the same situation with all forms of audio transport.
its the same even on the PCI-E bus.

 

[cishida]

 
Yes, but in digital audio timing is also important - is not quite as black and white as either getting a "proper" signal or not. You can be somewhere in between.
 
1. Variations in the timing of those 1s and 0s can be audible (jitter).
2. If an error occurs - say the signal for a set of 1s and 0s are unreadable -- there is no time to ask for re-transmission. The receiver just has to play whatever it gets and move on. These errors can also be audible.

 
The data transmitted over your USB cable or ethernet cable or via Wi-FI is not perfect - there are errors. It's just that the non-audio data transfer protocols automatically recognize errors (checksums etc) and retransmit when necessary. So for the non-audio data transfers buying a better cable may make transmission faster (by avoiding loss and errors) but won't make the end transfer "better" - the protocols ensure that the data will be reconstructed at the end.

 

[cishida]

 
Yes exactly,
But this is the same situation with all forms of audio transport.
its the same even on the PCI-E bus.

 
Pretty sure the PCI-E bus uses some sort of error detection / re-transmission. And I'd be surprised if there's a special PCI-E protocol for audio data.

 

[Letmebefrank]

This thread is completely off course and heading to BFE very fast. That said, I do find it interesting to hear both sides lol

 

[Jimster480]

   
Pretty sure the PCI-E bus uses some sort of error detection / re-transmission. And I'd be surprised if there's a special PCI-E protocol for audio data.

There isn't one, and yes PCI-E has some sort of error detection. But so does USB, and USB 2.0+ moves just as fast as PCI and PCI-E (especially PCI-E 1x)

 

[winders]

 
There isn't one, and yes PCI-E has some sort of error detection. But so does USB, and USB 2.0+ moves just as fast as PCI and PCI-E (especially PCI-E 1x)

There is no error detection/correction in USB audio......

 

[cishida]

 
There isn't one, and yes PCI-E has some sort of error detection. But so does USB, and USB 2.0+ moves just as fast as PCI and PCI-E (especially PCI-E 1x)

My understanding is that the digital audio protocols (i.e. the protocols used between source components and DACs), including usb audio, do not do re-transmission.

 

[Porteroso]

USB itself does not have error detection for an audio feed. I am not sure at which layer that is added (things like file transfers would have error correction), but for a dac gaining complete control over its USB input, it would not have error detection.

 

[Jimster480]

 
My understanding is that the digital audio protocols (i.e. the protocols used between source components and DACs), including usb audio, do not do re-transmission.

No they do not. But USB protocol has re-transmission period.