Author Topic: ADC/DAC losses  (Read 7854 times)

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
ADC/DAC losses
« on: August 05, 2014, 11:17:50 PM »
I believe ADC's (of all varying types) rob the music of "something".   It's not easy to put your finger on it,    digital just lacks a certain weight and body that isnt there. 

Why is it so?    Is it our playback systems,  or is it the ADC?   A combination of both perhaps?   

The Killerdac is designed to put back, what digital music lacks.   It does this incredibly well.   

But I have this irrepressible feeling that the ADC step is a significant part of the problem, that no one is really focussing on. 

If someone could design and develop an ADC,  that sounded as real and lifelike as a top quality tape machine does,   then this is a game changer.

It's possible.   We just need the desire, inspiration,  and intricate knowledge of the things that matter.   

I dont believe there is any ADC manufacturer that has cottoned onto this.  It makes me wonder what systems they listen to, when they are developing their devices.   Who knows. 
« Last Edit: August 05, 2014, 11:47:00 PM by ozmillsy »
It's all about the music,, not the equipment.

Offline rhlauranna

  • Hero Member
  • *****
  • Posts: 1202
  • Liked: 156
Re: ADC/DAC losses
« Reply #1 on: August 06, 2014, 01:25:55 AM »
yes, thank you, this topic is a very good idea, and I have no doubts that the results are nothing but spectacular, but deeply inside I fear the whole project will more likely remain a suberb idea - inspite of the fantastic results...

why? because most of us (nearly everybody) have (no longer) a machine to play back adequately (not regarding the exceptions) nor do we have any chance to put our hands on that music stored on mastertapes that we like that much, we are only able to lay our hands on the finally "produced" results, i.e. CD, SACD, DVD-A, Bluray-Audio, direct downloads in whatever resolution... there are only very few people who can live out their "imaginations" within the recording area...

well, Stefano pointed "Hemiolia" already out some time ago with some very interesting pixes and descriptions... and as he told me, he is very content with the results... you might want to have a look here:

http://twogoodears.blogspot.de/2013/01/hemiolia-tape-project-more-pixes-and.html

and Stefano - like you - knows what he is talking about, he is a fanatic  in tape recoding - not only owning a Telefunken M15(a?) but as well a professional Studer machine, not to speak of numerous highest end microphones from Neumann and other companies, from the 1940ies, 1950ies... (sorry if I cannot quote them all), making his own recordings for years... and I am always very impressed when he listens to "my" reproduced (commercial) music here with me saying immediately with what mics and with what "arrangement" it has been recorded)...

well, although not personally tested, I agree with you saying: "I believe ADC's (of all varying types) rob the music of "something"...

but for quite some time I do (no longer) agree with your saying: " It's not easy to put your finger on it."...  For me (us here) - meanwhile - it is "easy" to put the finger on it: it is the squares, the same "problems" and/or "imponderabilities" we had to experience in reproduction (generating) the sound we have to face - of course - while transforming from analogue to digital (ADC)... it appears just to be the other way round...

I do (no longer) agree with you saying "...digital just lacks a certain weight and body that isnt there..." no, that problem meanwhile definitely is "solved" by multiplying chips in combination with controlled power supply... otherwise "weight" and "body" and I have to add "natural precision and warmth" simply are not generated...

it is just the controlled and really strong power supply that "adds" that what you rightly feel as "lack" in digital... separated for 5 Volts and 12 Volts...

and within this context there are some dramatical things going on which I cannot yet commentate on with the integration of Shunt Labs...

..and from what I can see, all these things are (still) exactly missing on the transferring side ADC...
well, it is not that we did not try out, for example, Weiss-products, or Berkeley, or... we did as well test out others (and measured them)... on the reproduction side, to get an impression of that what is "reference" in recording studios when playing back...

...and we had the possibility to test the Weiss and Berkeley DAC for several weeks side by side in competition with all our stuff, and yes, here I agree with your saying: "...digital just lacks a certain weight and body that isnt there...", and yes, right you are, this way we do not want to listen to music, the origin already is simply generated "steril" and "castrated" and... "aneamic"...

but with our DACs this is not the case, perhaps I should say "no longer", because to get "there" took us nearly five years of intensive research and building and testing...

and with your saying: "But I have this irrepressible feeling that the ADC step is a significant part of the problem, that no one is really focussing on. " you bring it to the point...

well, way back I referred already about this topic on tuyen's topic... but as this is the recording side, we kept away from this, because it simply is not our "domain"...

you write: "If someone could design and develop an ADC,  that sounded as real and lifelike as a top quality tape machine does,   then this is a game changer."... yes, I have been "dreaming" of this for years, and for me it could be Doede to do this, but at the time he told me that he is not interested (not only because of lack of time)... but someone should construct this according to the experiences with the DDDACs (or Killerdac - I don't know) on the reproduction side...

and yes, I am deeply convinced, there are quite a lot of additional sound "resources" still to be put free... but, one has to keep in mind that the DAC-side alone is still not yet finished... there are some really interesting developments and discussions going on here:

http://www.diyaudio.com/forums/digital-line-level/224108-nos-192-24-dac-pcm1794-waveio-usb-input-240.html

nevertheless, we have made really great progresses within the last years, that great that it simply was not imaginable, that way, that when people listen now they nod with their heads, but the end is still not yet reached...

my proposition for those who want to face ADC - which is indeed the indispensably necessary next step - would be to "wait" till the "end" in the developments of the DACs is reached, at least within the DACs itself, and then head to ADC... in order to have it "easier" and "simpler" to "save" quite a lot of work in developing...

it appears as if it is "only" the other way round:





and here Doede's results regarding 1000 cycles and 20 cycles... absolutely top notch...








anyway, a fantastic idea and I will gladly read about the progresses...

« Last Edit: August 06, 2014, 01:53:47 AM by rhlauranna »

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #2 on: August 06, 2014, 07:39:11 AM »
It's interesting, isn't it Reinhard.   

There are 2 sides to the digital music coin.    ADC then DAC,   if what is stored in the digital files is not authentic,  then I feel we will always be striving for more.

Obviously I can't comment on how far you all have come, with the dddac,  one day I will visit and hear.     But I will say that the Killerdac crew have a very different approach,   their pursuit is to add colours until the music sounds real.    The really interesting thing is, this approach isn't a fixed formula,   each hifi system is different and unique,  and what the system needs (to sound real) can and does vary from system to system.   

As I sat and listened to my Hemiolia tape last night, it occurred to me that colouring the dac does work wonders,  but if the ADC was more true to the analog tape sound,  then this changes the equation.   We would still need to tweak our individual systems, can't get around that, they are all unique.   But we would now no longer be trying to add the things that digital music lacks.

   How do the various ADC manufacturers know that their devices result in an authentic capture?   You would think they are actually 'listening' to the results.   Listening on what?  What is their baseline, what is their reference, what is their approach for judging and comparing?   Impossible questions for us to answer, but would be good to have that insight.   

The other aspect to the recording side,  is this tendency to digitally fix everything.  I also feel the addiction to protools is contributing to the artificial nature of modern music.    The imperfections are part of the art, it's part of the human emotion and makes the performance authentic.     Fixing everything to clinical perfection,  is self defeating to some degree.   If there are big problems on the recording that the producer and artist can't live with,  then do it again.   


« Last Edit: August 08, 2014, 12:53:03 PM by ozmillsy »
It's all about the music,, not the equipment.

Offline rhlauranna

  • Hero Member
  • *****
  • Posts: 1202
  • Liked: 156
Re: ADC/DAC losses
« Reply #3 on: August 06, 2014, 09:00:27 PM »
"There are 2 sides to the digital music coin.    ADC then DAC,   if what is stored in the digital files is not authentic,  then I feel we will always be striving for more."

yes, yes, yes, right you are... personally I have come to the conclusion (by still endless listening and comparing...) that if we are given superb recordings (I mean direct PCM digital - even if some are still downsampled at the moment), for example the chinese drummer

Hok-Man Yim, Master of Chinese Percussion (2008) K2HD



and/or McCoy Tyner - Double Trios [1986]



two things - to my ears - become obvious:

1. the "general" superiority of digital "against" analogue regarding dynamics (by far), precision, "attack", naturality, sheer power and brillance... and

2. that all (with all respect) "attempts" to digitize analogue sources are limited by their native origin...

when I crank these two sources on my system to the limit, I have the feeling I have it "live"... well, get me right, not to get it loud, but to get the instruments being played in reproduction as they were played while being recorded...

during my vacation in The Netherlands I happened to have the opportunity to listen to an organ concert for about two hours in a "rather small" church (in comparison to Cologne cathedral - not to be disrespectful, just to describe things as they are)











first I thought, oh, what a nice opportunity to further "educate" my ears ... of course, but then some things came to my mind:

1. this organ (it is a "small" organ in comparision to Cologne cathedral - you understand my intended "sense") played very nicely and without any "faults"... to my ears the "best" way to get it reproduced is "digital", I do not know of any "good" analogue recording that could ultimately "match" that what I heard being played this way here from sheer dynamics and brillance... and

2. while listening and watching the place where the organ is put on the wall and how the sound distributes everywhere within the room, I became aware that absolutely no things have been "installed" and/or adjusted for "room correcation" and/or "room adjustments".... the walls are just walls, and that "steril" and "clean", no absorbers or whatever, same with the floor and the ceiling, and then that sound... the organ played that naturally within the church - and I sat on half a dozen places within the church, in front, in the middle, aside, in the back - the impression was nearly always the "same", and the sound of the organ did not become lower when further away from the instruments (as is with speakers)... so, less is more (natural)... to my ears a reproduction of this organ with a "perfect" digital recording should be hardly a problem (the last two octaves form Cologne cathedral organ is another story of course - I cannot handle this like Klaus because of my room length...

you say: " Obviously I can't comment on how far you all have come, with the dddac,  one day I will visit and hear.     But I will say that the Killerdac crew have a very different approach,   their pursuit is to add colours until the music sounds real.    The really interesting thing is, this approach isn't a fixed formula,   each hifi system is different and unique,  and what the system needs to sound real can and does vary from system to system."

I am convinced that we are "looking" for the same thing... well, at the moment a direct exchange/visit seems not possible, but you seem to have several opportunities:

1. from what I read here on the site and on the DIY-site there are already a few users with DDDACs in your country, and the one or other with ultimate efforts (further ahead than we are here...) I don't know where they "live", west. east, north, south, but perhaps you might have a listen and/or comparison with them and/or invite them to your sessions....

2. I cannot see any attempt to adapt or try out - to me - the "greatest" discovery in amelioration of the digital sound, i.e. "multiplying" chips, with the Killerdac you all use... from my point of view this should be the "easiest" way for you to try out to get even more "gain" in your sound or not... is there any reason why this did not happen during the last years ? get me right, this is just a question for which I have not found an answer so far and which makes me wonder a little bit...

you talk of "colouring"... if you feel there is "need" for colouring, then - to me - something is lacking right in the origin...

you write: "...it occurred to me that colouring the dac does work wonders..."

when I think of the organ in the dutch church, there have been made no attempts at all (within the sourroundings) to take whatever "influence/colouring" on/of the sound... why? well, to my ears it is not necessary... the organ simply played as is... and it is just right...

it appears: if the source is right, no ameliorations/colorations/tweekings/adjustments or whatever changings are necessary at all...

so, if one has all the frequencies, and has them just right in the beginning, and the equipment to amplify and reproduce, why should one take influence on the sound ?

and you say further: "But we are now longer trying to add the things that digital music lacks." not always to repeat myself, but perhaps you have become aware on the DIYer site how many of them state that they with their personal systems have never ever listened to music that good then with the DDDAC (you understand me: I am not an advertiser, I just try to give hints...) and from what I read there is the one or other having as well a Killerdac...

yes, and you are of course absolutely right regarding all the other imponderabilities on the ADC and DAC-site...

anyway, there is still a lot more do discover and experience, and that's what the joy is all about, isn't it ?

« Last Edit: August 07, 2014, 08:17:44 PM by rhlauranna »

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #4 on: August 07, 2014, 10:58:13 PM »
that if we are given superb recordings (I mean direct PCM digital -
My understanding is that virtually all ADC chips are delta sigma variants of some kind,  with internal PCM conversions happening.     1 of the reasons why purists love DSD,  is that the data is pure, it is the straight delta sigma capture with no manipulations happening. 

Things may have changed lately, and there could be genuine full 24bit word multibit ADC chips now.  Not sure.   ???

My concerns isnt necessarily to do with the chips though,  it's also to do with the electronics around the chip and how things have been implemented.   There just seems to be a consistent problem with digital recordings/transfers.

Quote
during my vacation in The Netherlands I happened to have the opportunity to listen to an organ concert for about two hours in a "rather small" church (in comparison to Cologne cathedral - not to be disrespectful, just to describe things as they are)

Live music is an important reference.   Cathedral organs can certainly set some perspective.

Quote
I am convinced that we are "looking" for the same thing... well, at the moment a direct exchange/visit seems not possible, but you seem to have several opportunities:

1. from what I read here on the site and on the DIY-site there are already a few users with DDDACs in your country, and the one or other with ultimate efforts (further ahead than we are here...) I don't know where they "live", west. east, north, south, but perhaps you might have a listen and/or comparison with them and/or invite them to your sessions....
Send me a PM with a contact, and I'll make every attempt if it is in any way practical?   WA is a bit far though.

Quote
2. I cannot see any attempt to adapt or try out - to me - the "greatest" discovery in amelioration of the digital sound, i.e. "multiplying" chips, with the Killerdac you all use... from my point of view this should be the "easiest" way for you to try out to get even more "gain" in your sound or not... is there any reason why this did not happen during the last years ?
Yes, absolutely.   The heart of the Killerdac is the TDA1541A chip.   And the good ones are not cheap.   Cost and availability are an issue.   I also dont have the skill, to pursue such developments.  I am more of a consumer hack.

Quote
you talk of "colouring"... if you feel there is "need" for colouring, then - to me - something is lacking right in the origin...

Well, yes and no.    This is a philosophical view of mine.   My feeling is,  everything is coloured.    No 2 systems sound identical in every single way.   Even if you copy every single component and cable,  different rooms come into play.     

Everything is coloured,  and every system needs to be dialled in to some degree.  IMHO.    You can play with cables,  and you can play with room treatments,   and you can tube roll in your amps,  if you run tube amps,   all these techniques are things we do,  to get the tonal balance right for our given system.   They are compensating colours. 

Like all of this,  we also have the ability to add colours to the DAC  by changing components within it.    Why do different brand caps and resistors make a difference when they have the same rating?   They are colouring the sound.     And I must say,  it is extremely powerful to be able to tune in with changes to the front end source component.

Quote
it appears: if the source is right, no ameliorations/colorations/tweekings/adjustments or whatever changings are necessary at all...
That is 1 way of looking at it.   I will say,  of my 3 front ends,   digital vs vinyl vs R2R,   that R2R is the best, and certainly "seems" right.    So I use it as my reference.


Quote
anyway, there is still a lot more do discover and experience, and that's what the joy is all about, isn't it ?
Absolutely.  Where will it ever end?   It never ends.  :-\
« Last Edit: August 07, 2014, 11:00:30 PM by ozmillsy »
It's all about the music,, not the equipment.

Offline rhlauranna

  • Hero Member
  • *****
  • Posts: 1202
  • Liked: 156
Re: ADC/DAC losses
« Reply #5 on: August 08, 2014, 02:58:17 AM »
this is a good title for the movement !!!

My understanding is that virtually all ADC chips are delta sigma variants of some kind,  with internal PCM conversions happening.     1 of the reasons why purists love DSD,  is that the data is pure, it is the straight delta sigma capture with no manipulations happening.

o.k., but that is not what I meant... I mean that there is "quite a large" difference (a lack if you want) right in the origin... any analogue recording has its limits... for quite some time I am no longer in recording machines, but I have in mind from some reports back then that the Studer A80 and the Nagra IV machines, when precisely adjusted and tuned, were able to "catch" up to 100dB during the recording process, but not more because of tape saturation, so rather "equal" to that what a CD (96dB) is capable of performing, if "fed" right...

but even if we switch to 24 bit in ADC for such an analogue recording there is no "more" than this 100 dB (which indeed is quite a thing, I do not want to discuss this "small", no, no...)... but the "pure" digital PCM-recordings (without being analogue in recording chain), even if formerly "only" 20 bit (20x6dB=120dB, have already been capable in storing much more information, and from what I hear this - to me - definitely shines superiorly through, even if downsampled and "only" played back via ripped CD... that is the difference that I mean...

My concerns isnt necessarily to do with the chips though,  it's also to do with the electronics around the chip and how things have been implemented.   There just seems to be a consistent problem with digital recordings/transfers.

I know, what you mean and within certain areas one is simply bound to what is used... for me the electronics have to be the most "neutral" possible in order just to not "colour" the sound in order not to cure the symptom later on, but to erase the cause right in the beginning (as far as that is/might be possible)...

I also dont have the skill, to pursue such developments.  I am more of a consumer hack.

same with me... what a pity...
 
My feeling is,  everything is coloured.    No 2 systems sound identical in every single way.

...as is in nature, no two instruments or voices or whatever sound identical in every single way... yes, of course, choose the best for your ears, but there is still so much in the origin...

that R2R is the best, and certainly "seems" right.    So I use it as my reference.

sure, but regarding the music that I like I will not be able to play anything of it as a master analogue recording... I am already satisfied with Jean's Beatles glasmaster CD... I will see if I can get that...

just a question, did you ever play back a "pure" 24 bit digital master on your system to hear the differences? that would be interesting...

you asked me for some connections regarding DDDACs in your country... well, as it is an official site and as they officially talk about their experiences, I well might quote their nicknames: Chanh (Perth Western, very very far ahead of us...), Tuyen (you know how creative he his), PET-240 (Brisbane Northside)... if you want, you might contact them here...

http://www.diyaudio.com/forums/digital-line-level/224108-nos-192-24-dac-pcm1794-waveio-usb-input-243.html

ha, and if you should meet, that should become a real festivity, shouldn't it?

« Last Edit: August 08, 2014, 06:12:19 AM by rhlauranna »

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #6 on: August 08, 2014, 08:44:37 AM »
Yes I have.  I own a large collection of high resolution music.

I have sacd players, I have 24bit dacs, I've tried all the formats over the years.

In my experience there is far more to be gained through our hardware choices, than there is in worrying about software formats.

All digital formats suffer the same problem,  they lose weight and body, they become anemic, often grey.  And there is 2 parts to this,  the ADC step and the DAC step.

We're all focussing our energy in addressing the DAC step.   But is that the real source of the problem?   It's only half of the problem

No one is focussing on the ADC.   Sure there might be new chip developments happening,  but these manufacturers are focussed on specs, they are focussed on theory and numbers.    What companies are focussed on taking the chips and making actual devices?   Korg, Weiss, Prism, etc?   What are they doing?    My impression is they seem to be focussed on the wrong things.
« Last Edit: August 08, 2014, 10:39:11 AM by ozmillsy »
It's all about the music,, not the equipment.

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #7 on: August 08, 2014, 12:48:50 PM »
... for me the electronics have to be the most "neutral" possible in order just to not "colour" the sound in order not to cure the symptom later on, but to erase the cause right in the beginning (as far as that is/might be possible)...
This point you touch on WRT to neutrality, is really important.

How do you assess if something is "neutral" ?    Please avoid talking about square waves.

When you listen to the dddac,  what is your measure for determining that the dac is actually neutral?
It's all about the music,, not the equipment.

Offline zenelectro

  • Hero Member
  • *****
  • Posts: 793
  • Liked: 151
Re: ADC/DAC losses
« Reply #8 on: August 08, 2014, 12:54:45 PM »
Yes I have.  I own a large collection of high resolution music.

I have sacd players, I have 24bit dacs, I've tried all the formats over the years.

In my experience there is far more to be gained through our hardware choices, than there is in worrying about software formats.

All digital formats suffer the same problem,  they lose weight and body, they become anemic, often grey.  And there is 2 parts to this,  the ADC step and the DAC step.

We're all focussing our energy in addressing the DAC step.   But is that the real source of the problem?   It's only half of the problem

No one is focussing on the ADC.   Sure there might be new chip developments happening,  but these manufacturers are focussed on specs, they are focussed on theory and numbers.    What companies are focussed on taking the chips and making actual devices?   Korg, Weiss, Prism, etc?   What are they doing?    My impression is they seem to be focussed on the wrong things.

Oz,

There is actually a lot of work being done in the ADC arena, most hi end people are just not aware of it.

The first and absolutely crucial thing to understand is that chip makers will not design an ADC chip that -adds- colour or is euphonic.
Their goal is to accurately reproduce what is going in. I think that is completely understandable and it would be somewhat ludicrous
to do otherwise.

Tape does -not- do this. Tape changes, in a nice way, what goes on to it and this is very well known, at least in pro circles.
Many hi end download labels claim analog tape to be the most accurate storage format. This is simply not true. In many cases
it is the best sounding format but it is not the most accurate format.

OK, once we have established that, then I think it's appropriate to look at what 'damage' each format does to the source, IOW
how each format actually changes the source.

Tape has a lot of interesting effects from reduced resolution, to what some engineers call a 'glue' effect when using many tracks
of a multi track. It also has a LF 'head bump' , or a LF boost, depending on speed and tape type. It also adds a lot of distortion,
mainly 3rd harmonic. Many classic rock records done on analog tape actually exploit the tape compression when hit hard (high levels).
You can hear all this, if you know what to listen for, I certainly can.

Digital is a very different animal. More resolution but also because of that added resolution many very subtle artifacts manifest and
it's often hard to even get a handle on what they are. Here are some of these artifacts:
- Delta Sigma modulator artifacts with non steady signals. ESS (Sabre) identified these and tried to address them in their products
- Digital filter artifacts. In the ADC these are from the decimation filters. All audio ADC's run at around 6MHz with anywhere from 1 to a few
bits and that digital signal is decimated down to whatever PCM format is required.
- Noise floor artifacts. These can be increased noise above 20kHz or actual noise floor modulation with the signal. AKM semiconductor has
done quite a bit of work on this side of things.
- Jitter at the ADC end. Once it is encoded you can't get rid of it, doesn't matter what trickery you do in the DAC. You can probably cover it up a bit.
Just about all pro ADC's need to be able to synchronize to other units when multiple units are running. They do this with word clock, you can
use one ADC as the 'master' and send it's word clock to the 'slave' ADC. Some people also use a 'house clock' and slave everything off it.

This is all good and well but the problem is the ADC chip doesn't run off, for example a 44.1kHz word clock, it requires a much higher freq clock
usually 11.2896MHz or double that. So  that HF clock must be made up  from the 44.1kHz word clock. They do this with a PLL (Phase locked loop)
which basically means that the internal 11.2896 clock is 'tracking' or 'following' the 44.1kHz external word clock. Designing a very low jitter PLL
system that can track an external WC is very difficult and if you are talking about levels of jitter that say a zen clock has, well there is no
PLL system that can come close. But it's hard to sell a professional ADC that doesn't have a WC input and slave-ability. In an ideal world we
would have settled on one frequency, say 22MHz master clock and just pump that clock around the studio slaving any ADC's directly off it.

One of the reasons I like the Korg MR unit, it has two fixed clocks that can be upgraded to very low jitter units. Much better than a PLL.

OK, you suggest ADC and DAC chip makers don't do much listening to their products, this is not true at all. A -lot- of listening goes
on with these chips. But again - we have to go back to our initial discussion, they are not trying to make their chips sound like an
analog tape machine. They are trying to make the chips reproduce exactly what goes into them accurately. This makes perfect sense
to me.

They will leave any euphonic license to the product designers. And contrary to opinion here there are quite a few very high quality
pro ADC boxes that do add some analog spice from very subtle to not so subtle. In the very subtle variety I would say Forssell
Technologies, not so subtle would be Burl Audio (bomber) JCF Latte etc.

Other top end pro ADC's just want the box to be transparent, an example would be Prismsound, Pyramix Horus, DigAudio Denmark etc.

And if you do a bit of forum trolling, sound engineers are pretty spread across the board WRT which they like to use. Some like ]
very transparent, some like some color.

I can't really say much more, it gets a bit beyond the scope of this forum but suffice to say just slandering pro audio ADC chip
makers and ADC box makers is, IMO complete folly. There are lot's of people out there doing lot's of work trying to get a handle
on the understanding of digital artifacts and how they manifest, it's a complex business. Add to that the whole cost competitive nature
of products and those compatibility issues mentioned and you start to get a clearer picture of why we are where we are.

It's also worth considering that a decent 24 track 2" analog tape machine probably cost well over $50,000 back then and you get
some idea of the challenges. People these days scoff at paying much over $5k for a hi end converter and they usually want 8 ins and
outs for that.

I would suggest just be patient, I'll have that Korg going in a while and I will address most of the issues. The ADC in the Korg is a
true 1 bit design and whilst it doesn't have absolute cutting edge DR figures of over 120dB I don't think that's important. At DSD128
it is a very good converter with very few of the PCM artifacts discussed above.

I could write pages and pages on this whole subject but I'll leave it at that.

As a finishing note, it's worth considering that next year AKM is releasing a hi performance range of converter chips including
DAC's ADC's and Sample Rate / format converters (SRC). They -all- do PCM at 32 bit up to 768kHz (yep) and quad speed DSD256
(11MHz). The format converter converts any format, (DSD or PCM) to any other  (DSD or PCM).
So I expect this whole hi rez / DSD episode is just beginning and we may yet see some very interesting products.

cheers

Terry

Offline zenelectro

  • Hero Member
  • *****
  • Posts: 793
  • Liked: 151
Re: ADC/DAC losses
« Reply #9 on: August 08, 2014, 01:13:37 PM »
This point you touch on WRT to neutrality, is really important.

How do you assess if something is "neutral" ?    Please avoid talking about square waves.


Easy, do a 'round trip'. Analog signal -> ADC -> DAC then compare to original. This has been going on in pro circles
for years and converters these days are getting a lot more neutral than you might think.

The issue is, you guys -don't want neutral- and subjectively most other people don't either.

So the question becomes, where do I want to add the spice. The answer is. well,  anywhere you like.
Just look at historic recording chains: 

- Mics - One of the most highly sort after = Telefunken U47 valve mic = very rich colored sound, not close to neutral.
- Mic Pre amps - Telefunken V72, same goes, not transparent at all but lovely rich warm sound.
- Recording format - Discussed above, analog tape.
- Mixing consoles - Many very old consoles like Heil and those EMI valve Redd (used for Beatles) are very sought after
for same reasons.

After a while you get to understand that the process of recording is pretty much like baking a cake and a very
skillful combination of mic techniques, gear and - great performance to get the result. As such you can also understand
we have many different 'cooks' (engineers) with different gear, experience, budgets and just sonic preferences.

Personally, I find the whole recording process much more fascinating than just replay. It's kind of the big picture
and if you can see all the challenges that are involved with just DAC's then you can understand how complex it
gets when everything has to be considered in the whole chain.

There's a lot to know, but it's all very fascinating. 


cheers

Terry   

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #10 on: August 08, 2014, 03:11:18 PM »
The first and absolutely crucial thing to understand is that chip makers will not design an ADC chip that -adds- colour or is euphonic. Their goal is to accurately reproduce what is going in. I think that is completely understandable and it would be somewhat ludicrous to do otherwise.

Don't get me wrong, I think it is definitely the job of the ADC to transparently capture what it is given.    My point is,  I don't believe they do. 

The capture and subsequent playback,  lacks life and realism.

Quote
Tape does -not- do this. Tape changes, in a nice way, what goes on to it and this is very well known, at least in pro circles.

I'm certainly aware of the practise to run (already mixed) recordings through a tape machine, to capture the colour of the tape machine.

Quote
Many hi end download labels claim analog tape to be the most accurate storage format. This is simply not true. In many cases it is the best sounding format but it is not the most accurate format.

The whole accuracy thing is a bit of a rabbit hole.      On my humble home system, R2R sounds the most "true to life" , and by a country mile.     Whether it is accurate or not,  it is the most lifelike,  it sounds real.

Quote
OK, you suggest ADC and DAC chip makers don't do much listening to their products, this is not true at all. A -lot- of listening goes on with these chips.
In re-reading my posts, I can see where you are coming from.       I do question things like,  what they are listening on,  how they are listening,  and what they are doing to compare - what is the baseline?

Let me give you an example, on what I do to compare.     My reference is my R2R, because it sounds the most lifelike.    Lets put aside whether the whole tape recording process is distorting the original sound or not, lets just assess what we hear.   

1) Baseline:  Otari MTR-10 -> Radford Amp -> Tannoy Golds = sonic bliss
2) ADC:   Otari MTR-10 -> Korg MR1 (dsd64 record) .  Korg MR1 (replay) -> Radford Amp -> Tannoy Golds = disappointment,  sounds sh!t.  But hang on, it's a hand held recorder, with a mickey mouse dac and output stage.  Unfair.
3)  Take the Korg DSD64 files ->  Convert to 24/88.2 ->  MacMini -> Metrum Octave dac ->  Passive Pot -> Radford Amp -> Tannoy Golds = underwhelmed,  anemic, grey, yada yada
4)  Take the Korg DSD64 files ->  Convert to 16/44.1 ->  Burn CD -> CD94 transport -> Killerdac -> Passive Pot -> Radford Amp -> Tannoy Golds =  a lot closer to #1,  bit still not quite there.

So,  I have to say,   will it ever be there?    Will the whole digitization process ever be truly transparent?    Is it unfair to expect that it is transparent?     #4 gets a lot closer to #1,  because the Killer is adding the colour that the digital music lacks.   

Are my tests mickey mouse,  are my systems mickey mouse (in comparison to what the vendors use to judge) ?       

I'm genuinely interested to know the systems and processes that the ADC manufacturers use to assess whether they have a transparent product?      If they are listening to thin and anemic sounding equipment,  then this is a problem.    I'm not saying they are,  but you get me.


« Last Edit: August 08, 2014, 03:13:45 PM by ozmillsy »
It's all about the music,, not the equipment.

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #11 on: August 08, 2014, 03:19:10 PM »
Easy, do a 'round trip'. Analog signal -> ADC -> DAC then compare to original. This has been going on in pro circles
for years and converters these days are getting a lot more neutral than you might think.
It's good we are saying the same things,  and reassures me that I am not off on another planet.     That's exactly what I did,  note the last post.      We'll see how far we get,  when the MR1000 is upgraded.   That really will be interesting.

Quote
The issue is, you guys -don't want neutral- and subjectively most other people don't either.
What *I* want,  is true to life sound.      I don't get this with stock SS equipment.     Anemic/thin,  grey,  these colours are not true to life.
It's all about the music,, not the equipment.

Offline Tuyen

  • Hero Member
  • *****
  • Posts: 554
  • Liked: 211
Re: ADC/DAC losses
« Reply #12 on: August 08, 2014, 04:01:11 PM »
What's interesting is that these 'colours' that bring the 'life' you are after, aren't even able to be depicted as frequency response variations in the outputs, as we have discovered that the output FR from a KillerDAC is in fact basically just as 'flat' as any other DAC out there (albeit bit stronger rolloffs at both ends).     

My thinking is that there will always be something 'lost' after every (ADC->DAC or vise versa) conversion.   So because of this, one can never expect it to sound exactly the 'same' or as good as the original.   Only just 'good enough'!  :)

Offline rhlauranna

  • Hero Member
  • *****
  • Posts: 1202
  • Liked: 156
Re: ADC/DAC losses
« Reply #13 on: August 08, 2014, 08:24:26 PM »
so, you want it "live"...  ? with what colours ???



then let me tell you, you are not alone...

when I was young I came across this report in a magazine from 1976 (!!!)... (please don't tell me about the quality of the SS at that time... Jean was just on his way to invent Class A, anyway, together with Crown DC300 and Audio Research and Mark Levinson amps it belonged to the best at that time without any doubt... we had them then...

and besides Jean it was this guy from which I learned to develop and keep an eye and "care" for my most beloved squares (it is - to me - always listening  a n d  measuring and listening  a n d  measuring... and finally my ears decide of course...) I still own a copy of another report where exactly the squares of this system are presented... and I have to tell, these were the best at the time... but by far not that what we are used to today...

but what is of most interest: Richard S. Burwen played the drums personally and recorded himself on his different tape recording machines and played the recordings back in the   s   a   m   e   room, and to me that's what it is all about in comparison of playing "live" and playing a recording: it has to be done in the same room, otherwise it is just not "comparable", isnt't it ?





another guy from Japan,  Mr. Kobayashi, at the same time (1978) did the same (playing drums and recording and playing back in the same room with Studer A80 and active (at that time !!!) 5 way system, with 6 meter bass horns lying in the garden (with double D-1250 from YL mounted on each, which I was about to buy if the Goto SG146LD had not appeared on the market a little later) and finishing at the entrance of his living room...

so you might imagine that those two systems (1976 and 1978) - although I never had the chance to listen to them - blew me away... and Mr. Kobayashi was quoted to have the reproduction there where it should be exept of the highest highs (experimenting with some Fane ion drivers from Englang at the time...)

...and the quote "to have the reproduction there where it should be" from then on made me inquiet for the rest of my life...

you will find the system from Richard S. Burwen in detail - which meanwhile has become digital as well - here or elsewhere when you google (the other one from Mr. Kobayashi I will try to find)

http://www.scrounge.org/speak/burwen/

and here

http://www.burwenaudio.com/Biography.html

http://www.burwenaudio.com/Sound_System.html


not to put any oil into any fire, but you might want to read about the results from Richard S. Burwen regarding his researches in tubes (for more than 40 years) which might be of special interest for the general state you are in... feel free to scroll through the different chapters...

« Last Edit: August 08, 2014, 09:12:33 PM by rhlauranna »

Offline zenelectro

  • Hero Member
  • *****
  • Posts: 793
  • Liked: 151
Re: ADC/DAC losses
« Reply #14 on: August 08, 2014, 11:40:49 PM »
Yep, have read all the Burwen stuff - a bit of a genius.

I also considered trying the Burwen Bobcat software but have since moved on to other things. http://www.burwenbobcat.com/

I'm not sure how all this relates back to the topic - ADC accuracy and I'm also not sure why you have an obsession with 'squares'.
As stated before, record and playback a square wave into the best analog tape machine and it won't come back a 'square' not even close.

Thats just bandwidth. You can argue that your non oversampling DAC can play squares, but there are no squares encoded on PCM
even at 192k, there is always some bandwidth limit. That is basic electronics.

If you want absolutely the best 'squares' cut to the chase and go DSD128 / 5.6MHz - it has the widest bandwidth and will encode
/ decode the best looking square wave.

Terry

Offline rhlauranna

  • Hero Member
  • *****
  • Posts: 1202
  • Liked: 156
Re: ADC/DAC losses
« Reply #15 on: August 09, 2014, 09:05:02 AM »
nice to read you, and thanks for your comments...

well, it is not my intention to produce redundant information material here (see Tuyen's thread for that matter), but I will gladly answer you of course:

... well, I have to agree, you are right and I understand what you mean, and it might indeed be a little off topic, but from what I read here my impression has grown that the general discussion tended already before from expressis verbis "ADC/DAC losses" to chasing for the "best" sound in general and that the arguments are not always that specifically focused and reduced...

...personally I was glad to realize this because all things in HiFi depend on one another, and I learn a lot, some results are measurable, some are measurable and listenable, and some are only listenable...

all things considered regarding "ADC/DAC losses" we should be able to precisely measure what happens, how much loss is on ADC-side (from what, if not a perfect signal, i.e. square ? i.e. a 50 cycle frquency for example), and how much loss is on the DAC-side (from what, if not a perfect signal, i.e. square... ?)... what do we put in? what do we get out ?

and then we have to face that generally there are quite a lot of losses within the amplification chain, not to talk about the speakers/drivers... and then: what do all these losses in ADC/DAC, in amplification, in speakers have for an influence on the sound ? I would like to see/listen to these results singularly, but this is not possible or is it ? So measurements seem the only way out...

perfect squares are ideal for measurements (put for example 50 cycle from a sine-generator into an ADC/DAC and/or pre-amp and/or amp and/or channel devider and see how it comes out on a scope ) and in reproduction (do the same with 50 cycle and see what happens - and there are your losses)...

by the way, what ever could a physician/technician/developper reach "more" in electronics but a perfect signal, i.e. input equals output ? if there is more than a perfect signal than I would like to know... so, this is at least one of the ultimate parameters...

Why? Because perfect squares represent velocity of the signal (vertical), stability (horizontical) and ultimate precision (no overswings). One simply cannot reach these aims with tubes - I have always "heard" this, and for those who are in doubt Richard S. Burwen now has explained physically why it is exactly like that...

...reaching perfect squares (this took nearly 30 years) in combination with just the right Class A controlled power supply (I just finished the last step for the DDDAC for the 5 and 12 Volt several months ago) in my modest HiFi-experience has dispelled all those things that you guys here are complaining about: the need/urge to colour, the impression of things being not o.k., being aenemic, being inprecise, lacking life, feeling/hearing losses...

sometimes I ask HiFi-fans: what is your aim in HiFi ? Mind you that I get an answer ? Here on the site they have the courage to express it. Bravo !: "I want to have it live!" and they tell what they are struggling for... Yes, that's an aim. And as an ecomomist I have learned how to reach aims. At least I am convinced in having reached this aim in HiFi: perfect squares !

well, the more "right" you get things right from the beginning, the more you get the music that is stored the one or other way... and this means you get the music the way you ever wanted it... no matter at all what music you play and how "loud"...

For more than thirty years, whenever it comes to the discussion of "squares", even the hightest gurus tended and still till this very day tend to "neglect" and/or run down them with some sort of indignancy, presuming/pretending that squares have no or hardly any influence on the sound... I know the discussion very well, but for me squares are the greatest key in HiFi at all... and I am really a lucky man that Doede finally filled this gap with his perfect squares right in the beginning where they were missing so much...

not to become indiscreet, but to illustrate what's happening, this (my) results brought Jean Hiraga to loudly scream out when first listening to these squares resulting in no longer listening to his own system at his home for more than two and a half years now, claiming having made so many faults...

... it is that good for him that he just announced and insisted to bring with him the editor of the magazine that he is writing for coming autumn... and I can confirm you that they then will get the ultimate shock with the controlled 5 Volt and 12 Volt power supply for the DDDAC... this result is even overwhelming for me, the sound is that natural...

...and I had already listeners who "could have died" for that sound...

... if I reproduce for example a "simple" spoken narrative voice, it is that natural that I am surprised and sometimes frightened if there might be someone in my room... that's one of the reasons why I am so "obsessed" with squares, knowing that the end still is not yet reached, oh, what fun...

...so, the problem is to bring the 400 horse powers that one hopefully has, i.e. "perfect" squares (from the DAC) on to the street, i.e. through the amplification chain to the drivers, without squeeking slicks (without any losses)... in order to be able to listen to the pure signal...

to precise: "Perfect" squares guarantee a running through (for example through a pre-amp) without altering the signal... this can easily be proved by playing one and the same source within a system chain with a pre-amp or without... from what I partly read here, the squares seem not to be perfect, otherwise the discussion whether to "use" a pre-amp or not would not become that evident....

the more the squares become perfect and are amplified "right" the less the need to taylor to taste is given... tubes are not precise, they are not able to produce/amplify squares, and with all respect to all the tube afficionados, the supposedly positive effects that different kinds of tubes deliver do not bring it to the point (see the precise physical explanations and theoretical background at Richard S. Burwens site...)

You say: "As stated before, record and playback a square wave into the best analog tape machine and it won't come back a 'square' not even close." But it should come out, or not ?

You say: "You can argue that your non oversampling DAC can play squares, but there are no squares encoded on PCM" And the input signal ? Neither does have squares ?

You say: "even at 192k, there is always some bandwidth limit. That is basic electronics." yes, of course, although there is still room for further improvements, at that level - to me - it may remain as is already... it is like looking at an Ultra High Definition Television, no more complaints, just the opposite, knowing that 8 times Ultra High Definition is not only already in the making but already being distributed in Japan...

You say: "If you want absolutely the best 'squares' cut to the chase and go DSD128 / 5.6MHz - it has the widest bandwidth and will encode/decode the best looking square wave." sure, but I told already, it is unfortunately not only "done" having the highest rates everywhere (i.e. having the most horse powers), but to bring them on the road without losses... we have come more or less close to that but we are still away from it... and second, the music that I like does not exist in that high resolution... and it is my music that I want best way possible....

well, and the DDDACs with all ameliorations would not have been built that way without my ideas/intentions/feelings - even heavily insisting in building "against" physical laws... see Doede's discussion about multiplying chips - he was the greatest resistor in topping more than 60 chips...

anyway, I find the discussion here and there very interesting and I learn a lot from it... although I am neither a technician nor a DIYer, so basically I have no influence on the buildings and constructions, but I can "see" measurement results and I have my ears, and both of these have brought me to the point where I am now...

and I will be glad to further report on the theme...




« Last Edit: August 09, 2014, 09:28:59 AM by rhlauranna »

Offline bhobba

  • Sr. Member
  • ****
  • Posts: 330
  • Liked: 9
Re: ADC/DAC losses
« Reply #16 on: August 09, 2014, 01:23:03 PM »
I believe ADC's (of all varying types) rob the music of "something".

Of course they do.

It simply depends on how sensitive you are to it.

The DEQX guys have a device that takes analogue input, converts it to PCM, does digital processing for room correction etc, then converts it back:
http://www.deqx.com/product-deqxmate-overview.php

Can you imagine what that would do to a Killer Dac?

Yet some go Ga Ga over it - and even get into arguments about it:
http://www.positive-feedback.com/Issue67/deqx_mate.htm
'No Dave, this is how it should sound. I know this cut, CD, album, whatever… and this sounds right. The other three options are messing with the sound. They ain't right.'

Would love someone to get one and see what it does to a Killer DAC.  Bit chock with audio gear I need to sell right now to indulge that whim - but sometime down the road - one never knows.

There is even the possibility those shenanigans wont rob the Killer of its magic - but I doubt it - still one never knows does one.

Thanks
Bill

Offline zenelectro

  • Hero Member
  • *****
  • Posts: 793
  • Liked: 151
Re: ADC/DAC losses
« Reply #17 on: August 09, 2014, 02:48:13 PM »
nice to read you, and thanks for your comments...

You say: "As stated before, record and playback a square wave into the best analog tape machine and it won't come back a 'square' not even close." But it should come out, or not ?


I think it's a good idea that you get even a basic understanding of electronics before launching into technical discussions. You don't appear to
understand really what it is that square waves do and what they are good for checking. How can you participate in such a discussion if you
don't really understand the basic principles of what we are discussing. You are wasting my time and your own just reiterating what someone else
has told you.

Quote

You say: "If you want absolutely the best 'squares' cut to the chase and go DSD128 / 5.6MHz - it has the widest bandwidth and will encode/decode the best looking square wave." sure, but I told already, it is unfortunately not only "done" having the highest rates everywhere (i.e. having the most horse powers), but to bring them on the road without losses... we have come more or less close to that but we are still away from it... and second, the music that I like does not exist in that high resolution... and it is my music that I want best way possible....

well, and the DDDACs with all ameliorations would not have been built that way without my ideas/intentions/feelings - even heavily insisting in building "against" physical laws... see Doede's discussion about multiplying chips - he was the greatest resistor in topping more than 60 chips...

anyway, I find the discussion here and there very interesting and I learn a lot from it... although I am neither a technician nor a DIYer, so basically I have no influence on the buildings and constructions, but I can "see" measurement results and I have my ears, and both of these have brought me to the point where I am now...

and I will be glad to further report on the theme...

I am sure DDDAC sounds great, there has been a lot of work put into it.

The principles are very basic and have been known for many many years. Using many DAC's in parrallel is a very well used technique to lower
noise. A good example is ESS Sabre, in mono mode it is effectively 8 DAC's in parallel. In DDDAC's case one advantage is also probably
just the lower output impedance when more DACs are used. This can also be done with a simple analog circuit.

The principles of zero oversampling DACs are also very well known and the fact that they can reproduce a square wave is one of their attributes.

Happy square wave hunting! :)

cheers

Terry

Offline ozmillsy

  • Hero Member
  • *****
  • Posts: 2167
  • Liked: 245
Re: ADC/DAC losses
« Reply #18 on: August 09, 2014, 05:28:50 PM »
What's interesting is that these 'colours' that bring the 'life' you are after, aren't even able to be depicted as frequency response variations in the outputs, as we have discovered that the output FR from a KillerDAC is in fact basically just as 'flat' as any other DAC out there (albeit bit stronger rolloffs at both ends).     
Absolutely Tuyen, and this point alone is worth a dedicated thread.   There are all kinds of distortions, and alot of them dont translate into a measurable change in FR. 

Quote
My thinking is that there will always be something 'lost' after every (ADC->DAC or vise versa) conversion.   So because of this, one can never expect it to sound exactly the 'same' or as good as the original.   Only just 'good enough'!  :)
What I hear with standard gear is not 'good enough'. 

I dont know that anyone has unlocked the key to truly transparent digitization of music.  And I think there is definitely opportunity to explore this further.

Zen, you say we want it coloured, I dont really accept this (but I understand why you think this).   What I want is to be able to digitise my analog sources, and have them sound *exactly* like they do on replay without a digitization step in the chain.  (Or really really close to it).

The only way to be sure, is to actually compare and listen.
It's all about the music,, not the equipment.

Offline rhlauranna

  • Hero Member
  • *****
  • Posts: 1202
  • Liked: 156
Re: ADC/DAC losses
« Reply #19 on: August 09, 2014, 08:34:03 PM »
I think it's a good idea that you get even a basic understanding of electronics before launching into technical discussions. You don't appear to understand really what it is that square waves do and what they are good for checking. How can you participate in such a discussion if you don't really understand the basic principles of what we are discussing. You are wasting my time and your own just reiterating what someone else has told you.

thank you, it is free for everybody to read or leave out what I write for nonsense... I feel fine here, and I do learn quite a lot, and if you don't like my contributions, just leave them out... that simple...

you might keep in mind - especially in electronics - that there is a proverb in medicine from Paracelsus since the middle ages: "The one who cures, is right"... even against physical laws and knowledge... and that even Doede had to "learn" during our researches... why ? because the sound got better - against initial presumption...

...as I do know nothing about electronics - you are right -  I always try to quote professionels: I quoted Richard S. Burwen who owns several dozen patents in electronics worldwide, and others verbally, who invented Class A and more...

...and as you seem to not want to read my ignorant and unscientific descriptions regarding squares you might prefer the ones from professionals...

http://www.dddac.com/dddac1794_test_specs.html

and by the way, there is further no need to theoretically argue/presume/confirm anything regarding DDDAC if you actually have not yet listened to it and/or compared it and/or specifically tested it out...  then please don't waist my time...

...what I learned in HiFi: it is always the result that counts, not all theory...

...my aim is: to make/get things better, everywhere... the one or other way... that's why I am here for, and to learn from you all...

...and yes, thank you for your: "Happy square wave hunting! :)"... after 50 years in chasing for the best in HiFi, I can confirm you: I finally have them... what an infinite luck, exactly this is the fulfillment of my dreams.... !!!!!!!!!!! it must not be yours...
« Last Edit: August 09, 2014, 09:55:05 PM by rhlauranna »