Denafrips

[The Beave]

So I got 11 posts after my post....Good, I did my job, you guys were starting to get a little too quiet on a holiday weekend!
Keep it up, posting that is, I’ve got a while before my SDG1 GETS HERE!!
Ill look now for the research part of the site, but if I need help in finding it, I’m going to ask you guys for help.
Just not a lot of math talk please, I hate math, ever since algebra, which I failed.
Beave

 

[The Beave]

The problem here being that 'resolve' is a buzzword designed to pray on the angst of audiophiles. In reality there's no bits gone missing.

It's a bit like people who think synthesized ascorbic acid is poisonous. It's not. It's chemically identical to the stuff found in nature.

Ok Killingbeans, then I’m thoroughly confused.
I just spent time “researching” this on amirm’s thread on Sigma Delta dac’s, and right there and here is his quote:

“ but the levels within the 4-bit DAC must be accurate to 16 bits if we desire a 16-bit result. Trimming and digital compensation are used in commercial multi-bit DS DACs to provide the required precision.”

I read that to mean that, indeed SD dac’s DO throw away bits from the 16 bit signal, only to Reproduce them by TRIMMING AND DIGITAL COMPENSATION, ie SYNTHESIZING what the original bits were.
This is NOT the same as decoding those original bits! It’s a synthesis of those bits and that is TWO COMPLETELY DIFFERENT THINGS.
A ladder dac, even though it’s inherently noisier, does indeed decodes all 16 discreet bits of the digital signal.
A SD dac does not, that’s what I just read, the SD part, in amirms thread on sd dac’s which I just quoted above, not my words, his.

A visual analogy is useful here, take two pictures and put them up on a hd monitor split screen.
One the original picture of a human being, next to it a CG created version of that same picture, do they look the same?
Of course the don’t, they CAN’T. The real picture looks, well, real. The CG version of it has all the characteristics of the picture but it clearly looks fake, as in not real, or analog. It clearly looks like a Computerized version of a real photograph of a human being.

This same EXACT analogy can be transferred over to the different dac technologuse and how they do things.
But from what I just read HERE, SD dac’s DO indeed throw away bits from the original stream.

Not really trying to argue or threadcrap, just again trying to illuminate for you curious here what some people find intriguing about ladder dac’s, nothing more nothing less. So please nobody take offense as none is intended.
And, yes, I am doing research here.

Love the discussion!
Beave

 

[The Beave]

Nailed it.

Alas there will always be folks who believe that compared to chip-based DACs, R-2R sounds "better" or "different". Such as some people in this here thread...

Well, be nice now because this thread IS about Denafrips. Unless you want a boring site where everyone thinks alike?
I think that would very boring, don’t you?
Now for your penance you are to buy a Denafrips Ares 2 and force yourself to listen to it for a month like I did, and then report back to us what you found, like I did. Unless you sample the other single malt scotches of the world you have no right to say only yours is the best. That’s reverse ignorance in the worst way.
Since I’m a tequila man, I’ll try any tequila at least once, and if you’d like to buy me a shot I’d gladly partake.
And thank you for being so nice!
Beave

 

[The Beave]

That's not a problem for me if this is their pleasure. But these guys could not come in the science domain. It's like to consider the earth is flat and try to prove it to me.. At the moment, there is no more possible debate because the basis are gone.

Totally spot on! And I’m understanding the scientific aspect, but somebody started this thread to explore why the interest, and I, because I have have indeed slept with the enemy, thought I could bring some answers to your questions. Now granted, there isn’t a whole lot of science when dealing with SYNERGY and how it equates to musical satisfaction in ones living room. But there is such a thing, even though it can’t be scientifically proven.
But neither can be Love..............
Beave

 

[Killingbeans]

I read that to mean that, indeed SD dac’s DO throw away bits from the 16 bit signal, only to Reproduce them by TRIMMING AND DIGITAL COMPENSATION, ie SYNTHESIZING what the original bits were.

No. The trimming and digital compensation does not touch the data stream. It is implemented/designed to keep the multi-bit DAC accurate *. There is no synthesizing taking place. It does exactly the same as an SD chip that only has a 1-bit DAC, but gives higher SNR without resorting to higher sample rates.

The main point here is that the process exchanges bit depth for sample rate. Data is not lost at any point. The chip does not "guess" on any values at any time. That would require artificial intelligence. An SD chip is a complicated thing, but it's not THAT complicated

I must point out that I'm no expert, and that other users in her will be able to explain it in much, much more detail, and probably also correct me if I'm wrong in some of my assessments, but they will likely agree that your interpretation of Amir's article and the analogy, you made from it, is misguided.

*Is it actually single bit DACs in parallel? Not sure.

 

[The Beave]

No. The trimming and digital compensation does not touch the data stream. It is implemented/designed to keep the multi-bit DAC accurate *. There is no synthesizing taking place. It does exactly the same as an SD chip that only has a 1-bit DAC, but gives higher SNR without resorting to higher sample rates.

The main point here is that the process exchanges bit depth for sample rate. Data is not lost at any point. The chip does not "guess" on any values at any time. That would require artificial intelligence. An SD chip is a complicated thing, but it's not THAT complicated

I must point out that I'm no expert, and that other users in her will be able to explain it in much, much more detail, and probably also correct me if I'm wrong in some of my assessments, but they will likely agree that your interpretation of Amir's article and the analogy, you made from it, is misguided.

*Is it actually single bit DACs in parallel? Not sure.

Very interesting!
Ok, I can jump on that train, your explanation makes sense to me. So I’ll wait while others chime in to help me understand What actually happens to those 16 original bits as it hits a Sigma Delta dac. I understand that the actual science of it is extremely complicated, that’s a given, but again I defer to my ears, and when I heard the ESS Sabre 9038 chip in the OPPO 205 I knew something had been elevated by a big margin. To my ears it was the first time I really heard digital audio without any audible negatives. But even here, on this site there are so many people liking this dac over that dac. If they test close to the same benchmark there sould only be one opinion amongst all of you, that being they all sound great, but I don’t see that here at all, quite the opposite. So how does one correlate that with science?
How can so many dac’s test so close to each other but there are so many that pick one over the other.
THAT to me doesn’t make scientific sense at all. I cant believe that one can hear 2db of Sinad from one dac to another and find the one that tested better the one. But that’s what I find here, the numbers WIN!
And that’s kinda weird to me. especially in this hobby, but I know I know this is science............
So what about the emotions of music?
Beave

 

[Killingbeans]

but I don’t see that here at all, quite the opposite.

You haven't looked closely enough. This place is crawling with tests pointing at surprisingly unassuming gear as being more than adequate.

How can so many dac’s test so close to each other but there are so many that pick one over the other.

Price, build quality, enclosure aesthetics, user friendliness, input/output options, features, reliability and so forth.

And a bit of irrationality in the form of SINAD fetishism.

But that’s what I find here, the numbers WIN!

The numbers tell something about good solid engineering. They are admirable, and in that sense they do win. But ASR is not just about numbers. It's much more about how usefull these numbers are when coupled with our knowledge of psychoacoustics and audibility.

 

[PaulD]

So what about the emotions of music?
Beave

This is VERY off-topic for this thread... Suggest you start a new one.
Emotions are in listeners, not in music - that's just sound (potentially organised). I have seen a lot of completely rubbish papers on music and emotion, some of them even get published, it is appalling. A friend is in a research group at Juliard doing some decent work in this area but nothing is published yet. (Decent, as in there are psychologists involved who have a solid understanding of human emotions, as opposed to computer scientists I've seen trying to write papers on the topic.) The emotional response to music will be highly individual and therefore difficult to generalise or study with any rigour or accuracy.

 

[Killingbeans]

This thread might be ripe for resurrection: Soul and emotion

 

[The Beave]

You haven't looked closely enough. This place is crawling with tests pointing at surprisingly unassuming gear as being more than adequate.



Price, build quality, enclosure aesthetics, user friendliness, input/output options, features, reliability and so forth.

And a bit of irrationality in the form of SINAD fetishism.



The numbers tell something about good solid engineering. They are admirable, and in that sense they do win. But ASR is not just about numbers. It's much more about how usefull these numbers are when coupled with our knowledge of psychoacoustics and audibility.

Very well stated!
So off for the night, it’s been a good day here.
I wish you all a good nights sleep!
Beave

 

[The Beave]

This is VERY off-topic for this thread... Suggest you start a new one.
Emotions are in listeners, not in music - that's just sound (potentially organised). I have seen a lot of completely rubbish papers on music and emotion, some of them even get published, it is appalling. A friend is in a research group at Juliard doing some decent work in this area but nothing is published yet. (Decent, as in there are psychologists involved who have a solid understanding of human emotions, as opposed to computer scientists I've seen trying to write papers on the topic.) The emotional response to music will be highly individual and therefore difficult to generalise or study with any rigour or accuracy.

Understandable Paul.
I’ll bow out now, my best to you.
Beave

 

[majingotan]

What actually happens to those 16 original bits as it hits a Sigma Delta dac.

LPCM data is oversampled and converted to 4-5 bits with very high sample rates by demodulation and oversampling and digital filtering then that forms pulses (pulse density modulation) then it goes to the DS DAC and then analog filtering. Oversampling R2R DACs take the LPCM data, oversamples it to 8x the rate for example then goes to digital filtering then takes the bit depth into each ladder resistors producing a total voltage then that goes to analog filtering. If the source bit rate is higher than the DAC bit rate, the least remaining bits are truncated while all of that bits are preserved on DS DACs albeit in PDM form (or DSD for 1-bit PDM files)

 

[The Beave]

LPCM data is oversampled and converted to 4-5 bits with very high sample rates by demodulation and oversampling and digital filtering then that forms pulses (pulse density modulation) then it goes to the DS DAC and then analog filtering. Oversampling R2R DACs take the LPCM data, oversamples it to 8x the rate for example then goes to digital filtering then takes the bit depth into each ladder resistors producing a total voltage then that goes to analog filtering. If the source bit rate is higher than the DAC bit rate, the least remaining bits are truncated while all of that bits are preserved on DS DACs albeit in PDM form (or DSD for 1-bit PDM files)

God Love you! In layman’s terms, thank you for that explanation!
That makes sense to me.
now I’ll bow out!

Beave

 

[majingotan]

God Love you! In layman’s terms, thank you for that explanation!
That makes sense to me.
now I’ll bow out!

Beave

It's an oversimplified way but the actual methodology and mathematics behind it are much more complicated. I can at least add that oversampling allows the DAC to simply the complexity of filtering the analog signal

 

[Dichotome]

I find it very difficult to decide where I should stand on the whole objectivism vs subjectivism debate. On one hand you have very well respected audio engineers who appear to have plenty of integrity like Nelson Pass and Bruno Putzey who are of the view that whilst measurements are important they don’t tell you EVERYTHING about how a piece of kit will sound. Then on the other hand I have direct experience which suggests measurements tell you all you need to know. I own Neurochrome DIY amps that are designed by an extreme objectivist. They measure better than almost every other amp you care to mention and they sound . . . . . SUBLIME! The most transparent and neutral amps I’ve ever heard.

One explanation the subjectivists sometimes advance is that there are factors which we haven’t yet learned to measure but which affect the way a component sounds. Maybe there is something in that; the Illustrious team at Purifi ( which includes Bruno Putzey) reckon they have come up with a new measurement that has helped them to develop their new amp modules. Perhaps other critical metrics will be discovered in the future?

 

[VintageFlanker]

I find it very difficult to decide where I should stand on the whole objectivism vs subjectivism debate.

There is no debate in my book. As simple as people who know vs people who don't...

One explanation the subjectivists sometimes advance is that there are factors which we haven’t yet learned to measure but which affect the way a component sounds.

Sure. This factor is called marketing...

Or maybe manufacturers are designing their gears with magic or something?

 

[Killingbeans]

the Illustrious team at Purifi ( which includes Bruno Putzey) reckon they have come up with a new measurement that has helped them to develop their new amp modules.

You are confusing cause and effect. Magnetic Hysteresis Distortion have been known about for a long time. Bruno simply discovered that it was causing some of the nonlinearities in class D amplifiers, that have been measurable all along.

 

[Dichotome]

You are confusing cause and effect. Magnetic Hysteresis Distortion have been known about for a long time. Bruno simply discovered that it was causing some of the nonlinearities in class D amplifiers, that have been measurable all along.

OK new in one sense but not in another. The semantics doesn‘t change the fact that they have identified a factor the effects of which were not previously fully understood. Allegedly at any rate. I don’t have a problem accepting that we haven’t yet reached a perfect understanding of what makes a component sound “good”. (Supply your own definition of good.). If you believe otherwise then fine.

 

[Killingbeans]

The semantics doesn‘t change the fact that they have identified a factor the effects of which were not previously fully understood.

Yes, but the effect was not immeasurable until then. Purifi did not suddenly discover something that was hidden in the waveform.

I don’t have a problem accepting that we haven’t yet reached a perfect understanding of what makes a component sound “good”.

Me neither. But I believe all the things that makes "good" different from "transparent" can be linked to psychology and the way evolution has shaped our auditory system in more or less fortunate ways. All the things that are worth mentioning from a technical point of view can already be measured IMO.

BTW, the creator of the Neurochrome amps is an active member here on ASR

 

[Frank Dernie]

Ok Killingbeans, then I’m thoroughly confused.
I just spent time “researching” this on amirm’s thread on Sigma Delta dac’s, and right there and here is his quote:

“ but the levels within the 4-bit DAC must be accurate to 16 bits if we desire a 16-bit result. Trimming and digital compensation are used in commercial multi-bit DS DACs to provide the required precision.”

I read that to mean that, indeed SD dac’s DO throw away bits from the 16 bit signal, only to Reproduce them by TRIMMING AND DIGITAL COMPENSATION, ie SYNTHESIZING what the original bits were.
This is NOT the same as decoding those original bits! It’s a synthesis of those bits and that is TWO COMPLETELY DIFFERENT THINGS.
A ladder dac, even though it’s inherently noisier, does indeed decodes all 16 discreet bits of the digital signal.
A SD dac does not, that’s what I just read, the SD part, in amirms thread on sd dac’s which I just quoted above, not my words, his.

A visual analogy is useful here, take two pictures and put them up on a hd monitor split screen.
One the original picture of a human being, next to it a CG created version of that same picture, do they look the same?
Of course the don’t, they CAN’T. The real picture looks, well, real. The CG version of it has all the characteristics of the picture but it clearly looks fake, as in not real, or analog. It clearly looks like a Computerized version of a real photograph of a human being.

This same EXACT analogy can be transferred over to the different dac technologuse and how they do things.
But from what I just read HERE, SD dac’s DO indeed throw away bits from the original stream.

Not really trying to argue or threadcrap, just again trying to illuminate for you curious here what some people find intriguing about ladder dac’s, nothing more nothing less. So please nobody take offense as none is intended.
And, yes, I am doing research here.

Love the discussion!
Beave

Don't forget at the beginning all DACs were discreet like R2R. It is the obvious way to convert the bits to analogue. Whilst obvious and simple to understand it is very expensive to make because the accuracy of the larger bit resistances have to be extremely high so as not to swamp the contribution of the lowest bit resistance. This is extremely expensive to achieve.
A much cleverer, but less easy to understand for the layman, method was developed for manufacturing engineering and cost reasons and when its performance exceeded that of the expensive discreet DAC chips it took over.
The old type have only survived in inevitably expensive to very expensive implementations for non-technical enthusiasts who try to understand how their hobby works. They can perhaps just about understand how this sort work but seem to be not sufficiently mathematically "literate" to understand the more sophisticated but engineering-wise simpler solutions.
The whole point of good engineering development is to get better perfomance al lower price. This doesn't sit well with a fandom which equates more expensive with better...
I have about a dozen DACs here either separate or in CD players. The discreet ones were much more expensive, even in their day. It is not just the conversion technology that counts but the whole device, from circuit board layout, earthing filtering and analogue stages. In the end the only connection to the outside world is the output terminals and the only thing there is the magnitude frequency and phase, there is nothing else. We are able to measure these to a level well beyond anything a human can hear.
It doesn't matter how the conversion is achieved nor how it is filtered or amplified, as long as the output is accurate.
Things like what sort of DAC method or make and model of chip or exotic componentry are really irrelevant if the output is accurate. It is extremely difficult and expensive to get an accurate output from a discreet DAC so they are an order of magnitude more expensive for similar or often inferior performance.
However IME and IMO they all exceed the level necessary to get all the information from an actual music recording since the SQ limit is in the recording and the music's dynamic range not the DAC.