Recommended Posts

38 minutes ago, kelly200269 said:

Exakt systems should be ‘deadly’ quiet, with everything kept in the digital domain until the transducers.

In practice, noise is an issue even in the "digital" circuitry. All electrical circuits are, in effect, analogue (and therefore generate and carry noise), even when they transmit and process digital data. The data is not affected by the noise, but keeping it out of the analogue domain is a major challenge.

I think that your assumption that Exakt systems are inherently quieter than passive systems is not something that the Linn designers would go along with. I believe that that they would say that the noise has to be designed out of both kinds of system, and that the job is not necessarily easier for the one system than the other. Of course, the customer forking out £30K for a pair of Akubariks will expect them to be inherently quiet. But that doesn't happen just because they are Exakt devices.

David

Share this post


Link to post
Share on other sites
37 minutes ago, DavidHB said:

In practice, noise is an issue even in the "digital" circuitry. All electrical circuits are, in effect, analogue (and therefore generate and carry noise), even when they transmit and process digital data. The data is not affected by the noise, but keeping it out of the analogue domain is a major challenge.

I think that your assumption that Exakt systems are inherently quieter than passive systems is not something that the Linn designers would go along with. I believe that that they would say that the noise has to be designed out of both kinds of system, and that the job is not necessarily easier for the one system than the other. Of course, the customer forking out £30K for a pair of Akubariks will expect them to be inherently quiet. But that doesn't happen just because they are Exakt devices.

David

Whilst I acknowledge, and even recognise the concept of ‘noise’ in the digital domain, it’s impact is minuscule compared to the impact of noise in the analogue domain.

Yes, I personally find some USB and even Ethernet cables appear ‘quieter’ than others. But these differences are very, very, very small compared to ‘noise’ differences in analogue circuitry, and in the analogue domain.

With an Exakt system, it should be effectively ‘noise free’ until it reaches the DAC’s/output stages of the speakers. And I’m sure it is. Well. It should be if you’re ‘forking out’ £30k for a pair of Exakt Akubarik’s. Any ‘noise’ experienced in this type of system will be due primarily to the analogue output stages powering the transducers.

So yes, whilst I acknowledge that digital ‘noise’ does exist, it is an order of magnitude less than noise in the analogue domain.

Edited by kelly200269
  • Upvote 1

Share this post


Link to post
Share on other sites

My Katalyst Akubariks are practically silent with the volume turned up to 100. I have to put my ear right up to the tweeter/super tweeter to hear a little bit of white noise.

Share this post


Link to post
Share on other sites
3 hours ago, kelly200269 said:

Whilst I acknowledge, and even recognise the concept of ‘noise’ in the digital domain, it’s impact is minuscule compared to the impact of noise in the analogue domain

Let's be clear. Noise does not exist in the digital domain, because noise is a physical manifestation of imperfect signal transmission, and 'digital' has no physical existence, but is simply a construct of the human mind. We are accustomed to say that computers and other digital devices deal only with 0s and 1s, but it would actually be more correct to say that they deal only with instances of 'not voltage' and 'voltage' that are used to represent 0s and 1s respectively . The devices that transmit and process these representations of binary numbers do not 'know' that they are numbers. They're just voltages in electronic circuits, like any other voltages in any electronic circuit.

Noise is only attached to a signal in the analogue domain. But it can, and often does, come from associated circuits that deal with the voltages that represent digital data. The noise is always quite separate from the digital data (not least because it is, in effect, analogue), but keeping it out of the circuits that deal with, in particular, the analogue outputs of, say, a DSM or Exaktbox, is a significant engineering challenge. The lack of identifiable noise in akamatsu's Exakt Akubariks is an indication that their noise performance is, as one would expect, exceptional. But it has absolutely nothing to do with the fact that noise is non-existent in the digital domain.

For the avoidance of doubt, it is of course the case that the noise content of any analogue signal that is converted to digital will remain part of that signal and become part of the noise content of the eventual (analogue) output. But, when in the digital domain, that pre-existing noise becomes to all intents and purposes an integral part of the digital data stream. It isn't 'digital noise', which does not exist.

David

Edited by DavidHB
  • Like 1

Share this post


Link to post
Share on other sites
30 minutes ago, DavidHB said:

It isn't 'digital noise', which does not exist.

Of course ‘noise’ is produced in the digital domain:

https://www.soundonsound.com/techniques/digital-myth

There any many elements of a digital conversion/transmission process, that can produce, and add ‘noise’ to the digital signal. This may then be evident as various sonic artefacts, and not necessarily (or even likely) audible as analogue ‘noise’.

  • Like 1

Share this post


Link to post
Share on other sites
8 hours ago, kelly200269 said:

Of course ‘noise’ is produced in the digital domain:

Not as I understand the term "noise".

I go back to pre-digital days. To me noise is a set of artefacts exogenous to and independent of the signal, and typically does not vary with the signal. As "digital" is just a mental construct (the physical reality is a very large set of voltage changes in a circuit), nothing exogenous to that construct can actually exist.

Distortion, on the other hand, is a failure to render the signal correctly when it is being processed. Clearly, distortion will vary with the signal. The interesting paper to which you link lists ways in which the signal can be imperfectly rendered in the course of conversions between analogue and digital; they are, if you like, failures or limitations of the mental construct. Whatever they are called (and the terminology has been used very carelessly at times), they are all forms of distortion, not noise.

I believe that the distinction I make is significant in the present context. The OP asked us to describe what is audible from our systems when no signal is present., in other words the noise floor. The discussion he intended was not therefore about what is more correctly called distortion. The issues that are (rather well) described in your reference paper lie outside the subject area of this topic.

That said, I would commend the paper to anyone participating in this discussion as it does usefully describe a range of issues confronting those who engineer digital audio systems. The conclusion, that there is no inherent reason why such systems cannot be engineered to perform as effectively as their analogue counterparts, is pretty far-sighted for a document written in 2007, that is at about the same time as the original KDS was launched. It is interesting to speculate how the author would now write that conclusion in the light of the 13 years of intervening experience. After all, today's DSs, DSMs and Exakt systems do actually work rather well.

David

Share this post


Link to post
Share on other sites
12 hours ago, kelly200269 said:

With an Exakt system, it should be effectively ‘noise free’ until it reaches the DAC’s/output stages of the speakers. And I’m sure it is. Well. It should be if you’re ‘forking out’ £30k for a pair of Exakt Akubarik’s. Any ‘noise’ experienced in this type of system will be due primarily to the analogue output stages powering the transducers.

So yes, whilst I acknowledge that digital ‘noise’ does exist, it is an order of magnitude less than noise in the analogue domain.

The Exakt Akubarik's (with old DAC) certainly aren't noise free. very low noise yes, but it's still there.

12 hours ago, kelly200269 said:

Whilst I acknowledge, and even recognise the concept of ‘noise’ in the digital domain, it’s impact is minuscule compared to the impact of noise in the analogue domain.

Agree with that, however in an all digital system it's still there and doing things to improve it does make a difference. it might be very marginal compared to the analogue domain, but all the tiny improvements do add up.

There's also lots of different types of noise here that affect things in the different ways in the digital domain.

As an example I'd used one of @Johannes's tips which address very real problems.

The Exakt link cable makes a difference - why is that? Linn state not to use CAT 7+ cable because it can adversely impact the timing on the digital cable. Having tried this you can hear the difference, things are grainier with the wrong cable. What's happening? my guess would be the DAC is trusting the DSM's timing signal which is being thrown off ever so slightly by increase capacitance due to the shielding. the graininess might be then be DAC jitter.

His other advice like don't have dirty power sources in the same power strip as the DSM also has good reasoning and does stop noise being created in the DSM.

I think part of the issue is that people don't recognise lots of these kinds of noise in their system, mainly because they usually don't get introduced (we're good at spotting changes), instead they tend to be there from the beginning.

The start of this thread was just focusing in on the background noise floor as a way to compare systems. That doesn't mean that DAC jitter noise from the wrong Exact cable couldn't exist on the quietest of system which has no background noise floor....

  • Like 1

Share this post


Link to post
Share on other sites
4 minutes ago, Phobic said:

Linn state not to use CAT 7+ cable because it can adversely impact the timing on the digital cable. Having tried this you can hear the difference, things are grainier with the wrong cable. What's happening? my guess would be the DAC is trusting the DSM's timing signal which is being thrown off ever so slightly by increase capacitance due to the shielding. the graininess might be then be DAC jitter.

An interesting thought, but I don't know how it squares with the fact that the master clock is in the Exakt engine; as I understand it, the clock in the System Hub is, in effect slaved to the master. Maybe the issue is noise rather than jitter. As the Exakt Link protocol has never been published, it is difficult to speculate.

David

Share this post


Link to post
Share on other sites
2 minutes ago, DavidHB said:

An interesting thought, but I don't know how it squares with the fact that the master clock is in the Exakt engine; as I understand it, the clock in the System Hub is, in effect slaved to the master. Maybe the issue is noise rather than jitter. As the Exakt Link protocol has never been published, it is difficult to speculate.

David

yes it's purely speculation on my part. there maybe different clocks at play here as well, e.g. a transmission clock for data from the sender, and a master DAC sampling clock which needs synchronisation to the data stream. If you do it this way then you only need to synchronise against the data stream in 1 place, at the DAC.

If instead the DAC is sending the clocking signal to the DSM, and the DSM is then using that as a transmission clock - I think you potentially introduce additional failure modes.

Better in my brain to send a transmission clocking signal along with the data stream, that way you avoid both the DAC and the DSM having to syncronise against receiving and sending the data respectively. 

anyway, who knows! however Linn have implemented things errors can have an affect.

Share this post


Link to post
Share on other sites
1 hour ago, Phobic said:

yes it's purely speculation on my part. there maybe different clocks at play here as well, e.g. a transmission clock for data from the sender, and a master DAC sampling clock which needs synchronisation to the data stream. If you do it this way then you only need to synchronise against the data stream in 1 place, at the DAC.

If instead the DAC is sending the clocking signal to the DSM, and the DSM is then using that as a transmission clock - I think you potentially introduce additional failure modes.

Better in my brain to send a transmission clocking signal along with the data stream, that way you avoid both the DAC and the DSM having to synchronise against receiving and sending the data respectively. 

anyway, who knows! however Linn have implemented things errors can have an affect.

The Exakt link page in Linndocs (https://docs.linn.co.uk/wiki/index.php/Exakt_link) is interesting. Perhaps you looked it up, but I don't think I've seen it since it was last amended.

I had always thought that Exakt link used the the Ethernet layer, but it appears this is not the case; only the cable and connector specifications are the same, and indeed a subset of Ethernet cables is preferred. I think that

"The Exakt-link has been designed to operate with the CAT-5 cable electrical specifications for signal timings ... We recommend using the 100Base-T4 / 568B ethernet cable pairing"

may well be significant, as it suggests that particular cable specifications and pairings may affect signal timings. So you could be correct about jitter.

David

Edited by DavidHB
  • Like 1

Share this post


Link to post
Share on other sites

The more important noise level to me when trying to get maximum enjoyment from the music is the ambient noise level of the listening room.

That is when the HVAC, dehumidifier, washer, dryer, dishwasher are not running and noises from outside the home, lawnmovers, power washers, weedwackers, snowblowers are not running as well.

The low end of this ambient noise level in my home is about 37db, when the forced air heat kicks on it goes up about 10 db.  For reference, I generally listen at average levels less than 80 db.

Edited by John76
  • Like 1

Share this post


Link to post
Share on other sites

great point on the ambient noise.

I'm quite lucky to live in a quiet rural area, ignoring the hifi side of things I have quite a problem with noise generally, if I can hear anything clicking, ticking, humming, banging etc it drives me nuts.

I have to hunt things down and eliminate them!

We've got heavy studio curtains in the living room to help keep noise out as well as to provide some room treatment.

the new listening room I'm going to make next year will need to consider this as well - not yet sure how I'm going to do things but potentially rockwool in the partition wall I'm going to build is an option.

Edited by Phobic

Share this post


Link to post
Share on other sites
1 hour ago, John76 said:

The more important noise level to me when trying to get maximum enjoyment from the music is the ambient noise level of the listening room.

That is when the HVAC, dehumidifier, washer, dryer, dishwasher are not running and noises from outside the home, lawnmovers, power washers, weedwackers, snowblowers are not running as well.

The low end of this ambient noise level in my home is about 37db, when the forced air heat kicks on it goes up about 10 db.  For reference, I generally listen at average levels less than 80 db.

The biggest ambient noise I have is the people living upstairs walking across their wood laminate floor. This is annoying to say the least, with a continuous thud noise. Fortunately they are not in town all the time.

I also live close to a hospital, and can hear the helicopters landing on the roof once or twice a day. Fortunately our building is well insulated from noise, even though there is traffic noise nearby

If in doubt, I turn up the volume.

1 hour ago, Phobic said:

great point on the ambient noise.

I'm quite lucky to live in a quiet rural area, ignoring the hifi side of things I have quite a problem with noise generally, if I can hear anything clicking, ticking, humming, banging etc it drives me nuts.

I have to hunt things down and eliminate them!

We've got heavy studio curtains in the living room to help keep noise out as well as to provide some room treatment.

the new listening room I'm going to make next year will need to consider this as well - not yet sure how I'm going to do things but potentially rockwool in the partition wall I'm going to build is an option.

Building a new listening room will be an interesting learning curve.

A friend of mine had some large B&W 802 Nautilus speakers, and contacted B&W to see what the ideal listening room size and shape would be. They sent him plans to build the ideal room for his speakers. He was building a new house, so he had a clean slate to design the room whatever he wanted, with his wife’s approval.He reinforced the walls with bigger building studs, better insulation, thicker Sheetrock, and a dedicated electrical spur for powering is electronics.

How did it sound? Not very good as it turns out. We spent hours on speaker positioning and placing his Velodyne subwoofer, but it never really sounded that great. That’s said, the sound stage was massive, and with his Bryston monoblocks certainly created a convincing live concert feel.

I think proper bass traps would have been very helpful for his room, with diffusers and some absorption panels

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.