[Baldr]

So, Mike, dare I ask...the ultimate goal of all of these standards is simplify to supply data in a perfectly unmodified format to the DAC for conversion. Computers have been exchanging data in a bit perfect way for decades. There is no challenge to this. Yet, it seems to be a serious challenge for audio digital interfaces.

If USB 5.0 is "USB fixed", does that mean to say that the DAC is finally getting this bit perfect bitstream, as intended, and there is no room for improvement?

One would think that eventually data would be supplied as intended (as computers have done for decades) and we'd be done with the big circus caused by inherently inadequate digital audio interfaces..

One would think..
I have found that timing intervals below 10 ps can have significant impact on jitter. BWD enables simplest latching for proper timing, which changes from DAC to DAC and is one of my secret sauces.

 

[wingsounds13]

So, Mike, dare I ask...the ultimate goal of all of these standards is simplify to supply data in a perfectly unmodified format to the DAC for conversion. Computers have been exchanging data in a bit perfect way for decades. There is no challenge to this. Yet, it seems to be a serious challenge for audio digital interfaces.

If USB 5.0 is "USB fixed", does that mean to say that the DAC is finally getting this bit perfect bitstream, as intended, and there is no room for improvement?

One would think that eventually data would be supplied as intended (as computers have done for decades) and we'd be done with the big circus caused by inherently inadequate digital audio interfaces..

Bit perfectness is most certainly not the issue. The typical DAC might see a bit error rate around 1 bit error per day. A high error rate might be one per hour and a transmission system just shy of failing may see a few bit errors per minute. If we were getting enough bit errors to change the character of the audio reproduction, we would not be hearing subtle changes, but gross errors in the form of ticks, pops and just plain NOISE all but drowning out the music. Bit errors can and will hit any magnitude bit in the sample, high or low - they are not restricted to only low bits.

What we are getting from all of our interfaces is electrical noise that gets distributed all through the DAC and affects the output, and timing errors (jitter) that does have subtle but easily noticable if not identifiable as such effects on the sound quality. Different interfaces (optical, coax, AES, USB and I²S) all have different sensitivities to noise and timing issues and none is perfect or immune to these issues.

I don't doubt that the Schiit USB 5.0 interface is excellent, but it is still not perfect. Some received noise will still filter through. Noise generated within the interface itself still exists. Timing errors within the interface will still influence the sample timing of the DAC. There is ALWAYS room for improvement though with each step this room and the improvements get ever smaller. Thus is the life of the audiophile and the makers of audiophile equipment.

Yes, this is all old hat. Sorry if dredging up old stuff like this may be boring or frustrating, but the post that I am responding to seemed to me to merit discussion of the subject.

J.P.

 

[Pietro Cozzi Tinin]

Hey guys!! We share a common hobby, or is it love, or is it that to which we listen? The point is that this would be a boring thread if we all had the same rankings/feelings/opinions (pick one - or more). It would be like an audio show where only Schiit showed up. I occasionally am caught with filters off (well, OK, frequently) and may well have pissed people off. If so, I honestly regret it. All I have to do is look into the mirror to realize the folly of taking myself too seriously. Perhaps participants here could consider this as well, so we will not resemble those on the highest floors of audio shows who are two heartbeats away from a heart attack selling their car priced systems because they mistake their opinions about a hobby as religion. This is for fun, whether some of us transiently offend or not. Let's have fun (and learn - I do) for our own sake.

What can you say?
Love is a battle field.

 

[Pandahead]

One would think..
I have found that timing intervals below 10 ps can have significant impact on jitter. BWD enables simplest latching for proper timing, which changes from DAC to DAC and is one of my secret sauces.

Mike I have an honest, curious question about this and if it's formulation in my layman's brain is erroneous please tell me and throw me a crumb so I can understand a little better.

I gather that it's not like each pulse shoots up 100% vertical and rings a bell like a carnival game. The pulse must be angled slightly and that angle can vary with what is demanded of the power supply? And drift around a bit in relation to where in time it's recognized? Is my notion in trying to grasp this real and measurable, a part of the less than 10ps you mentioned? If so BWD simplifies things so you can manage this better?

If this notion is pure nonsense I will not be offended if you tell me, as I have been trying to grasp this for a while. Thanks.

 

[Ableza]

@Pandahead I'm not Mike, but here's a laymen's description for you. In general a digital signal change is instantaneous. The change in voltage happens at the speed of electricity, which is nearly the speed of light. A digital state change from zero to one and back CAN potentially have a slope if there are manufacturing issues with the device but in all practical applications it is instantaneous. The state change itself is not the problem (if there is one.) The problem can come from handling (encoding and decoding) the bit timing: accuracy in knowing how fast the state is changing in terms of bit rate and being able to count them all, not accuracy in terms of the actual voltage change from zero to one on a micro level.

 

[Clemmaster]

No, the rate of change is not the speed of light.
The load of the transmission line (inductance/capacitance), together with the output impedance of the power supply will slow down the transition and create a slope.

 

[rkw]

A digital state change from zero to one and back CAN potentially have a slope

No. The real world is analog. There is ALWAYS a slope, and you can see and measure the rise and fall time of a digital signal even on a cheap oscilloscope.

 

[Ableza]

Well, I am not talking about transmission lines or analog. I am talking about inside a device. Where digital lives.

 

[GearMe]

No, the rate of change is not the speed of light.
The load of the transmission line (inductance/capacitance), together with the output impedance of the power supply will slow down the transition and create a slope.

No. The real world is analog. There is ALWAYS a slope, and you can see and measure the rise and fall time of a digital signal even on a cheap oscilloscope.

So...are you folks saying that the slope on the state change in the quality of products that we are interested in (think Schiit -- not an outlier, low quality example to make a point) is the source of potential problems as well? Or is it, generally speaking, just the encoding/decoding as stated by Ableza? Or...both?

 

[NoahFence]

I'm having a hard time getting into classical. I turn up the quiet violin parts so I can hear all the little details and then bam the horns come in and blow my face off. I'm all for dynamic range but it's a bit too much in the classical that I've heard.

Classical music is too loud? OK, so it's dynamic range.
It's not just you. I listen almost exclusively to classical music and for orchestral music, especially on ear phones, I frequently have to turn up the volume on the soft passages. Then the loud passages come across much louder than I am comfortable listening to. (I am protective of my ears.) Fortunately with earphones I am close to the volume control!
So for myself alone I would prefer a slightly compressed dynamic range for orchestral music, but with digital recording I suspect the ability to do this without degrading the overall sound is limited. And others want to play those fff passages loud.... Try chamber music, I love that as much as full orchestra.

 

[Pandahead]

It's amazing in a way how it all works and yet we can hear when it's not so good. I realize that on the miniscule scale of the electronics that it should be essentially instantaneous, especially when talking about 1ps. It's kind of mind boggling to me all the minute factors of what Mike designs that produces better sound than the major dac chip makers. Sometimes I think USB set us back 25 years!

 

[NoahFence]

I'm with Pandahead in that I don't really understand the tech nical details about timing and jitter, etc.

The state change itself is not the problem (if there is one.) The problem can come from handling (encoding and decoding) the bit timing: accuracy in knowing how fast the state is changing in terms of bit rate and being able to count them all, not accuracy in terms of the actual voltage change from zero to one on a micro level.

If you could enlarge on that a bit, I personally would appreciate it. If I understand you correctly, the problem is not the bit/voltage per se, It is just when to start the voltage and how soon to turn it off.

The load of the transmission line (inductance/capacitance), together with the output impedance of the power supply will slow down the transition and create a slope.

Would this problem not be (more or less) identical with reproducing an analog signal? How is digital different in this regard?

Thanks for your patience in responding to those of us not well versed in these issues
NF

 

[rkw]

Well, I am not talking about transmission lines or analog. I am talking about inside a device. Where digital lives.

It doesn't matter, inside or outside a device. You can instrument a circuit board and see rise and fall slopes in the signals. "digital" is only a concept. The real world operates in analog. Digital is interpreting analog signals as representing digital data.

 

[Ableza]

It doesn't matter, inside or outside a device. You can instrument a circuit board and see rise and fall slopes in the signals. "digital" is only a concept. The real world operates in analog. Digital is interpreting analog signals as representing digital data.

Of course it all happens in the "analog" world, but I am addressing the original question: does any possible slope on the digital pulse have meaning in digital audio. No, it doesn't. The rise and fall of the voltage signal is effectively instantaneous and the slope (if there is one) of the pulse is meaningless in digital audio data transmission, especially when compared to other effects such as sync and timing.

 

[rkw]

Of course it all happens in the "analog" world, but I am addressing the original question: does any possible slope on the digital pulse have meaning in digital audio. No, it doesn't.

We had gone a bit off-topic. Getting back to the original question...

I have found that timing intervals below 10 ps can have significant impact on jitter. BWD enables simplest latching for proper timing, which changes from DAC to DAC and is one of my secret sauces.

I gather that it's not like each pulse shoots up 100% vertical and rings a bell like a carnival game. The pulse must be angled slightly and that angle can vary with what is demanded of the power supply? And drift around a bit in relation to where in time it's recognized?

This is about jitter. @Pandahead wondered whether jitter is caused by rise/fall slope, and fundamentally it is not. Jitter is caused by drifts in timing due to other factors. Here is an explanation: http://www.tnt-audio.com/clinica/jitter1_e.html