Why digital is superior to analog

[mshilarious]

Here's my invention du jour: take that VHS as full-width analog tape idea and add a 24/48 digital track to the mix. Shouldn't take too much of the tape width for that. Add a sync to the analog signal and have the digital signal lead a bit so it can be buffered and then sync to the analog. Compress the digital rather strongly, to about peak -12dBFS. On playback, downward compress the analog signal and mix the two. Effectively, you have parallel compression with the analog signal sitting on top but the digital signal dominating the quiet parts for dynamic range that equals or maybe exceeds CD. And whatever exists of the analog signal above 24kHz is still there to fill in that gap. Obviously this is a playback technology and not production.

Odd thing is because you don't have to build a prototype anymore to patent something I could patent that (until somebody challenged it, but at least my compansion algo would probably be defensible). But that's an awful lot of trouble, so I'm putting that one in the public domain as of right now!

The A/D debate now resolved, you're all banned!

 

[SouthSIDE Glen]

There is a paradigm shift that's going on right now, and to me it is for the better. More and more people are actively searching for stuff that THEY like, instead of waiting to be fed through radio and MTV which have all but become jokes.

By and large IMIE (In My Internet Experience) you see white kids into basic metal or "classic" rock that doesn't get much beyond the Beatles and Floyd, and black kids into vanilla hip hop and dance mixes. Why? In large part because that's all you'll find on radio or in he clubs in most metro areas and college towns.

There's diversity available out there on the Internet, to be sure. But that diversity is in availability more than in usage. By that I mean that those that do bother to search out beyond the Soma gas of Clear Channel, tend to search out specific vertical silos rather than explore the available diversity. I swear, in the past few years here, I have seen more over-categroization and over-specialization in the types of music that many readers play or are "into" than I have seen in my entire previous life. People tend to search what they know, not what they don't know, and reinforce what they already like rather than try something new. In a similar vein, you'll never catch a Democrat actually listening to what Rush Limburger has to say any more than you'll catch a Republican considering Al Frankenstein's opines. There's a whole lot of preaching to the choir going on around the Internet.

For the guy or gal who's into *music* and not just this or that super-sub-genre played only in Dm on alternate Tuesdays, there's still a gold mine of talent out there and plenty to listen to. But very few are doing that; partly because of laziness, partly of comfort, but also largely because they have been led to believe that's the only thing worth listening to because the rest of the CC Soma gas pumped out on radio and TV is garbage.

G.

 

[SouthSIDE Glen]

Here's my invention du jour: take that VHS as full-width analog tape idea and add a 24/48 digital track to the mix.

Lets have a virtual format war, Jon! I'll counter that with an analog laserdisc .

G.

 

[Dogman]

It's worse than that. We can't even unregister from homerecording.com until we settle the digital/analog issue once and for all. It's in the fine print in terms of use. That's the only reason I'm still here.

You have been assimilated.

 

[mshilarious]

Lets have a virtual format war, Jon! I'll counter that with an analog laserdisc .

G.

You're on! My experience with optoelectronics tells me your noise performance won't be very good though!

OK, let's each do a simulation of our virtual formats and post .mp3s and let the public decide which is better! Low bitrate preferred, of course

 

[mshilarious]

not just this or that super-sub-genre played only in Dm on alternate Tuesdays

Dm is the saddest of keys

 

[mshilarious]

I find that most of the really young kids are into atonal piano singer/songwriter stuff:

Lieder op. 472

 

[miroslav]

Just because YOU don't use the instrument (yes, the instrument) to it's full capability it has nothing to do with any deficiency in the instrument.

Not sure what you are going on about here..?

There was no mention of any "deficiency" (I even said it was DA SHIT back in the day)...rather there was just the comment that it's not a very hot item these days...

 

[SouthSIDE Glen]

You're on! My experience with optoelectronics tells me your noise performance won't be very good though!

I'm no optical engineer, and I'll probably prove it with this post , but...I'd think that is simply a technology issue which could be addressed or solved a number of different ways.

Heck, if an optical varistor is good enough for an LA2A, an optical reader for source data shouldn't be too hard to engineer.

BTW, while we're at it, we need to end the binary revolution as we know it and either bring back analog computing, or at least move from bi-state to tri-state digital computing. Binary is soooo archaic.

G.

 

[mshilarious]

BTW, while we're at it, we need to end the binary revolution as we know it and either bring back analog computing, or at least move from bi-state to tri-state digital computing. Binary is soooo archaic.

Computing is just math, not sure why base 3 is better than base 2. But I suppose if transistor count does eventually become physically limited then a high/mid/low system is a potential solution to the problem. Lots of electrical challenges there though; CPU cores are already very low voltage and there is the MOSFET threshold voltage to overcome with enough margin to reliably distinguish tri-state. That would seem to give up enough in power dissipation to make the benefits questionable.

But I'm not really an EE, I just play one on TV

 

[miroslav]

I agree with you on this. But what does growing up with tape and analog gear have anything to do with it? Growing up with tape and analog gear should not and is not a prerequisite for someone to be a good musician or a recording engineer in the past, now or in the future. Of course having knowledge of these systems is always good as it expands the horisons of individuals. More knowledge never hurt anyone and I am all for knowledge. And hopefully this will be taught in music engineering courses as part of "music technology of the past" history curriculum.

Not sure what you are trying to say here...you agree with me…but then pretty much disagree?

Growing up with tape and analog gear when equipment options were very limited for the "home recording" crowd, and the term "project studio" had yet to be coined...it forced me to work in a way that required better production planning (from start to finish) since there were few options for "fixing in the mix" with unlimited editing, comping, processing like you get with digital.
Working in that limited, very structured and "plan-ahead" environment forces one to consider many things before hitting RECORD...and THAT leads to the building of a much more solid foundation than what most kids get these days by dropping some DAW app on their home PC and assuming the "power" of computers and software will "fill in" their missing knowledge and lack of any foundation....IMO.

Yeah...there are all these "recording schools" now days (none existed back in my younger days) where you CAN pick up a lot of that foundation, though how many newbies actually go to those schools to learn recording techniques, and how many just dive deep into editing/processing/mixing/mastering with NO foundation at all?

And just having "creative/artistic" talent is NO guarantee that diving into a complex DAW will somehow be any easier or yield better results.
There are a lot of great musicians that are all thumbs in the studio environment (analog or digital).

So yeah...IMO...working back in the early days of "home recording" with tape and limited analog gear was a BIG + for *learning*, and I’m sure it was that same way for many other folks here that got that experience, and it is most of those people who seem to now have a better handle on working with digital…that is, if they ever made the transition to digital!

Sure…someone can just “dive” into a DAW for the very first time with NO previous recording background and still churn out some quality product…but on the whole, judging by the many YouTube, MySpace, SoundClick offerings I’ve heard over the last few years, the majority of home/project studio enthusiasts would IMO benefit a LOT from some limited, “guerilla” recording experiences.
Get a 4-track, a couple of mics, one spring ‘verb or an old Echoplex, and a basic little mixer with a couple of bands of EQ…now go record and mix some music.

 

[SouthSIDE Glen]

Computing is just math, not sure why base 3 is better than base 2.

A geometric increase in information density and speed. Each bit can now carry three values instead of two. A three-bit word in binary can represent one of eight different values; a three bit word in trinary can represent one of 27 different values, already more than a tripling in information density, and that's just in three bits. In 8 bits, binary can represent 256 values; an 8-bit trinary word, OTOH, 6561 values - more than 25 times the density of binary. And so on.

Ideally we could push that all the way to the base 10 we're used to learning on our fingers, where an 8-bit word can represent one of 100,000,000 values.

And of course, then the question is, why not just go "all the way" back to the "old-fashioned" analog computing. Analog computing was first pretty much abandoned for digital when noise was indeed a major issue, but technology has changed a lot since WWII. If you could get to quantum resolution, you could have virtually infinite data density per bit.

And again, the physical engineering is just that; it's physical engineering. It's not like for trinary one would have to invent a new room-temperature superconductor or a quantum Heisenberg compensator; just the ability to manipulate tri-state voltages for starters. And I say "just" like I know how to do it myself: I don't . I am not a materials EE either, but I wouldn't think it'd be a large leap to accomplish tri-state low voltage semiconductor technology.

Then again, if that were true, then we'd probably be doing it already; .

G.

 

[mshilarious]

And again, the physical engineering is just that; it's physical engineering. It's not like for trinary one would have to invent a new room-temperature superconductor or a quantum Heisenberg compensator; just the ability to manipulate tri-state voltages for starters. And I say "just" like I know how to do it myself: I don't . I am not a materials EE either, but I wouldn't think it'd be a large leap to accomplish tri-state low voltage semiconductor technology

I don't think it's a matter of materials engineering as simple efficiency and also the speed of the transistors in question. With binary you can have trigger voltages set roughly at the midpoint (ish) of the voltage, and rise/fall time is equivalent. At tristate it will take longer to go from high to low than high to mid. Therefore your speed can't be higher than the high/low time, but during that interval you have a transistor that is conducting and thus dissipating heat--more heat than bistate because you have to increase the voltage to achieve the required resolution. I'd guess that is also the problem with higher base math; you are trying to hit voltage thresholds at a very fast rate so you have to be able to cope with rise time and overshoot, etc.

Those are just my wild guesses, I just presume there is some reason tristate isn't used. Since OP seems long gone, what else is there to do?

 

[SouthSIDE Glen]

At tristate it will take longer to go from high to low than high to mid. Therefore your speed can't be higher than the high/low time, but during that interval you have a transistor that is conducting and thus dissipating heat--more heat than bistate because you have to increase the voltage to achieve the required resolution.

It sounds like that's assuming that the interval between states remains the same; for example if the difference between a binary 0 and 1 is one volt (to keep it simple), the assumption is that for a tri-state, the interval would remain one volt and that the difference between 0 and 2 is two volts.

But when Intel and the rest move to the next level of integration where the difference from 0 to 1 is 0.3V, we now can potentially have a tri-state system where the total interval is no more than the previous integration level of binary. But since the density of information is geometrically increased, there is a large net gain in efficiency, perhaps even over the next level of integration. Or so it seems to me.

Large roadblocks in a non-material way to a trinary system is that it is a whole 'nother logical animal from binary. Simple AND or OR gates or flip-flops or other building blocks of binary logic are no longer sufficient for efficient logical operations. Trinary logic is more complex, but not super-greatly. Still, it would require a whole new discipline and set of design algorithms to build a CPU based upon trinary "WHICH" gates (or whatever they may be called). Sure you could build them out of simpler binary gates, but that increases the circuit complexity to the point where it would probably offset any logical gains in efficiency. In other words, it's not just a matter of being able to "handle" tri-state voltages, but to design and build single-stage trinary logic circuits.

That also means a whole new vocabulary for the machine language for programming trinary computers, which also involves a new programming mindset. This also means a whole new OS kernal over the top of it all. All of which would be backwards IN-compatible with any existing binary-based software or hardware. So we're probably talking a major re-tooling of the entire industry in many ways, actually. That alone is quite an impediment, I should think .

But really, my personal theory is that because the word "bit" is really shorthand for binary digit, that if we moved to a trinary system, we'd have to call them "tits", and the powers that be just can't take the idea seriously past that point .

G.

 

[lumbago]

Here's the Massenburg presentation, for those interested, from a couple of days ago

https://www.youtube.com/watch?v=wzJbjHc6bRE

 

[miroslav]

He makes a lot of valid comments.

At least it clarifies that he isn't against digital...he is against crappy digital - aka MP3s.

 

[noisewreck]

This thread is making me more and more cranky the more I read some of the responses. Yet for some odd reason, I can't get away from it.

SOMEONE HELP ME!!!!!!!!!!

Fuck it. I am gonna go listen to some Merzbow just to clear my head and wash my ears.

Oh, BTW I hate Bruce Springsteen. His "music" just irritates me.
Same with Peter Gabriel.
Can't stand most of The Beatles.
I find Black Sabbath laughable, and Led Zeppelin self-indulgent.
Snoop Dogg, P-diddy, et al... can't believe they make millions on that nonsense.
Pop... is that music?

In fact, I hate 90% of the music that has topped the charts of any era, and/or has gone gold/platinum/whatever.

Most american (pop, and by that I mean what is and has been on radio, MTV, VH1, etc) music sucks, and pretty much all of russian music reeks of mothballs AND sucks.

BTW, after Merzbow, maybe I'll listen to some Fennez.

 

[Tim Gillett]

Here's the Massenburg presentation, for those interested, from a couple of days ago

https://www.youtube.com/watch?v=wzJbjHc6bRE

I dont know about you others but listening to a George Massenburg lecture on the subject of quality sound where his voice track has massive clipping, plus breath popping all the way through just doesnt seem quite right.

A better production would have included some lecture theatre ambience too rather than just dead silence in between his breaths. Sounds like he's talking to himself in a booth.

And it's not just the headworn mic either. These days you can get good vocal quality from headworn mics. Just listen to a good sports commentator call.

Something about theory and practice not coming together here. Nothing to do with analog or digital either.

Tim

 

[tc4b]

I think your thesis would be better if it was more focused. Instead of trying to prove digital is "better," a mushy, nebulous, subjective term, try to nail down something specific. Such as:

- the lower cost of digital recording and it's impact on modern music
- digital recording and the label-less recording artist
- the effect of the digital recording revolution on the profession of studio engineers

I use digital because it fits my lifestyle, not sayin it's better. One thing I like about it: infinite # of 'undo's' on a given track without any sound degradation. This is a plus for musicians who don't have a lot of time to practice. For a real pro, though, how many undo's would you need?

 

[MCI2424]

Mostly I'm curious if he just cut and pasted my on-line musings and took credit for it like I find a lot of other people have. I would just think great minds think alike, but they don't even put it in their own words... it's mine verbatim. Very annoying.

Oh brother.................