Are the Lights On or Off?
It has become quite common in high-end audiophile circles to go all “analog” on parts of the signal path that are carrying digital information. Talking about noise or electrical interference in a USB cable as it carries electrical signals from the output of your computer to the input of your DAC, ignores the basic tenants of digital representations of music or images or video or anything else. I’m not sure how this “analog” thinking crept into the land of computer audio but it seems to have taken a strong hold in some circles.
There some very experienced and knowledgeable electrical engineers out there explaining how electrical signals pollute digital streams or packets of digital media and others with equal experience and knowledge that dismiss the claimed issues of noise in the outputs of their DACs. Who’s right? You have to go back to the basics of digital coding to begin to grasp both sides of this issue. And while I used to build Heathkits, know how to read a schematic, and spent 5 years making a living as an electronics repairman, my knowledge is insufficient to parse through all of the finer points of the processes involved…but as an audio engineer and musician, I can share my experience and knowledge. I teach digital audio to a new group of audio engineering students every fall. I do know some things.
There may be an infinite number of voltage levels tripping through a circuit board, USB cable, Ethernet cable, or DAC…the digital information doesn’t care. The process of converting or representing an analog electrical signal in simple binary…ones and zeros…bits only needs to know if the lights are on or off? That’s the beauty of digital. John Swenson, an engineer interviewed over at Audiostream.com stated, “… the wire can pick up huge amounts of noise and still be ‘read’ properly. This is the promise of digital and why it is used in the first place.” Let’s be clear about what he’s saying. The room in which you’re processing two states…binary data or lights on or light off…can be filled with smoke, have a large number of bouncing ping pong balls obscuring the space, or have dim lights or bright lights…all the digital receiver (in this case you!) has to be able to perceive is whether the lights are on or are they off.
I’m going to ignore the issue of sampling rate and jitter because a good quality DAC takes care of reclocking…it’s only the data that matters.
I’ve repeated my mantra on the “bits are bits” debate. If a successful transmission of digital data can be accomplished by a reasonable quality USB or Ethernet cable or by a WiFi connection, the medium doesn’t have any chance of impacting the fidelity of the reproduced sound. Of course, this will only be the case if the reconstruction of the waveform is done by the same DAC. Naturally, different DACs from different designers will have their own sonic signatures.
Why do I get so righteous about this rather simple concept? Because I’ve done repeated tests on cables and discs…expensive or treated or not…and found that the digital bits from one path vs. a challenging path when compared using polarity reversal results in a complete null signal. If they perfectly null each other then they were identical. And even the most ardent “bits are NOT bits” person has a very hard time explaining how things can sound different between two sets of data that are identical.
That doesn’t mean they won’t keep trying.
37 thoughts on “Are the Lights On or Off?”
I think a more precise and simple explanation for some of this discussion should be explained in terms of the powers supplies. Almost all appliances run on direct current and have some type of alternating current conversion process. This is a critical area that most people can understand. Digital transmission is another discipline all together, and only in the last 40 years has this science been examined. Like all engineering, there are plenty of opinions, and in some cases, patents to work around when designing and selling new idea’s.
Mark, Mark, Mark. Cherry picking quotes is so…intentionally misleading. Here’s the actual context of John Swenson’s words:
“What digital does is quantize those values and say ‘anything below a certain value (the threshold) is low and anything above the threshold is high’. This is the fiction part. In reality it can still be any of those values, there is nothing about the wire that physically forces it to be one value or another. A common spec today is to say low is 0V and high is 3.3V with a theoretical threshold of around 1.6V (half the low and high).
Why do this? The primary reason is you get a large ‘noise margin’. What happens if you have a wire that is supposed to be low, but has picked up some noise (ignore for the moment how it got the noise), as long as it is less than the threshold it is still thought of as ‘low’. Hence the wire can pick up huge amounts of noise and still be ‘read’ properly. This is the promise of digital and why it is used in the first place.
What is the down side? You started out with one wire that could have an almost infinite range of values and replaced it with a wire with only two values, if you want more values, you need more wires. But once you have more wires, you can put those wires in a situation where they pick up lots of noise and you haven’t changed the value at all.
As explained above the magic is all in the receiver, it performs the thresholding process. But there is more to it. The transmitters are part of this. They are supposed to output only one of two values (high and low, 0 and 3.3 for one standard). Thus most of the time the voltages on the wire actually do closely represent the theoretical standard.
Now on to the fun part, what actually happens? First off you cannot have just one wire, a voltage is a DIFFERENCE between two wires, so there HAS to be a second wire. In digital systems this is typically what is called a “ground plane”, a piece of copper on the PC board that covers the whole thing. This is the “reference”, all the voltages on the other wires are compared to it. Many people treat the ground plane as if it is equipotential, the same voltage everywhere. But it isn’t. This is a major issue in all the subsequent discussions so it is important to understand. There is resistance and inductance in the ground plane, not very much, but it is still there. Any current flowing through the plane will cause a voltage to develop between different parts of the plane. This is called “ground plane noise”.
Of course if read even more of what John wrote, you’ll understand this issue even more.
Repeating ‘bits are bits’ will only get you so far in understanding the relevant issues.
OK fine. I simply quoted the fact that John agrees that the concept of digital encoding is because any noise isn’t relevant to the communication of the needed two state information. All of the explanations about low and high and possible mistakes and noise don’t matter if the data that I get out of the device is identical…and I mean bit for bit identical…to the data out of a straight digital cable. If they are the same…and I’ve shown it to be true at least for a high-priced USB cable…then I don’t need the device.
The issue with ‘bits are bits’ is we are not talking about a digital system. We are talking about a mixed signal system. The devices that generated this discussion sit between the source, typically a computer, and a Digital to Analog Converter. Focusing on the data, the cable, or the REGEN/Wyrd/JitterBug, misses the most important fact – in the end we are dealing with an analog signal.
I think it would be beneficial to step away from these devices and their claims and simply look at the issue of noise in mixed signal systems. Doing so will demonstrate how noise at the input of a DAC, for example, can corrupt the analog signal. Here’s a relevant quote/intro:
Noise In Mixed Signal Systems
It is also common, however, as stated earlier, for electronic systems to begin and end as analog circuits, but in between have digital logic subsections. Such a system is called a mixed signal system, and noise is a concern both at the input to the system and at the output. Noise at the input is converted into errors in the digital logic circuit; the data then picks up further errors as it is processed within the digital logic circuits, and when the conversion from the digital to the analog domain is finally made, these accumulated errors become noise once again at the digital logic circuit’s output. There are some mixed signal system that combine digital and analog circuits on a single semiconductor strip substrate, and these can produce crosstalk, which originates within the digital logic circuits and creates noise within the analog circuit section. The most common of these are power supply noise and ground bounce”
That’s from “The Electronics Handbook, Second Edition” edited by Jerry C. Whitaker. Here’s a bit more:
But digital-to-analog converters used in this process are subject to all of the noise problems of analog circuits since they themselves are analog devices, and rely on resistive components, which have thermal noise limitations; thus, the digital information necessarily enters the digital system with noise present.
So “this ‘analog’ thinking crept into the land of computer audio” because we are dealing with a mixed signal system whose output is analog.
Sorry for the delay in getting this comment approved, I’ve been at the California Audio Show all day. And thanks for the additional information, which adds the “noise” aspect of analog signals into the world of digital coding. Yes, there is noise…but the most important question is whether that noise corrupts any of the digital information. And in quality equipment connected with quality wiring or even wireless systems, the answer is no. The noise is below the threshold of changing any information and even if a bit did change the error detection and correction routines would return the errant bits to their correct values.
My point is that digital representations of information…the simple maintenance of an accurate stream of bits is all that matters. If the bits out of my non-regenerated system are the same as the ones from a REGEN box, then would you acknowledge that the playback DAC has to produce identical fidelity? If I compare the two files by flipping the phase on one and not the other and they cancel to null, does that show that the two files are identical? If I take the two files and compare them byte by byte and they prove to be identical, then is the REGEN box or fancy cable supplying anything that can improve the sound? Your explanations and quotations don’t sway me…I’ve done the comparisons on other tweaks…and willing to do it with the current equipment under discussion. But I won’t purchase the boxes.
You’ve missed the most important point entirely. And this is not “my point”, it is basic electronics in mixed signal systems as the first quote in my last comment is from The Electronics Handbook.
Yes, there is noise…but the most important question is whether that noise corrupts any of the digital information.
Wrong. The most important question is whether noise affects the analog signal. Answer = yes.
So here’s a fundamental question. If the data coming out of the REGEN is the same as the data coming down a line without the REGEN would the digital information be identical or not?
It’s interesting that the AES Technical committee that I’m a member of has opened a discussion on the REGEN thanks to Bob Katz. The subject of his email is “REGEN Nonsense”.
“While we’re at it I’d like to change the subject a little and mention a new religion that insults my technical intelligence. It’s a bunch of guys at computer audiophile who are swearing by a USB regenerator that “re locks” the USB. I haven’t gotten a straight answer about how this could make any difference whatsoever in an asynchronous USB system but the amount of fighting and ignorance and insulting my good sense is blowing my mind.
As most of you who understand how digital audio works, when there is a high quality clock directly driving the dac and software commands data from the computer when needed this means that there is no clock per Se in use on the way into the USB. The USB simply carries data and a simple handshake from the smart circuit in the dac commands that data whenever it needs to refill its little buffer.
But a whole religion has built up around a device called USB regen. Which has — would you believe, some kind of reclocker. It’s a solution looking for a problem and they have bamboozled a bunch of audiophiles into believing they hear a difference.
Hold their feet to the fire Mark. We depend on you to help maintain some sort of semblance of integrity to the High End audio market.
Perhaps this has been said before however I did not find it clearly described in your post today.
You are correct regarding the binary state of digital data. I really liked your image of the smokey ping pong’d room and the light on or off. The light is on or it is off and if you are not sure then it is off.
The “analog” voltages that represent a binary state is very broad, meaning is it designed to be unambiguous. The key part that is not obvious in your post is the various means of validating the data. The disambiguation of state by the range of voltage that represents a 0 and the range of voltage that represents a 1 is the first validation.
Data is transmitted from source to destination using well defined protocols that group data bits such that they too can be verified. The data is bundled such that an error in the data can be detected and corrected without getting it sent a second time.
Think of data bits being wrapped in a package then sent to somewhere that in turns wraps it onto another packages, on and on and each has its own validation and correction mechanisms.
Some protocols (TCP) go even further and declare that the data will go through correctly come hell or high water (within reason) and will force the data to be resent if the data package cannot be validated.
The digital bits will arrive intact or they will not arrive at all. There is no such thing as degrading the integrity of a digital bit.
There can be errors transmitting and receiving digital data and they will be corrected or not delivered at all.
Getting the digital bits from digital point a to digital point b is not subject to most environment variables like analog data as long as the voltage or light pulse can be delivered. Extreme environments will probably need some way to shield the medium to prevent uncorrectable errors. The keyword here is extreme
Each physical layer has well defined parameters including the maximum length of the physical medium allowed for reliable transmission and the integrity of the interconnections. Most USB, Thunderbolt, Ethernet, Coax, fiber will do just fine and the bad one will cause errors that will have unambiguous effects, nothing that requires a discriminating ear (or eye) to hear or see.
A stable, reasonable physical medium that meets the requirements of the defined standards is all that is needed.
Please remember that Mark and I are talking about getting the digital data from point a to point b. We are not talking about transforming that digital data once it arrives at the other end of the physical connection.
Digital bits that are transmitted by the New Horizons probe out beyond Pluto traverse a hostile physical environment and those bits arrive and are used to produce high resolution images (and all that other stuff). Getting your bits intact to and from your components within a few feet of each other is not that challenging unless you’re next to an MRI scanner or 100K or higher radio or TV transmitter.
Buy the pretty and shinny status cables if that meets your aesthetic or ego requirements AND they meet the technical requirements specified by the standards committees. Buy reasonable but inexpensive cables that meet the technical requirements specified by the standards committees and put your $$$ into the ADC’s and DAC’s and other equipment where quality really does make a big difference.
Thanks Wayne…I love the reference to the New Horizons probe.
Firstly, I’m neither an audiophile nor have golden ears (though I’d like to believe that I missed the Golden level in the Phillips Golden Ears test because of the equalisation section). I just like listening to music. I enjoy reading your daily blog – every morning right around 6.30 am here in India ! You have helped tremendously in building my knowledge around high resolution audio.
Here’s my observation – my laptop’s motherboard shorted and had to be replaced. It is an HP laptop (i5, 4 GB RAM). Before that I always had an ‘unclear’ audio when using USB DACs like Oppo HA2 and HRT’s Microstreamer. I certainly wasn’t expecting any improvements when the motherboard got replaced by an HP original of the same specifications. I was so surprised when I noticed that the sound was clearer, esp the bass frequencies. I’m not an EE and have no explanation, but I do know that the sound is clearer now. There has been no other change in my system – same DACs, headphones/speakers, cables,etc. If these DACs fit your definition of “poorly engineered”, I guess I should upgrade my DAC first! It would help if you can elaborate on this definition though.
It appears to me that there are differences in hardware which impact sound quality. In my case, it was the laptop motherboard which had a poor quality part which degraded sound quality.
So, I suggest that you and Michael sit down in one room, listen to some music on the same system and then come to a conclusion whether “bits are bits or there are other variables too”. Your readers don’t have the time or the money to keep experimenting with different gadgets and hence, we are looking for expert opinions to help us in our decisions when we want to build a good audio system. If the experts themselves keep arguing on what works and doesn’t work, where will the average consumer go for help?
Thanks for your experience. The DACs that you have are certainly capable of delivering very good sound. It’s pretty clear that your computer hardware was messing with the digital data prior to it being sent through the USB connection. That’s certainly possible.
Let me try a different approach since we’re not getting anywhere with the fact that we’re dealing with a DAC, not data.
Let’s instead imagine I just reviewed a pizza and I liked it better than the last pizza I reviewed. You read my reviews and discover both pizzas use the same mozzarella and conclude my reviews are nonsense because those pizzas must taste the same since they use the same cheese.
I point out that there’s more to a pizza than cheese to which you respond, “cheese is cheese”.
I’m going to post about this on AudioStream next week so I’m done here.
I’m heading to the show today. But quickly, your pizza strategy doesn’t work…cheese is analog. A beaded necklace of black and white beads would be closer to the digital issue I’ve talking about. If the necklace gets scratched or dirty the ability of a person to still see the sequence of B&W beads is not diminished. This is digital information…maybe it’s a mixed signal system because of the dirt…but the critical issue is whether the black and white beads maintain their integrity through the noise.
You say I’m dealing with a DAC not data…quite the contrary, all I’ve been hammering on is the validity of the data. The DAC…any DAC…is going to interpret the stream of data as they will. If the same stream of data comes along, they will produce the same output.
You’ve missed the forest for the cheese.
I am fairly new to this site but, just like Gurpreet, have begun to live receiving your morning newsletter.
Also, like Gurpreet, I love listening to music but have no electronic or sound engineering knowledge. I am not in to audio tweaking, in part for lack of funds (boy can people spend a lot of money on tweaks!), and because of the lack of understanding of what the tweaks are supposed to do.
However, if Gurpreet’s new motherboard improved his sound, does that mean the old motherboard was sending the wrong bits? This may be way oversimplifying things, but I am trying to learn.
I know how carried away audiophiles can get – in the 80s I remember reading an article in a UK audiophile magazine expounding the virtues of placing wet cloths over every metal surface, including over the taps (faucets) in the kitchen next door!
Yes, it means that the previous motherboard was not delivering the data accurately.
“I’m going to ignore the issue of sampling rate and jitter because a good quality DAC takes care of reclocking.”
What if noise affects the DAC’s clocking?
Ever since Meitner and Gendron’s 1991 AES paper we’ve known that ground plane noise changes the reference voltage to which the signal is compared to determine whether the “lights are on or off.” Unless things are really bad the noise won’t cause a misread of a “1” as a “0” or vice versa, but what it does do is create jitter. (For a graphic depiction of the mechanism, see figure 1 here, which is taken from the M&G paper: http://www.amr-audio.co.uk/large_image/MAC%20OSX%20audio%20players%20&%20Integer%20Mode.pdf .)
So the basis for things like the Regen isn’t voodoo, it’s just paying attention to things the great engineers “present at the creation” of digital audio were writing about decades ago.
You and plenty of others may want to believe that devices like the REGEN and others make difference in the delivery of digital audio through a DAC…I don’t. I listened carefully to the REGEN box and found not audible difference. Bob Katz, a well-know author and mastering engineer, wrote to me today and told me he thinks I’m being too nice about the issue. He said I could quote him, “let them come back with measurements and then we can talk.”
If it works for you…go for it. I don’t recommend it.
There appears to be cognitive dissonance on display here evidenced by:
– if Gurpreet’s HP motherboard was not sending accurate data (the wrong bits) would this not have an effect globally on it’s operation? Would it not have affected other digital devices connected to this USB port?
– In your previous listening test of the Regen two out of three people concluded it made an audible difference – were they mistaken or was the digital data changed in traversing the Regen? This could be easily tested by attaching a USB printer
The REGEN didn’t change the bits…this was acknowledged by the owner of the company. That two other audiophiles heard a very subtle difference is something I can’t explain. I don’t believe there was a difference. I listened to the same output that they did…and I trust my ears.
So, you have now two datapoints which indicate something other than the bits being responsible for audible differences:
– Gurpreets’s HP motherboard
How do you explain your position that in digital audio only issues which affect the data accuracy will be audible?
Your comment about Gurpreet’s m/b “the previous motherboard was not delivering the data accurately” doesn’t hold water as he reported no other issues that affected the data accuracy when using his laptop, only an audible difference from his DAC when the m/b ws changed.
I’m afraid your logic doesn’t hold in the face of evidence such as this – no measurements needed to show flawed logic.
But if you look at John Westlake’s measurements of the Regen you will see that it removes a large amount of low frequency fluctuation riding on the HF USB signal. Engineers can quiet easily envisage how this LF noise can pollute the ground of the USB receiving device showing the measurements for the effect of this noise at each step through a USB audio device & it’s final affect on the analogue output will be tricky & require expensive equipment.
The engineers and scientists that I respect and my own personal experience rule the day for me. The will always be believers but I’m not one of them.
“The engineers and scientists that I respect and my own personal experience rule the day for me. The will always be believers but I’m not one of them.”
Can you not see that you are doing exactly that – believing in those engineers who favour your view as opposed to those engineers & scientists who don’t? Can you not see that you are believing “your personal experience” (whatever that really means) despite the facts & logic that are in front of you?
Sorry, but again you demonstrate that you are very much a “believer” & not logical or scientific just based on what you have posted here.
John, you can disagree and push back…no problem. The science and engineering of digital systems is an area that I know very well. Having experienced many of the audiophile tweaks, I’ll continue to steer clear of devices that make claims that seem too good (or too subtle) to be true. Other, including you it seems, can invest in whatever turns you on.
” I don’t believe there was a difference. I listened to the same output that they did…and I trust my ears.”
My understanding was that this listening was done blind so are blind testing results now a matter of belief? I thought it’s use was to remove any such “belief” i.e any bias? Two people report an audible difference in a blind test that you organised & set-up but you didn’t hear any difference. From this result you come to the conclusion: “I don’t believe there was a difference – and I trust my ears.”
A person reports their DAC improves when their laptop’s motherboard is changed. Your only explanation is that the laptop was sending inaccurate data to the laptop. It’s one possibility but one based on what evidence? Wrong digital data sent to a DAC would not just result in degraded sound – it would result in noticeably audible glitches.
I really can’t understand your stated position in all of this – your statements aren’t logical – please help me understand.
Mark, it’s not about our individual freedom to buy or not buy whatever audio devices we like – that’s a given, within reason, in a democracy.
I’m not disagreeing – I’m simply pointing out how your own logic disagrees with the facts.
What I had hoped you might have been interested in was the science & engineering involved in the mixed signal nature of digital audio – something that has already been mentioned here by others.
John, I read the papers and researched the issues associated with “mixed signal systems” and I’ve chatted about it with several equipment and software designers…as well as other audio engineers. And you know I sat in my studio listening to my recordings through a box that is supposed to enhance my listening experience without changing the bits through one of the world’s finest DACs (my Benchmark DAC2 HGC which has ultra low jitter that is completely independent of the incoming clock on the USB line). I didn’t experience any change…and I really didn’t expect to based on my knowledge of digital systems. The “mixed signal” issue is a red herring…at least in this context. If there’s a contradiction in this explanation for you, then let’s just agree to disagree.
Mark, I’m not denying your personal experience – just pointing out the flaws in your logic & conclusions.
BTW, there is no incoming audio clock on the USB line if the DAC is working in asynchronous USB mode – the clock in the DAC acts as the master audio clock & the Regen has nothing to do with this clock.
My point exactly, the data is the same and the clock is new at the DAC. The sound that is output will be identical with or without the Regen…it’s only logical.
I’m sorry, Mark but your reply demonstrates that you definitely haven’t “researched the issues associated with “mixed signal systems”
So let me get this straight…if I don’t end up agreeing with you, then I don’t know what I’m talking about. Have a nice weekend.
Mark,, it’s not a matter of agreeing with me – I’m only reiterating some of the simple science & engineering behind digital audio.
I know the simple science and engineering behind digital audio and I have many very smart associates and friends that share my doubts on the Regen and other audiophile accessories. I’m content with the conclusions of our side of the debate.
OK but your conclusions are just beliefs & can’t explain the positive blind tests results of your two friends or explain the change of Gurpreet’s motherboard affecting the sound of his USB DAC.
Both of these results you have failed to explain with any logic – instead talking about your research in mixed signal & your many friends that agree with your beliefs.
Beliefs & logic do not make good bedfellows, I’m afraid
Have a good weekend
There’s only two possible points of view here.
A. All dacs are impervious to anything else coming Down the wire along with the signal.
B. They aren’t.
If you believe that A is true then a good quality power supply doesn’t matter, and a quiet gnd doesn’t matter, and emi and rfi don’t matter.
Or if you believe B. Then you accept some dacs can be negatively affected by the quality of the components they are connected to. And ‘bits are bits’ and believing B need not be mutually exclusive.
I believe a good dac should be impervious to source and environmental influence, I owned a couple that were, I currently live with one that obviously is not.
Just because a manufacturer comes up with a bullshit explanation for why a device works doesn’t exclude it from actually having an effect by other means
There’s more to it than that…the level of the design and build plays a significant role as well.