R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: 1 ... 15 16 [17] 18 19 ... 23   Go Down

Author Topic: George....what's the resolution of analog?  (Read 99467 times)

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: George....what's the resolution of analog?
« Reply #240 on: June 30, 2004, 03:22:35 pm »

I find this thread very amusing. It is a good example for how comments on a subject drift into and out of focus. When I saw the question about “resolution of analog” my first knee jerk reaction was – resolution is not a terminology for analog. It is terminology for digital. Of course I am viewing things from engineering science and math standpoint. I now see that an ear person may use such term for what the ear can resolve, and indeed it is a valid way to look at it.

Indeed, the way I would approach the question is from an EE standpoint, and I would start with “what is the resolution of digital”. An EE would immediately focus on the number of bits, which sets the dynamic range (noise floor) with about 6dB per bit (thus 16 bits is 6*16 = 96dB, 20 bits is 20*6 = 120dB dynamic range). Clearly, this is a starting point that assumes that all the bits are acting as desired (good bits). The reality is that non of the 24 bits AD and DA can yield 6*24=144dB dynamic range. One may view the lower bits are “fuzzy”, mis behaving, at best randomly bouncing up and down, at times also moving about with some influence from the signal itself… As far as I am concerned, a bit that “moves around a lot” relative to it’s weight is not following the music, thus is useless, and is a “marketing bit” 

There is always some constant noise out there with the signal, be it an analog system or digital one. The lower bits in an attempted design of say 24 bit AD or DA are “overwhelmed” by the analog noise, reducing the dynamic range to less than 24 bits. On the other hand, a real 20 bits performer (converter) contains so little noise relative to say a 16 bit system, that the noise figure – thus the dynamic range of the digital system may be dictated by 6*16 = 96dB without any limitations due to the analog.

So in a sense, I would tie the answer to the question regarding the resolution of analog to what it would take to accommodate some specified digital criteria. What is the resolution of analog? It is tied to “what is the dynamic range of analog” or “what is the noise of analog”. The answer always depends on the specific gear.  Figuring the noise is just engineering stuff. The noise is there because electronic parts generate noise, and because unwanted signals couple into the signal path. Even a simple resistors, just a piece of say carbon, has electrons bouncing inside randomly and generating noise. Of course, with low resistance (such as metal wire or good conductor), such electron motion (current) can not build much voltage. But with high resistance you get surprising amount of noise, certainly plenty enough to make 24 real bits a non real proposition in this day and age… And of course, semiconductors is a whole other level of complexity… and so no…

So this kind of analog noise is what limits dynamic range. Some of that noise appears with the signal. Some of this noise limits the dynamic performance of what the gear is capable of, be it analog or digital.

A lot of what I said is a bit simplistic. I can not cover everything in a massage. But this is basically an engineering approach (EE type, not a recording and mastering type).

Every real digital bit provides real value in terms of dynamic range as well as lower distortion (certainly in terms of low level signals by themselves or riding on a high level signal) is a good thing. Lower noise is also good thing, and in that sense, and even “meets” the “ear type definition of resolution” – hearing small signals by themselves or when other music is played. I can appreciate the ear type people that that wish to include frequency issues with the word resolution. After all, resolution is an everyday word, not just to be used by audio converter designers. For example, video guys have their own meaning attached to it.

But I am sorry to see so many people misled by the notion of increased sample rate as contributing anything to resolution. The marketing guys, and some ear types with less technical savvy seem to equate high resolution audio with increased bits and sample rate. The fact is – it is about bits, not about sample rate. More real bit increase dynamic range. Faster sampling beyond a certain point does nothing. But that is all covered in the 192KHz thread…    


BR
Dan Lavry
Lavry Engineering
Logged

Johnny B

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1134
Re: George....what's the resolution of analog?
« Reply #241 on: July 01, 2004, 03:02:29 am »

Dan,

Are you saying that real improvements could be made by going to 32-bit converter chips?  

Sure some of those 32 bits could be wasted, but those extra bits could come in handy, at some point down the road, ya never know... LOL.

Logged
"As far as the laws of mathematics refer to reality,
they are not certain; as far as they are certain,
they do not refer to reality."
---Albert Einstein---

I'm also uncertain about everything.

Level

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1811
Re: George....what's the resolution of analog?
« Reply #242 on: July 01, 2004, 03:38:24 am »

I like dynamic range and more intervals between dynamics=more fluid, analoguish. We have the computing power, lets use it!

If you don't like it, decide on what you DO like with your overall sensations of what is 'best' for you and your client.
Logged
http://balancedmastering.com

"Listen and Learn"
---Since 1975---

steve parker

  • Full Member
  • ***
  • Offline Offline
  • Posts: 134
Re: George....what's the resolution of analog?
« Reply #243 on: July 01, 2004, 08:48:38 am »

"more intervals between dynamics=more fluid, analoguish"

but there aren't more intervals between dynamics im any meaningful sense.
extra bits just make the rounding error half of the next bit back leading to this rounding errors noise being lower.

steve parker.
Logged

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: George....what's the resolution of analog?
« Reply #244 on: July 01, 2004, 10:29:10 am »

Johnny B wrote on Thu, 01 July 2004 08:02

Dan,

Are you saying that real improvements could be made by going to 32-bit converter chips?  


Believe me, he's not saying that.

Nika.
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: George....what's the resolution of analog?
« Reply #245 on: July 01, 2004, 02:28:22 pm »

Johnny B wrote on Thu, 01 July 2004 08:02

Dan,
Are you saying that real improvements could be made by going to 32-bit converter chips?  
Sure some of those 32 bits could be wasted, but those extra bits could come in handy, at some point down the road, ya never know... LOL.




One must keep in mind that there are 2 separate issues here at play:
1. The number of bit you can get out of a real wold AD or DA, which has a lot to do with analog noise. Lets face it, we do not have real 24 bits performance gear, nor do we have 22 bits. Face it - the best mic pre with a shorted input makes enough noise to "walk all over" a real 24 bit AD performance (if you could have one). 24 bits would require 144dB signal to noise ratio, at the microphone. You just can not have it yet.

2. The number of bits used for digital signal processing (DSP). This is another issue all together.

Of course the initial reaction is: why 32 bits DSP if the signal is already below 24 bits? Why will we not have the analog noise "walk all over" the very low quantization noise of the DSP?

Of course, if all you wanted is say gain change - a single multiplication of the data by a constant, than a 24 bit DSP machine is fine (in fact, with a lot of margin). The problems appears when you do hundreds or more computations. One such problem - each (single) computation makes for a tiny error. You keep doing more and more computations and the errors "pile up" and you get a combined error that is too large. Another such problem is the fact that many computations are a sum of many tiny elements that each one is very tiny (requires a lot of bits). You can later truncate the end result to say 24 bits, but first need to keep the individual elements to high precision.

This is a whole subject by itself. The point is - for DSP you do need to be ready for a lot of "extra" bits, depending on what you specifically need to do.

But 32 bits for conversion? Forget it! Say I get 1 volt out of a mic (that is a lot of signal!) for 24 bits, I need to have about 63 nano volts of noise (over the audio - I am assuming 22KHz bandwidth) the noise generated by a 2.5 KOhm resistor. So you are asking a mic to provide 1 volt into 2.5 KOhms - .4 mWatt free power and the limiting factor will be the resistor and nothing else. Try for 32 bits (192dB dynamic range): The noise requirement is now 0.25 nano volts, which would be generated by a single 4 Ohms resistor! So yes, make me a mic that can supply say .4 milliwatt of noise free signal into a 4 Ohms resistor, than have a mic pre amplifier with no noise at all and we are there Smile


BR
Dan Lavry  
Lavry Engineering

Logged

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: George....what's the resolution of analog?
« Reply #246 on: July 01, 2004, 02:35:40 pm »

danlavry wrote on Thu, 01 July 2004 19:28

Johnny B wrote on Thu, 01 July 2004 08:02

Dan,
Are you saying that real improvements could be made by going to 32-bit converter chips?  
Sure some of those 32 bits could be wasted, but those extra bits could come in handy, at some point down the road, ya never know... LOL.



Correction:

I said (for 32 bits)

So yes, make me a mic that can supply say .4 milliwatt of noise free signal into a 4 Ohms resistor, than have a mic pre amplifier with no noise at all and we are there Smile

I Should have said "make me a mic that can supply say 1/4 watt (1 volt into 4 Ohms) of noise free signal into a 4 Ohms resistor... That 32 bit was carzy enough at any power level, but .4 watt was the wrong number so I am correcting it. Sorry, i did it too fast thus the error.

BR
Dan Lavry  
Lavry Engineering



Logged

lucey

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1043
Re: George....what's the resolution of analog?
« Reply #247 on: July 01, 2004, 03:17:24 pm »

Nika Aldrich wrote on Wed, 30 June 2004 09:37

lucey wrote on Wed, 30 June 2004 06:04



Steve, I admit to a limited understanding of the subtle technicalities of digital recording.  I do know tone however, and specifically I know vibrations.


I am not surprised that you hear differences between analog and digital systems.  I've heard your music and work you've mixed, and my personal observation is that you tend toward very rich sounding analog material and equipment.


Anyone I know can hear the difference Nika, what are you implying?  

Yes I do produce music by recording analog instruments on analog gear, up until the final mastering step if possible.

(as for my work you've heard one or two older mixes that were poor by current standards.  you've heard no recordings made in the last 3 years, nor any mastering work)



Quote:


Quote:

To my ear any AD DA (even my amazing Pacific Microsonics) is fundamentally altering the waveform's quality.


Perhaps the PMi DOES alter the waveform's quality.  How have you tested it and what other converters have you tested - and how?

Nika


a) Everything alters the waveform, analog alters it quantitatively, digital alters the quality.   The Pacific is the most beautiful and accurate I've heard (have not heard Prism or Lavry, and dont need to a this point)

b) by listening
Logged
Brian Lucey
Magic Garden Mastering

"the economy is a wholly owned subsidiary of the ecology" - unknown

Nika Aldrich

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 832
Re: George....what's the resolution of analog?
« Reply #248 on: July 01, 2004, 03:47:30 pm »

lucey wrote on Thu, 01 July 2004 20:17

what are you implying?  

(as for my work you've heard one or two older mixes that were poor by current standards.  you've heard no recordings made in the last 3 years, nor any mastering work)


Brian,

Just to be clear, I was not implying anything at all.  I have heard your work and I noticed that the artistic flavor you work toward is a very "analog-y" sound.  The tone was very rich and full, saturated at times, thick, sometimes heavy, big and round sounding.  I am drawn to the conclusion that this is your personal taste and that you aspire toward this sound.  I am led to this conclusion because I know that you have a lot of equipment and that you are very particular about your investments in your equipment.  I also know some of your equipment and know that you can get various colors out of it depending on how you use it.  Since you know this equipment well I am led to the conclusion that you intentionally color the music a particular way - not because you have to - but because that is how you like it.

Let me repeat that - I respect your ears and your abilities enough that I draw the conclusion that you are pleased with your results and that you like the particular coloration that you get in your recording/mixing process.  Ergo, my earlier statement was to draw no valuation on the *quality* of your mixes and was only to explain that, based on what I have heard, I would assume you would prefer analog equipment because it is easier to get the sound you like from that equipment.  That sound is "more available" with analog equipment at this time.  

There is no need to defend your work, and I doubt there is much purpose in me hearing your other work.  I have heard enough of your work to understand what it is that you like stuff to sound like that I have to assume your recent work does not contrast that much in terms of the overall color.  I don't think my conclusion would be any different today.  Again, no valuation of the quality of your work one way or the other - just observation.  Believe me, I don't play the "yeah, well I've heard your work and you don't have any grounds to talk" game.

Cheers,
Nika.
Logged
"Digital Audio Explained" now available on sale.

Click above for sample chapter, table of contents, and more.

Scovi

  • Newbie
  • *
  • Offline Offline
  • Posts: 1
Re: George....what's the resolution of analog?
« Reply #249 on: July 01, 2004, 07:39:46 pm »

Oooo ... very juciy topic here folks. Thanks for the great read. My feeling is there are two questions/topics being addressed here. i.e "Which sounds better" and "which is more accurate"? I aggree with the concept previously presented of "which format is introducing more error?" Surely analog in the grand scheme of things is introducing more "error" if comparing input to output in the given system. But obviously there are some aritfacts of "analog errors" that we as listeners are either accustomed to, or just down right enjoy. The same does not seem to hold true for digital, in that we don't seem to have the same love for "digital errors". In fact, we seem to hold it to a higher standard ... and that standard is "it has to sound at least as good as analog" Anybody feel like we are there yet???
hmmm ... feel like I could ramble on for hours about this ... I'll spare you.

Scovill out
Logged

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: George....what's the resolution of analog?
« Reply #250 on: July 02, 2004, 09:40:11 pm »

[quote title=lucey wrote on Thu, 01 July 2004 20:17][quote

You said
"Everything alters the waveform, analog alters it quantitatively, digital alters the quality.   The Pacific is the most beautiful and accurate I've heard (have not heard Prism or Lavry, and dont need to a this point)"

I say:
Hi, I am Dan Lavry and I wish you listened to a Lavry Smile The fact that you hear “even” the Pacific M you like is not surprising to me. I know their AD is a module made by Analogic, and we (I was employed by Silicon General) used to compete with them in non audio applications with the same sub-ranging architecture. It is a good approach, though somewhat noisy, and will have some unique color. When it comes to their DA’s, I know all about it. I designed that segment DA many years ago at Silicon General (the division was Analog Solutions). Later, a few folks started Ultra Analog and “took” that design and repackaged it. It eventually ended up at various places such as Mark Levinson gear that buys the DA modules. Years later, I reviewed the shortcomings of that approach (such as the weakness of PCM around digital black, the lack of deglitcher circuitry, the variations of components over temperature, component ageing and more), and I redesigned a much more segmented DA with far superior jitter rejection and a few other totally proprietary concepts.

Yet, how transparent is my best DA? How does it compare to analog? That is for the listener to decide. The first point to make is that there is analog and there is analog. Some is harsh, some is warm…. But I do appreciate your approach, to go as far as you can in analog. At some point, digital has some real advantages too – it does not scratch like vinyl, it does not demagnetizes or stretches, it is easy to copy and mass produce…. Other than that, I too love analog. We just can not ignore digital, and need to keep making it better. We have come a long way, in my opinion. I agree that digital has some unique problems and issues that do not come with analog.

BR
Dan Lavry  


Logged

lucey

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1043
Re: George....what's the resolution of analog?
« Reply #251 on: July 03, 2004, 10:34:59 am »

danlavry wrote on Fri, 02 July 2004 20:40


I say:
Hi, I am Dan Lavry and I wish you listened to a Lavry Smile



Well, the price of this was right for me, and I love it, so ...

As for tone.  The Model One was in the shop before i bought it from Trevor, and they do upgrade both analog and software, and totally reset the specs any time they see a box for service... so that's cool.

Yet assuming your process improved upon the AD and DA of the Pacific, which makes sense, as time has passed, what about the analog path itself?

The Model Two was discontinued due to parts supplies on the Class A stuff they needed drying up.

Are your boxes chip laden?  If so, with ICs?  And how do the two levels vary in analog and converter path?
Logged
Brian Lucey
Magic Garden Mastering

"the economy is a wholly owned subsidiary of the ecology" - unknown

Johnny B

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1134
Re: George....what's the resolution of analog?
« Reply #252 on: July 03, 2004, 05:47:08 pm »

Dan,

You used some of the following words:

truncation
conversion errors
not a full 24 bits
errors piling up

Or expressed those ideas for problems, right?

So what do you need at the chip level
for it to be done the right way?

In other words, couldn't the chips be made better so these errors don't exist and would not that be better? I dunno?




Logged
"As far as the laws of mathematics refer to reality,
they are not certain; as far as they are certain,
they do not refer to reality."
---Albert Einstein---

I'm also uncertain about everything.

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: George....what's the resolution of analog?
« Reply #253 on: July 03, 2004, 09:05:25 pm »

You said:
"Dan,
You used some of the following words:
truncation
conversion errors
not a full 24 bits
errors piling up
Or expressed those ideas for problems, right?
So what do you need at the chip level
for it to be done the right way?
In other words, couldn't the chips be made better so these errors don't exist and would not that be better? I dunno?"

Of course the chips can be done better, and the idea is to know where the next thing is to improve. It is a very big question!

Some of it has to do with the standard itself. The limitations of 16 bits (red book) call for processes such as dither to avoid truncations and noise modulation, but when you go for say 24 bits (DVD Audio), you can forget about such dither. With 441.KHz (CD) you are better off to noise shape the dither, but with say 88.2KH or 96KHz, you have a huge range where to “park” the truncation problems. So your noise shaping problem become an easy one to accomplish.

Some of it has to do with processes and resources. Say you need a hack of a lot of digital computations. It is more likely to be able to do a good job when you take an IC that is “specialized” at doing huge amount of processing and nothing but that, and put it together with another IC that does say just the DA function…

You see, the IC makers often try to cram a whole bunch of stuff into a single IC solution. The whole thing must be limited in space to something the size of a pin head, and it is running “boiling hot” at say 2 watts or so. Than someone “glues it” into a board that fits on a 19 inch rack… Well, yes, there are some support IC’s and often a display and some large connectors… But fundamentally, there is the question at hand: why cram a whole function into an single IC?

Much of the answer is about mass production and selling a complete cost effective solution. I am often amazed at who well the IC makers do things, given their constraints. But is it the best way to go? For mass production and cost effectiveness, the answer seems to be yes. For highest quality? Probably not so.

You see, The IC maker must produce solutions WHILE BEING CONSTRAINED to their own technology. True, their technology offers a lot of advantageous, but also there are a lot of things they simply can not do inside the IC. Anyone familiar with the game of Chess knows that there are only a few basic pieces (King, Queen, Bishop, Knight… Pawn). In electronics we have resistors, capacitors, indictors, semiconductors… I (an equipment designer) can use a whole range of resistor, capacitor and inductor values, and I can also choose them according to properties (the material they are made of and so on).  They (the IC designer) are playing the game with a lot less flexibility. They do not have ANY inductors inside the IC. They have almost no space for capacitors inside, only for very tiny value caps, of certain limitd material. For the most part, they often can not handle high voltage or current or much power and so on. Some of their world does offer advantageous the equipment designer is longing for (such as – all parts are so nearby, thus at almost the same temperature, thus they track well, laser trimming for great matching of parts, and so on).

We all end up knowing that the best design is a hybrid of IC and non IC technology. Going for quality, say a design calls for some poly capacitor. The smart IC designer will provide a pin or two to connect an external poly cap (The Cap may be much bigger than the whole IC) .The best IC technology for say a sigma delta DA may be best done with some semiconductor process, 5V supply and so on, that is NOT best for some other function. So you have a different IC process for an OPamp, say for a +/-15V supply, and so on…

To complicate things, all of it, inside and outside the IC, is based on theory, understanding of networks, circuits, signals, math and more. These days, I believe you will get the best quality when the IC guys do what they do best, and stop short of trying to go beyond their limitations. At that point, the equipment designer takes over, and with freedom from the limitations imposed on the IC designers and chip makers, goes for a complete quality product, whatever it takes.

Obviously, this is not what takes place when the main goal is mass production and low cost solutions.  So I do not expect the IC maker to lead us in the direction of highest quality. Their bottom line is driven by large quantity production! It costs a huge amount to make a new IC, and they need the volume. I am not sure we should wait for the IC makers to provide all the leadership. I was sad to see them sell 192KHz audio sampling AD and DA. They know it is BS from science and engineering standpoint. Often it is the little guy (like I) that has to fight those king kong size companies to send them in the right direction. I made a lot of progress, man does it take a lot of non paid for work Sad

BR
Dan Lavry
Logged

Johnny B

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1134
Re: George....what's the resolution of analog?
« Reply #254 on: July 03, 2004, 10:42:40 pm »

Dan.

Thanks for the decription of the problems and tradeoffs the IC makers and people such as yourself face. Maybe I did not ask the question well.

OK, you know what the problems are, would the ideal chip have more bits, if so, how many? How would you get rid of some of the problems, besides making it easier for a design artist such as yourself to slectively hang stuff off the chip?

Do we really need or desire a full 24 bits? What would happen if you had that? Would you still be in business, would your life be a little easier from the design aspect?

 
Logged
"As far as the laws of mathematics refer to reality,
they are not certain; as far as they are certain,
they do not refer to reality."
---Albert Einstein---

I'm also uncertain about everything.
Pages: 1 ... 15 16 [17] 18 19 ... 23   Go Up