How can this influence SQ?

[manisandher]

A couple of months ago, I posted a particular finding of mine in the following thread: http://www.computeraudiophile.com/content/Highest-quality-AES-EBU.

 

I found that I could drastically change the sound coming out of my DAC simply by changing the AES voltage setting in my RME AES-32 PCI interface. (The AES-32 is being slaved to the DAC, so is NOT generating the all-important wordclock.)

 

The AES-32 has two settings for the peak-peak voltage of the AES signal: 5V/professional and 2V/consumer. The 5V sounds full and rich, if a little 'thick'. The 2V sounds crisp and clear, if a little sharp and edgy.

 

In the aforementioned thread, I speculated that:

 

1) the 5V setting could be increasing the SNR

2) my DAC might be reacting differently to the two AES voltages

3) the RME card might be struggling to provide a true 5V signal at high (MHz) frequencies and might be happier outputting a 2V signal

 

Wanting to explore 3) a little more, I've tried underclocking the CPU (21x down to 9x) and reducing the CPU voltage (0.925V down to 0.7V). Both of these have absolutely no affect whatsoever - the difference between the two AES voltages is still clearly audible.

 

Do you have any idea as to what could be responsible for such a difference in sound between the 5V and the 2V setting?

 

Mani.

 

[Mike in MD]

You could try making a simple voltage divider to see if the 5v signal brought down to 2v sounds like an unchanged 2v signal. That would tell you if the effect is in the DAC or up-stream in the computer.

 

You would have to be careful not the change the input impedance of the DAC as seen by the RME output.

 

 

[Shadorne]

Mani,

 

My guess is that the higher signal level is interfering with the master clock signal coming in...

 

[Synfreak]

Hi Mani!

 

Maybe there are some bits lost on the way with the 2 volt setting, forcing the DAC to hold the last sample, and this might lead to an more edgy sound?

 

Maybe you can try to loop the signal back to the RME card, and record it.

If you could perform a "Bit-by-Bit" test, it should show off if there are some errors in the stream.

 

Cheers

Harald

 

[JR_Audio]

This is normal and has several reasons (loss when recharging the isolation material with higher voltage but shorter transition time through the hysteresis at the input side). My finding is that depending on the AES/EBU cable, something between 2.5 and 3.5 Volt sounds “perfect”. Higher voltages tend to be too muddy in the treble and lower voltages tend to be too bright and nervous in the treble.

 

Juergen

 

 

[manisandher]

Thanks for all the replies and interesting thoughts.

 

"Higher voltages tend to be too muddy in the treble and lower voltages tend to be too bright and nervous in the treble."

 

Juergen, well this is exactly what I'm finding.

 

Although I know very little about HF transmission, I find it incredible that this is happening at all. Could you help me understand a little more what you mean by, "... loss when recharging the isolation material with higher voltage but shorter transition time through the hysteresis at the input side." please?

 

Is this a commonly accepted effect? Could cables, with different insulation material and dielectric constants etc then add/subtract from the effect.

 

I was thinking that perhaps something between 2 and 5V would be 'perfect', so I'm glad that someone else can corroborate this.

 

Building on the suggestion that 'Mike in MD' put forward, I will now look into putting together some sort of voltage divider to reduce 5V down to 3V or so...

 

... unless someone has another suggestion?

 

Cheers,

Mani.

 

[JR_Audio]

Digital Cable and AES physical Level: In a certain amount, you can compensate with different AES/EBU cable with different isolation and different copper / silver for different physical AES/EBU levels, but only limited.

 

BiPhase Encoding: The “problem” with AES/EBU and SPDIF is the Bi-Phase encoding, that logical 1 is twice the frequency as logical 0. In digital circuits 2 times the frequency does have a different transition time as 1 times the signal and also with limited slew rate of digital circuits you have lower bandwidth at higher levels compared to lower levels with higher bandwidth.

 

These points are, I guess too complex, for regular CA readers and also would be far too long to explain all the points.

 

I you plan to change the voltage divider be sure to take always into account the line resistance of digital audio, but you can also tweak a little bit in that way that a 115 Ohm line with sound a bit brighter than a standard 110 Ohm line and a 105 Ohm will sound more smoother.

 

Juergen

 

[manisandher]

"These points are, I guess too complex, for regular CA readers and also would be far too long to explain all the points."

 

Yes, I understand this. Where do you then suggest I look to be able to learn more about this stuff?

 

I find all of this really facinating. I would suggest that anyone who is using AES3 to pipe data from a transport (be it a disc player or a music server) to a DAC should be interested in this also - if the AES voltage is too high or too low (but still well within the AES3 spec), you may not be getting optimal results. This is especially important if you use a 'legacy' DAC (like mine) that doesn't have a firewire or async USB input.

 

Why didn't BNC just become the adopted standard for piping digital data? The cynical side of me suggests that this was simply so that studios could use the same type of cable (110 Ohm with XLRs) for both digital and analog transmission. Sigh.

 

Mani.

 

[manisandher]

"The 'problem' with AES/EBU and SPDIF is the Bi-Phase encoding, that logical 1 is twice the frequency as logical 0. In digital circuits 2 times the frequency does have a different transition time as 1 times the signal and also with limited slew rate of digital circuits you have lower bandwidth at higher levels compared to lower levels with higher bandwidth."

 

I forgot to mention earlier that I'm using dual-wire mode for all of my material. 1fs material is upsampled to 4fs using XXHE's zero pre- and post-ringing 'Arc Prediction' scheme. I have the RME set to dual-wire for all native 2fs and 4fs material.

 

I'm not sure that the bi-phase mark encoding is therefore an issue.

 

Mani.

 

[JR_Audio]

My experience was that SPDIF on BNC doesn’t have a deep enough bass. But there is also a not widely known and not widely used AES Coax digital interface, which uses the balanced transfer technique of the AES/EBU Interface (I would highly recommend using mainly balanced digital audio interfaces), but with lower signal level (1 Volt loaded) but uses a 75 Ohm line with BNC connectors.

 

This is, in my opinion and experience, superior to the conventional SPDIF and also to SPDIF on BNC.

 

Juergen

 

 

[ar-t]

The only way to make SPDIF perform more poorly is to use the AES/EBU version.

 

Using a "pseudo-balanced" coax is a different matter. A lot of "high-performance " SPDIF set-ups use a technique similar to this.

 

All of this is moot, if the impedance on both TX and RX ends are not carefully controlled. Which, in 99.99% of the cases, it is not. That is why you get so many ideas on which cable, etc., works best.

 

[ar-t]

"Why didn't BNC just become the adopted standard for piping digital data? The cynical side of me suggests that this was simply so that studios could use the same type of cable (110 Ohm with XLRs) for both digital and analog transmission."

 

Not cynical, at all. That is the thinking of the people who came up with that. On top of that, the original version was much worse!

 

(I tried to 'splain to the head of that AES Working Group how wrong they were. A brilliant PhD, but no clue of how RF worked. Typical of audio: "It is still just audio, right? Just in digital format?" No, it is RF, and audio techniques do not work at RF. I could go on, but why?)

 

[JR_Audio]

Everyone has its own experience concerning different issues, for example on digital interfaces. You, ar-t, have yours, and you are explaining this on your website, but I have mine.

 

I do not want to open an endless discussion, and I do not want to abuse anyone, who has other opinions, I just want to add one comment and this is about my personal experiences.

 

Over the last 15 years I have developed lots of different digital interfaces with AES/EBU, with AES Coax, with SPDIF on RCA, with SPDIF on BNC, with Glassfiber (and also Toslink).

 

With this different interfaces for different products, with different parts, different logic families etc, I have taken my own listening experiences, that has not to be taken as “law”, they are just my impression. And when I read the original post, it matches my personal experiences, so I chime in.

 

So my answers, (clearly in a very simple and short explanation) does refer to the original post in order helping, because of the similar impressions on the digital interface.

 

I know, you at ar-t have a different meaning, so why not. I hope you do understand, what I was trying to say.

 

Juergen

 

 

[ar-t]

After I wrote that, I was afraid you thought all of that was pointed in your direction.

 

It wasn't.

 

The whole SPDIF thing is a kludge, originally designed just to be a test spigot. AES/EBU was a step in the wrong direction, and I believe they realized their most egregious mistakes, and corrected them. Still, trying to think of SPDIF as one would audio, just in a digital format, is the root of the problem. Folks who come at this from, say, a telecom background, understand the basic approach needs a different mindset.

 

What this really comes down to is what the average CA reader can do. Well, apart from redesigning your gear, swapping cables is about all you can do. Look at the bright side: it is almost certain to yield different results. And different, especially when you have paid for it, has to be better!

 

Right?

 

RIGHT!

 

[manisandher]

... I've learned a shed load.

 

Cheers,

Mani.

 

[ar-t]

Hmmm.........if you rearranged that, and misspelled a few words..............no, don't go there.........!

 

[manisandher]

What, 'shed load' isn't a common term in America? No, it has nothing to do with another similar-sounding term ;-)

 

I'm happy to spend $500-$600 to play around with things and see if I can get a more consistent sound from my DAC. Short of changing the input stage of the DAC, what should I do?

 

Let me know if you have any ideas.

 

Cheers,

Mani.

 

[manisandher]

The manual for my ADC/DAC strongly recommends the use of the 'ground isolators' that are provided with it. These are described as follows:

 

"The AES Signal Ground Isolators consist of a number of turns of precision 110 Ohm balanced, shielded cable fed through multiple ferrite cores, each having a different selected permeability. The cable is terminated with XLR-3 connectors. This construction provides 160 mH of inductive filtering effective over a very wide range of frequencies. A 100 Ohm resistor is also placed in series with the pin 1 ground connection, providing further isolation.

 

The AES Signal Ground Isolators should be connected to all AES digital signal inputs and outputs at the Model Two whenever they are in use. The AES Isolators are designed to be used in series with standard AES cables that are connected to digital recorders, editing workstations, etc."

 

From a theoretical point of view, how would using such isolators affect the HF signals flowing through the AES cables?

 

Again, I'm just interested in learning as much as I can...

 

Mani.

 

[JR_Audio]

I guess you better ask “Pacific Microsonics” how they made their AES input.

 

With Pin 2 and 3 it is normally very clear and that they are just connected to the pins of the digital input transformer, but for Pin 1 there are some different possibilities out there. They differ from company to company, but normally Pin 1 should also have no galvanic connection to the DAC, because it should only be shield the twisted 110 Ohm line, but some have connected Pin 1 of the receiver (DAC) to ground, which is wrong. Some have only a HF capacitor from Pin 1 to ground, which is in most cases a good compromise, between shielding and isolation.

 

Be careful that you do have a cable, where the case of the xlr connector is not connected to Pin 1, otherwise you will have a ground loop. Normally this should not, but I have seen some, that have made this mistake.

 

The best input schematic would be to have a digital input transformer with a static shield between the primary and secondary winding, and that this shield is only connected to pin 1 of the xlr jack. This is the way I have done a lot of AES/EBU inputs. So the static shield between the input transformer is connected only to the shield of the cable and on the sender side of the cable, connected to ground, By doing this, you have a very high common mode rejection of any common mode voltage and so the best isolation between the sender (transport) and the receiver (DAC). Only the differential signal of the AES signal is transferred to the DAC.

 

This is the reason why I do like balanced interfaces for digital line so much. With SPDIF you can't have this “features”. Good luck.

 

Juergen

 

 

[ar-t]

"The best input schematic would be to have a digital input transformer with a static shield between the primary and secondary winding,"

 

WRONG!

 

You guys really need to stop reading the Scientific Conversions App Notes. Do I need to introduce you to my TDR? Sounds like I might.

 

Having caused enough mayhem for one month (and it has just started), I'll refrain from saying what I think of that cable idea. They probably also need introduction to my TDR.

 

OK, in theory, it turns the HF common-mode crud into heat, and poof!....away it goes. So much for theory.

 

[JR_Audio]

Maybe we should meet for a cup of coffee on the next CES in Las Vegas, or may be on the next Rocky Mountain Audio Fest (but I am not sure, whether I am there).

 

This should be much more easier and also more fun, to share different experiences when meeting each other face to face, instead of writing.

 

I will stop here, so lets look forward.

 

Juergen

 

[The Computer Audiophile]

Ar-t - Your tone is antithetical to a good discussion where people share information freely. Computer Audiophile is a laid back site where this kind of banter does fly. I'm willing to bet your approach is not the best and only good approach in the world.

 

If your not familiar with some of the people posting here I highly recommend sending them a private message to better understand where they are coming from.

 

[ar-t]

OK. But my "tone" is that I am tired of seeing the propaganda put forward by Scientific Conversions, as gospel. Perhaps you are not aware of my work on Audio Circle, where I have dispelled their propaganda.

 

I do not proclaim my approach as best, whereas SC does. They have managed to bamboozle an entire industry they have the only answer. The fact is, as you have pointed out, there are others.

 

Sorry to upset the cart, but after 20 years, seeing it one more time is one more time I can do without.

 

But, I will contact JR. Can't divulge all our trade secrets to him, but I can steer him in another direction.

 

[The Computer Audiophile]

Thanks for the response ar-t. I'm with you 100% when it comes to dispelling propaganda. Only good can come from that.

 

[kana813]

Here's one of Pat's audiocircle threads about digital connections:

 

http://www.audiocircle.com/index.php?topic=45330.0