Home Recording's Dirty Little Secret

[Beck]

I'm not being hysteric about anything and I'm not attacking anyone so please don't fire up the douchebag brigade.

What people were doing with certain technologies 20 or 30 years ago is almost irrelevant to today. Not from a theory standpoint but from a practical standpoint. Tom Shultz was also a very talented electrical engineer and part of the point of home recording is that you don't have to have a master's degree just to keep the equipment running. For people who want or need to spend more time recording then dealing with equipment issues, an old analog multi-track is not a very good format.

I'm not even saying that high quality analog doesn't sound better then digital. I'm just saying its impractical for most budget situations and posting the opinions of guys who work with whatever equipment they want is irrelevant and condescending.

Ok, but it seems you're talking about analog like it's all one basket. 20-30 years ago is not what everyone is using, though some do prefer and seek out older classic gear. Analog is a mature but modern technology, the development of which peaked in the 90’s not the 70’s. Manufacturers were still making analog decks in the late 90’s, and at least one past 2k, and they (Otari) are still making some models on a special order basis. There are also companies like ATR that specialize in refurbishing classic Ampex machines.

Quantegy introduced a brand new tape formulation, GP9 in 1998 right in the mist of the digital frenzy, and ATR Magnetics introduced a new tape formulation just last year. The old BASF/EMTEC company was reestablished as RMGI in the Netherlands in the last couple years. There are analog decks in most of the larger more prestigious studios, such as Abbey Road, and also in lesser-known smaller studios around the world.

What I find very often in these forums is that a thriving analog subculture is hidden in plain sight, because it’s not on most people’s radar. And that’s because it’s not in the Sweetwater catalog or at Musicians friend or in pro-sumer mags like Electronic Musician, or at your local music store. Most people in home recording circles are simply lacking information no matter what their age.

Some older recording enthusiasts even look at analog from the perspective of their distant past because they went with digital early on and basically followed the evolution of the magazines they read. Well, analog was there the whole time and it even improved since they last used it.

When I say analog you may picture a dusty old clunker sitting in someone’s basement, but that’s not what I have.

Analog isn’t for everyone, but either is recording in general… some people aren’t cut out for it. But if they are reasonably intelligent, people can learn to use analog equipment as well as a DAW.

And in case someone missed it, I use both analog and digital… I have more in my bag of audio tricks, not less.

Not irrelevant and condescending, but informative. I already addressed the groupthink and bunker mentality that stifles a true free exchange of ideas in these forums. Some people have never had the opportunity to see or discuss the things I’ve brought to the thread, though some have been there and done that and just see a controversy. They have no problem with people obsessing over, channel strips, tubes, plugins and mic pres (which has been going on since the early 90's) to "fix" their sound, but I can't introduce tape into the conversation?

Dealing with equipment issues: Although one of the posters in this thread presumed to welcome me to the world of computers I’ve been a computer consultant for 12 years. If anyone needs help with their Pro Tools rig, send me a PM, I’ll be happy to help… well maybe not thrilled because I’m a Linux guy, so am more into things like Ubuntu Studio.

But anyway, about costs… nothing is more costly than PC based systems. Since people use PCs for other things they usually fail to factor in general PC issues and upgrades into the cost of maintaining a DAW. I fix peoples networks and PCs for a living… I know what it costs.

No question this is a PC-centric culture, but the analog users I know aren’t at all as some like to portray them… old fashioned, outdated, behind the times, which is an uninformed view… probably analog myth number one. It even appears one or two members here are of the mind that the more years of experience you have, the less you know.

 

[mshilarious]

But I was specifically referring to the width of the soundstage, not the dynamic range, frequency response or any other specification.

Soundstage is irrelevant. By the time a signal hits a recorder, analog or digital, it may as well be dual mono. It's two discrete signals (except for crosstalk . . . hmmm, is that better in digital or tape?). Let's dispense with the illusion of stereo, and just talk about a single mono track. If there is a flaw in digital, it can be exposed in mono. If there is a distortion in tape, same thing.

The truth is, tape-o-philes can't argue dynamic range, frequency response, phase response, or any other actual measurable specification, because they know they lose those arguments. So you have to make up things like "soundstage". Doesn't cut the mustard with me.

When people have fundamental differences in perspective they don’t tend to scan a person’s argument for things that actually make sense… because they don’t want it to make sense.

Wrong. You see, I am intellectually curious. When somebody makes a claim to me, I first think, can I test that? If the answer is yes, then I do it.

If you search my old posts, you will find very detailed analyses of attenuation of high frequency signals vs. sample rate; comparison of cassette tape, VCR, and a control sample of D/A/D, just yesterday, an analysis of the need for dither given a typical acoustical noise floor. Those are just three examples I can remember, there may have been more. Nearly every day, I am analyzing the performance of an analog circuit. That's how I feed the family.

So for you to imply I dismiss your arguments because I don't want to believe them is very mistaken.

And if you really want me to, I will analyze your silly evidence-free assertions too. I've done plenty of D/A/D transfers before; as I mentioned, I designed and built an analog summing cable (wasn't much of a circuit design, really, just a resistor) and did a detailed comparison of use of that product vs. digital summing. So I already know how well my converters perform on a variety of signals (yes, including program material). I didn't post that analysis for public consumption, but I will redo it if need be.

Are you, on the other hand, actually willing to do any work, in terms of a repeatable experiment?

Ok, about your Nyquest comment. The Nyquest Theorem is digital 101. I can’t remember a time when I didn’t know it. And it only addresses minimum resolutions necessary for achieving a given frequency response without severe distortion. The actual quality of that frequency response as perceived by the human ear (not test equipment) is open to the listener.

Wrong. Wrong, wrong, wrong, wrong, wrong, wrong, wrong.

First, learn to spell the man's name, he was a brilliant scientist and earned the courtesy. Don't dismiss his work as that of a digital hack, because he did it long before the era of digital recording, and the body of his work is the foundation of not only digital recording but analog audio circuitry as well.

Now, as to your misunderstanding of his theory: It describes the minimum resolution necessary for achieving a given bandwidth-limited frequency response without any distortion. The only caveat in a real world circuit is the effect of the antialiasing filter required to achieve that bandwidth-limited response.

For any modern converter, your assertion that 16kHz would be adversely affected is laughable. If you had said 18.5kHz or higher, I would have somewhat agreed, but in comparison with the Portastudio, that's just a joke. The Porta must be down 20dB at 18.5kHz, versus like 0.5dB for digital. I'll have to dig out my old graphs and check.

And what's even better is that if you switch up to a mere 48kHz sample rate, the 18.5kHz attenuation goes away, and the 20kHz is now down maybe 0.5dB. Portastudios can't do that!

Look up my old posts, the graphs are there. I've done the work already.

But advocates of higher digital resolution are already making that argument, so it’s as much a digital vs. digital argument as it is an analog vs. digital argument. The analog people could go away and it would still be an issue.

No, the real digital people decided long ago: something like 60kHz is an adequate rate to completely eliminate attenuation in the audible band (to 20kHz) caused by antialiasing filter attenuation. Anybody who says different has a warehouse full of 192kHz gear they are trying to move.

I don’t need to read anyone’s white paper.

Why? It demonstrates with real actual math everything I just said above, and what's better is that anyone with a converter can verify its conclusions for themselves. What's even better is that you really don't even need a converter; just spend an hour with an SRC plug and you can verify Lavry's theory.

That's science for you, right here:

http://www.lavryengineering.com/white_papers/sample.pdf

That being said, I don’t know that frequency response is the test to settle the issue. It is only a hunch among many thoughts I have as to why digital hurts some people’s ears.

Well, you know, I have already skewered your frequency response and distortion hunchs, so you need a better hunch, something we can test for ourselves.

Right now, all you got is that digital hurts because it doesn't distort like tape.

Let's try a test to see if you can distinguish a live source vs. a (reasonable quality) digital playback. Get a MIDI controlled grand piano, a mic, a pre, a converter, and an A/B/X switch, and see if you can distinguish between a straight analog feed and A/D/A. And make sure that you don't know which wire leads to A, B, and X

But your experiences and mine are just anecdotal nothings in the grand scheme of the analog vs. digital issue, which is universal.

Again, wrong. I have not argued a single anecdote, only testable, repeatable, verifiable experiments. You are arguing from anecdote, not me.

The results of your proposed experiments would be insignificant in light of the sheer number of people that already report a preference for analog for it’s pleasant handling of music compared to digital

To quote the best movie ever (EVER!), Joe vs. the Volcano

I'M NOT ARGUING THAT!

Here's the fallacy:

1) Analog sounds better than digital;

2) Something that sounds good is more technically perfect than something that sounds bad;

therefore:

3) Digital is technically flawed.


I am happy to agree with #1. Mainly because I don't care! I have no desire to own a large reel tape machine, and that is what is required to realize your goal of good tape sound. Anybody who thinks a Portastudio has the same good qualities as 2" tape . . . I don't know what you are smoking. I've had a 424mkII for 11 years now. I like the little thing. I still used it as late as 2004. But to suggest it sounds either euphonically or technically better than my HD24 (which replaced it for live work), that's really funny.

But I am perfectly happy to believe that 2" sounds better than going straight to my converters. I don't know whether or not it's actually true, but because I haven't done the experiment myself, I keep my mouth shut. That's fine, I don't make my living from tracking, and I don't mind not having those nonlinearities adding to my recordings. Other people's stuff I work on is already analog source or not, no sweat off my back either way.

I make and sell analog gear exclusively, so I am hardly tied in to digital (except that it has greatly expanded my pool of customers). Everybody could buy a Studer tomorrow, and I think I could still sell them mics, so long as the interwebs (digital!) are still in existence so they can find out about me!

The giant flaw in the argument is #2. That just not true. Tape-o-philes that argue it are gonna get their pants pulled and maybe even an atomic wedgie! Which, I have to say, in 2008, they deserve!

I’m equally exasperated by members that know just enough about digital to be dangerous, but not much really beyond popular misconceptions. But, what we got is what we got on these anonymous forums.

Anonymity can bite me. My name is Jon O'Neil, I live in Kill Devil Hills, NC, and I am the owner and proprietor of Naiant Studio.

What else you got?

When I consider how analog benefits my sound and how I meet such hostility by simply sharing that information for the benefit of fellow recording enthusiasts

See, I was perfectly happy to stay out of this thread so long as you simply argued #1 (above). You will get no disagreement from me on that point.

It's #2 and #3 that got me involved . . . so enjoy the hostility (wedgie!)

 

[CoolCat]

The best original bands can do nothing but play frat parties and post music to pedophiles on MySpace. It's quite sad.

.

bwaahahahahahahahahaaaaa

come on man, it ain't that bad....geeez hell?
there are people actually selling some of their CD's...ok, not much most likely.
but instead of recording and sticking the 4track cassettes in the closet or mailing them to Capitol Records , I've bought some "HR INdie" stuff...my kids actually buy t-shirts from un signed bands which wasn't easy to do before the internet age of MP3sharing...

I agree with the mindset...hahahahaa...the overflooding of mass everything..

to feel special or rare of a product... when it seems there's 900 billion doing the exact same thing...is strange.

...but there's the positive side too. Maybe selling 10 T-shirts is a start, maybe thats better than selling 0 T-shirts of the band..

MySpace is like the Distribution..........HR is the RecordDeal/Studio...... now all we need is a Guitar Hero that can turn goofing off into Original music that can be recorded while we jump around with plastic fake guitars made for 1yr odl to adults?

Maybe.... the next Beatles will come from MySpace?
why not?

does it have to come from the big corporations to be "pro"?

 

[Bob's Mods]

As only a humble home rec'ker, I have been reading both sides of this thread with great interest. I am a technical person by training, some 30 years and counting working in the Boston area ever shrinking high tech industry. As a small fry home reck'er I can only comment on my experiences. First, I have struggled with digital. In the beginning it probably was due to lack of experience in the recording arts and junk hardware and software. Lately I've got much better gear and software plus knowledge that has greatly improved my results. Are there deeper issues related to digital as Beck assurts? Well I can say I do not like digital effects pedals, modeling amps or modeling software and never have. Some synth modeling software is very good however such as Native Instruments stuff but I don't own it and have little experience with it and cannot comment on how it compares to the real deal. Is that because the digital genie is not what its cracked up to be or are these devices just a poor implimentation in a digital meduim? I would argue that a phyco-acoustical test of digitally recorded mixes vs analog recorded mixes be done simliar to how new drugs are tested to test human reaction to the meduim. This, I believe, is what Beck has been implying. This phyco-acoustical test would settle this issue once and for all.

As far the accuracy issue of digital in the upper audio range, at 44.1 kHz, yes the accuracy is less than in the lower range due to the fact that less samples per cycle are used to recreate that region of the spectrum. Its not that big a deal because most peoples hearing is 15 kHz or less as has been pointed out earlier. This problem is easily fixed by recording at a higher rate, 96 kHz for example. This would more than double the samples being taken at the upper end of the hearing spectrum and so improve the capturing accuracy during record and reconstruction of that range in playback. This issue affects fidelity more so. The full spectrum does not have to be perfect for us to enjoy the mix. You'll still love your favorite tunes heard on AM radio just as much as on an audioplile system. Its more a question of fidelity which is a personal preference rather than a defect in the production of the audio.

It should be noted that an analog recording is a faithfull recording of the subject matter (bandwidth limiting and smearing effects aside). Digital is actually a "synthetic" capturing and reproduction of data points taken of an analog signal, thus an interpretation of the original analog signal. This synthetic interpretation of the analog signal is sufficienct enough to fool our brains into thinking that it is the real deal. The pure analog version as recorded by tape is actually a true mirror copy of the subject audio. This is a hair splitting distinction but a real one just the same.

What if the Bealtes were to be transported into the future and started recording in a modern day digital studio? Would their mixes sound as compelling or would our ears ,that have been conditioned by years of tape mixes, reject them on the new meduim as just not sounding right? For me, after years of conditioning, having grown up in the 60s and 70s, listening to all the great classic stuff I love that was done on tape, my ears have a tougher time with digitally recorded mixes. I have no problem with classic recordings put onto to CD however. The change from vinyl to CD has not effected their feel to me in any way.

Although this is great subject matter for discussion, the reality is that part time home rec'rs have only one option, the digital box, to ply our hobby on. Yeah you can throw in a few a analog devices here and there, but for the most part, the home recording hobbiest is firmly planted in the digital world. There is just more good than bad. Digital, in the box, does not appear to be the perfect solution, as its made out to be, for everything that was once an analog only solution does it? Why is that? Is it poor implimentation of the analog to digital solution? Is there some inherent phyco-acoustical flaw in digital? Are we preconditioned to the tape sound due to its long years of use? Is the usefulness of the digital wonder box for an all in one solution just over blown hype? Is it the proliferation of lower quality recordings of the less technical/musical masses? Is it the quality of tunes today? For better or for worse (where have I heard that before???) digital is the new mass medium and its here to stay. All this ties in neatly to the original pole question.

Bob the Mod guy.

 

[SouthSIDE Glen]

I'm with MSH for 99.9% of his post, especially the part discussing Nyquist and the otoal refutation of the alleged soundstage issues. I already said my part about the soundstage, and back up NSH in his description as well. As far as Nyquist: Hell, even most digital advocates misuderstand just what Nyquist actually means, but when an analo-phile gets a hold of it, it is about as bad as listening to a traffic reporter trying to explain quantum mechanics.

The theorum itself *proves* (it is a theory only in the same way that Newton's theory of gravity is a theory) that in order to 100% accurately reproduce a signal with a maximum frequency of x, one only needs to sample at a rate of 2x. This means that in order to 100% accurately reproduce a signal with a maximum frequency of 20kHz, one only needs a sample rate of 40kHz. Period. Now, it turns out that there are some practical physical design issues complicating this simple formula in physical practice that have to do with the physical impossibility of strictly bandwidth limiting the signal. I won't go into technical detail here; it's published ad nauseum everyhere else, but it's for these reasons that we have a sample rate of 44.1kHz instead of simply 40kHz.

Any loss or distortion at 20kHz or below in the conversion to or from digital has absolutely ZERO to do with sample rate, and ZERO to do with inherant problems with the digital format itself. Any such degradations are entirely the domain of the quality of the physical design of the converter (including, I might add, the quality of the analong side of the converter), and can and are addressed by affecting the quality of that design. To blame "digital" for someone making a crap converter would be like blaming "analog" for someone making a crap tape head assembly.

The 1% where I disagree with msh's post is in step #1 of his reasoning. That is the weak point. Analog does NOT inherantly sound better than digital. Nor, to be fiar, does digital sound inherantly better than analog. It is, quite simply, a product of nurtured perception to say one or the other, and there is a very tight range of age and experience that says one or the other. If you were to take Mozart and transport him to 2008, and ask him to say which one sounded "better" to him, analog or digital, he'd most likely say they both sounded strange, just different flavors of strange, and have no preference for one over the other. Why? Because one is NOT closer to reality than the other. They are both artificial, just in different ways.

Which one is "least artificial" depends entirely upon the quality of engineer first, and the quality of his gear second. A top-notch digital production will blow away a phoned-in analog production, and vice versa. How about a top-notch digital vs. a top-notch analog? That will depend almost entirely upon the weakest link in the production chain of either one, along with any listener bias one way or the other.

G.

 

[Bob's Mods]

SG, accuracy is improved as the sample rate goes, Nyquist is only a minimum. You are capturing more data points of each cycle.

This seems to be turning into a "he says tomato she says toe-maa-toe" kind of argument. Beck, whatever his assersions, favors an analog sound subjectively while MSH is trying to defend that merits of digital with technical arguments.

Am I just too conditioned to tape based mixes that my unconscience mind struggles with the sound of digital? (deep one here). It is acknowledged by all parties that there is a difference.

Me, well I'm just a humble home wrecker!

Bob

 

[SouthSIDE Glen]

SG, accuracy is improved as the sample rate goes, Nyquist is only a minimum. You are capturing more data points of each cycle.

Sorry, Bob, but that is perhaps the biggest source of misunderstanding when it comes to how Nyquist works, and is simply not true. Nyquist has very little to do with "data points" in the conventional sense. It's not a connect-the-dots from sample-to-sample function.

Think of it this way; if that was how it worked, it would be imposible to reconstruct just about anything audible except maybe turntable rumble from a ~40kHz sample rate. At that rate, one is only getting 2 samples for every 20kHz wave. If reconstruction worked via the "data points" misunderstanding, if those samples fell at the zero crossings, there'd be nothing to reconstruct, it'd be a flatline. In fact if they fell anywhere else on the waves except the wave crests, they'd form a triangle wave of completly incorrect attack and amplitude.

A ~40Khz sample rate only give you 4 samples at 10kHz, 8 samples at 5kHz. Neither of those is enough data points to be able to guarantee reconstruction by the "data points" method.

None of that is an issue, because that is not how Nyquist works. Nyquist explains a statistical mathematical method based upon digital information theory - not analog data points - that allows and proves that waveform reconstruction is 100% accurate at twice [EDIT: Oops, I meant half, not twice] the sampling frequency. All increasing the sample rate does is increase the maximum frequency reproduced. The only reasons we have a higher sample rate like 44.1 instead of exactly 40 is because of physical design limitations, not because of improved sample rate resolution.

G.

 

[Bob's Mods]

I probably don't understand the inner workings of the theory completely...but certainly a connect the dots interpretation plays a role. In the upper ranges it hard to image accuracy being equal to the lower range. At 44.1 you are taking a snapshot of the waveform every 22.7 uS. Those are fixed points in time. At 100 Hz your period signal period is 10 mS. Thats 440 or so data points to reconstruct that waveform which is plenty to figure out what its all about. At 15 kHz, the period of your signal is only 67 uS. Thats only about 3 data points. Not enough to reconstruct a "pretty" version of it but enough to demonstrate there is something there. Its of such low amplitude and the ear is less sensitive at that range it does not matter a hell of a lot. The ear determines intelligability in the 300 Hz to 3000 kHz range anyway.
The A/D generates a 24 bit code (or something to that effect) for the voltage that exists at that time. If you look at the output of the D/A thats reproducing that data stream on a digital scope you can clearly see the digital stair steps created every 22.7 uS. As the sample rate increases you see a more rounded natural looking waveform. Smaller stair steps. There are funky looking waveforms in the upper region at 44.1 kHz. The ear seems to discount it though. There is a distortion in the upper range. As for the stair steps, we don't hear the stair step artifacts as the amplitude is too low with respect to the waveform being generated and the frequency of occurance is above human hearing anyway. If just gets filtered out. I may not have the complete picture of how the theory works but have a resonable understanding.

Bob

 

[mwthompson]

Hehe, great thread.

Beck: I've been doing some searching, but am having a hard time finding much substantial in the way of reel to reel decks. I want a simple stereo setup to copy mixes or even single tracks to as an effect. What should I be looking for?

 

[CoolCat]

in 5yrs or less it will be nano second, not micro.

the simulated sign wave, reconstructed digital is no different than a tape simulating the original sound vis magnetic particles and tape hiss.

i can understand the romance and "feel" for the analog dewds, it is for real...its like a natural thing to them. Like a wood worker some want to put their nose in the dust others want the newest fangled hitech zip saw do evertyhing without effort saw.

insightful was the blurb someone posted about in reality, majority won't be home recording with 2" reel to reels anyway, let alone maintain them to Abbey Road specs of 1964.

comparing Digital to Cassette tape 4-track machines, I doubt anyone gets the same romantic feeling for that crap...other than ease of use.

I have an old boss at work, he hates computers... it's really sad to see him try to function with one, I wouldn't like modern stuff either if it was that painful.

 

[Bob's Mods]

the simulated sign wave, reconstructed digital is no different than a tape simulating the original sound vis magnetic particles and tape hiss.

I agree, this is reasonable.

comparing Digital to Cassette tape 4-track machines, I doubt anyone gets the same romantic feeling for that crap...other than ease of use.

CC, this is probably where we should be here. Comparing what home recorders were using to what we got now. As much as I love the sound of tape, its not practical for the average hobbiest on a scale that makes a difference. What we really should be doing here is putting all the ugly warts of digital for home recording purposes under a microscope. When we look at the wonder box as compared to the PortaStudio of old....well there really is no comparison.

Bob...yes I've modded lots of stuff searching for the elusive sound of high end studio. Its only part of the solution.

 

[SouthSIDE Glen]

I probably don't understand the inner workings of the theory completely...but certainly a connect the dots interpretation plays a role.

Ah, but it really doesn't, not in the classical sense you're thinking. This is a very difficult thing to understand; I didn't get a handle on it myself until just a few years ago, and even now, I don't understand enough of the math (it's very advanced) to explain with technical precision exactly how the reconstruction does work.

I didn't in an earlier post use the allusion in to quantum mechanics lightly. The difference between how analog and digital information theories work is in many pedestrian ways quite analogous to the difference between classical mechanics and relativity or quantum mechanics. The mental translation between them is quite difficult, and some things that seem to make basic sense in the analog/classical world do not correlate to digital. Electrons are not really balls orbiting a an atomic nucleus, and analog waves are not really a connect-the-dot reconstruction of sequential digital samples.

At 15 kHz, the period of your signal is only 67 uS. Thats only about 3 data points. Not enough to reconstruct a "pretty" version of it but enough to demonstrate there is something there.

Think about that. It's not even enough to reconstruct an ugly version of it by the "classical" method.

Let's take it down to 10kHz. I don't think anybody - not even Beck - would argue that a 44.1k sample rate doesn't do a least a fairly decent job of reconstructing a 10kHz wave. At 10khz were getting just a fraction more than 4 samples a wave. The only way any kind of decent analog reconstruction of that wave could possibly be made is if those 4 samples happened to fall at the zero crossings and at the wave crests. And even then, it would be a triangle wave and not a sine wave.

Many reply that curve fitting is used to smooth the straight lines into a sine. There are a couple of problems with that; there is more than one curve that can fit any given set of four data points, each one radically different from the others. Imagine the four points, instead of falling on the crossings and peaks, fell at the four 50% amplitude points along the slopes of a complete 10kHz cycle. That could be represented by the true sine wave, or it could be represented by a trinagle wave of half the amplitude and twice the frequency, or a sine of twice the frequency and somewhat differing amplitude, or a square wave, or a sine that's clipped at 50%, or...well, you get the idea. Via the perceived classical method, there's just no way to represent to any degree of accuracy whatsoever a frequency response of 10kHz. Even 5kHz would be dodgy.

Digital isn't THAT bad . No, it's simply - well, perhaqps not so simply - that that way of looking at reconstruction is just not really correct.

The A/D generates a 24 bit code (or something to that effect) for the voltage that exists at that time. If you look at the output of the D/A thats reproducing that data stream on a digital scope you can clearly see the digital stair steps created every 22.7 uS. As the sample rate increases you see a more rounded natural looking waveform.

And this classical picture - as always - leads one into very popular misconception #2 about how digital actually works. The misconception that increased sample rate means increased resolution of a given frequency.

Those "stairsteps" are a very nice way of describing the encoding process on a basic level; i.e. as a way to describe what "sampling" actually means. But it should not be extrapolated into being an explanation of how recostruction works. Again, remember that at 10kHz, there are only 4 "stairs" that when shown without the original wave form can be interpreted to mean stairs going an any different directions or slopes (rather like an M.C Escher painting squashed into one stairway ). If there are eight different shapes that could potentially be drawn from those 4 "stairs", that would mean that there is only a 1-in-8, or 12% chance of getting it right. Even the most steadfast analo-phile wouldn't claim that digital gets a 10kHz signal 100% wrong 88% of the time.

Remember when we're looking at those stairstep pictures that we are looking them being overlaid on top of a sine wave of set frequency. Increasing the number of samples taken of that wave - i.e. the distance between the steps - does not increase the accuracy of reproduction of that wave of that fequency. If there are any irregularities in the wave, let's say a transient on the upslope of the wave for example, that requres an increased sample rate so that the transient does not get lost between the cracks between samples, that has nothing to do with the original wave of that set frequency. That transient occurs in a shorter length of time, meaning that it occurs at a higher frequency. The increased sample rate only increases the frequency response so the transient can be caught, but it does nothing to increase the accuracy or "resolution" of the lower frequency sine on which the transient is imposed.

Higher sampre rate = increased frequency range. It does not equal, however, increased "resolution" of the frequencies already covered under Nyquist at lower sampling rates. Because, again, that is not how it works. The "stairsteps" represent what "sampling" means, not how analog reconstruction occurs. Those stairsteps you see on the digital scope *are on a digital scope* looking at the data stream. Hook up an anlaog scope after the conversion to analog, read the actual resulting analog voltages, and you'll see no such stairstepping or straigt lines of connection; you'll see a standard, smooth analog wave.

Many people like to use digital photography as an analogy, equating sample rate to pixels-per-inch; the more samples/pixels you have, the sharper your resolution. While inderstandable, this is a false analogy. Sample rate is more analogous to the frequency response of the CCD; the higher the sample rate, the wider the color range (in this case >20kHz response would be going beyond the visible range and into ultravoilet ).

G.

 

[mshilarious]

The 1% where I disagree with msh's post is in step #1 of his reasoning. That is the weak point. Analog does NOT inherantly sound better than digital. Nor, to be fiar, does digital sound inherantly better than analog.

Oh, I will readily concede that. I don't know because I don't have a well-calibrated 2" machine in my studio, nor am I a trained operator of such gear. So to test it myself, I would have to borrow a tape machine and an engineer. I'm not interested in doing that, so I simply concede the point.

You are free to argue it

If you were to take Mozart and transport him to 2008, and ask him to say which one sounded "better" to him, analog or digital, he'd most likely say they both sounded strange, just different flavors of strange, and have no preference for one over the other.

Mozart had no need for recording; he had a musical memory as perfect as probably ever humanly achieved. Neither did Beethoven, who composed (and even tried to conduct) in total deafness.

Recording firmly lies in the province of we mere mortals, not in the realm of those, as my parents' old Time-life album collection calls them, the Great Men of Music. (Sorry about the sexism there, it's not my fault! For the record, I recall at least one of the recordings featured a chick--Wanda Landowska).



Bob's Mods, your understanding of Nyquist is limited because the representation of high-frequency signals by relatively few samples is not a barrier to accurate reconstruction of the original signal. As you alluded, the stairstep wave contains distortions that are harmonics of the high-frequency fundamental. But these are all removed by the D/A's anti-imaging filter, so it's not up to the listener's ear (or the following analog chain) to attempt to deal with the distortion--it's gone before the converter even lets the signal go.

 

[Bob's Mods]

My understanding of these things comes from a technologist point of view, meaning I know enough about their operation to fix them when they break. A deeper level of the theory behind them would be required to design them which I do not do. I can accept your explaination as they probably aren't as simple as the technologist's model. My view of their operation has given me doubts similar to those points you've brought up relating to the reconstruction of waveforms.

Bob

 

[Bob's Mods]

In Closing.....

I can conclude that I must except the limitations of home recording. Its not going to get me (and probably you) to a high end studio sound either past or present which was my original intent. The best I am able to attain is a very sound with much flexiblity to play the full band. Its an excellant demo recording however that is probably good for most untrained listeners and thats as far as the limited software and gear quality gets me. My original intention was to try an equal a good commercial studio. My conclusion is this is not possible without a significant increase cost. My advice for those who have been following this thread....if you want top notch tracks...do as Sonic Albert does, record them at a good studio....then bring them home and mix them yourself. You can build a reasonable cost mixing station cheaper than you can build the studio that actually records the tracks.

This has been an enjoyable thread and a widely read one as well.

Rock On,
Bob the Mod Guy.

 

[DavidK]

Maybe.... the next Beatles will come from MySpace?
why not?

Wherever they come from, they WILL come. Always have, always will. We just might not recognize them. Bach was all but forgotten until a hundred years after his death. Likewise, Ludwig Spohr was a huge composer in his day, as big as Beethoven. He's largely forgotten, and for good reason. Lots of classical pieces were total flops at the premieres and only gained attention years later. The next Beatles might be recording now, and we dont even know it.

 

[CoolCat]

yes, or the Next Eddie VanHalen is recording with his M-Audio FastTrack plugged in backwards...or the next Hendrix might be lighting his Firepod on Fire right now with BIC lighter !!

my kids went through freestyle...er...they couldn't really play or tune their instruments but they had fun, they jammed laughed and called it "fee style"....sometimes it was listenable too.

then they learned and started playing like everyone else kind of. its easier listening but it lost a lot of edge. there is still a thrill of a well done piece of music, or even a well recorded track...like "damn that bass sounds awesome..." or the accidental flanger piece in the chorus..

hopefully one thing with HR is that due to being "free$$" people can create and record new things as its not costing a bunch of money at a pro studio... to buy a suit another Mercede's or Yaht or Lear Jet.

its the same old underground thing imo...except instead of Indie labels its HR Labels...so the money is spread out amongst even more.

Maybe Indie/Underground companys will pop up more and more and break into the shelves of WalMart and BestBuy and they will allow new music to be stocked in their Mega Stores instead of just filling the shelves from a few major labels?
I'm not really sure how the HR could break into Best Buy....or WalMart....maybe it will just sit in MySpace land forever, or Ipod download land....what if CD's die out?

what if I got off my ass and did something today?

 

[Bob's Mods]

Studio Console Question, Neve, SSL

I have listened to a number of commercially recorded mixes recorded digitally. They sound pretty good. Just as enjoyable as tape. The commercial quality thing vs what a home recorder can get may be due to different reasons.

I've got a LynxOne card which is considered most likely on the bottom end of true professional hardware. I'm now finding it hard to believe its a digital vs tape thing as this thread has focused on.

What does a commercial studio have? It should have an acoustically tuned room. This will make some difference but its not the smoking gun. I'm sure some better mics. Again, here, I do not believe the most expensive mics in the world will make the difference bringing the home recorded sound up to a commercial level. They also got a fair share of outboard gear, some of which may make more of a difference. What they do have that I believe will make a difference is......their console. Many of the great classics were recorded on Neve consoles. Today I believe its the SSL thats in favor. These are very costly consoles. I believe the SSL has some nice built in bells and whistles. The designs of these preamps are unique and special, typically more complex than the usual basic op amp design of preamps that are commonly found in a home recording rig. I believe these console preamps with included EQs add a unique color to the tracks by focusing on bringing out the harmonic content of the recorded material. Something a basic op amp design just won't do. My question goes out to those have worked with commercial studio grade consoles AND simple home recording rigs. Do those studio consoles (just used alone with no outboard gear) really help to create the magic that a good studio imparts to tracks and mixes? If I brought one of those fancy consoles home and A/B'd against what I"m using in my home computer rig, would I hear the difference?

Bob

 

[TerraMortim]

And what's even better is that if you switch up to a mere 48kHz sample rate, the 18.5kHz attenuation goes away, and the 20kHz is now down maybe 0.5dB. Portastudios can't do that!

yes! I was just about to say that. If you can hear artifacts such as that at 44.1 (I can) then use a higher sample rate. I record everything 24bit 96kHz... It sounds to me like Beck can't wrap his head around anything beyond the typical redbook audio CD format (which is very outdated, compared to the technology used in even the average project studio today) I, personally don't like the sound of a 16bit 44.1kHz cd, and can hear the difference between recording a signal at 44.1 or 96kHz (unless it's something like a distorted guitar in some cases or something that has a relatively narrow frequency spectrum)

You see, Beck believes that he is part of this elite group of super-hearing gods. I haven't gotten my ears tested in a couple of years, but last time they were around 18kHz in both ears... let's just say that yes, I can absolutely agree that with a 44.1kHz recording I CAN hear some of the digital artifacts... but...we have 48kHz, 88.2kHz, 96kHz, and the impractical 192kHz. There ARE other release formats for digital audio, even though they are pretty uncommon... there's DVD-A which is burned to a normal DVD, which I believe can go up to 24bit 192kHz, there's super audio CD which I don't know all that much about but I do know it surpasses by a long shot the specs of the redbook format. Even the audio on a video DVD, while compressed to an AC3 is anywhere from 24bit 48khz to 24bit 192kHz (ac3 encoding, by my ears is actually not bad...sure beats mp3 compression to hell.. and from my experience is about the same size...maybe the tiniest bit larger...I think it's a lossless codec, anyway)

Anyway, just a couple of other points so I don't end up flooding the thread with post after post... Beck... no, you didn't introduce any discussion, you quoted some crotchety old goon who was playing a fiddle over how digital recording is destroying music and turning people into sonic/musical retards, and there is no concern for fidelity any more... it was a stupid quote, which I'd be willing to bet this sMear guy would probably laugh at himself nowadays. People say things they later recant all the time... while he might prefer the sound of tape, it's a bit silly to panic that music is being destroyed by any other method (unless..he really IS an idiot)

As was brought up in this post before.. you're comparing the top end analog gear to bottom end digital gear... comparing a studer multitrack in perfect working order to a cheap portastudio doesn't even make sense. There are plenty of analog cassette versions of the portastudio that are on the same level of "quality" might I add. Now that we've gotten past that. If you're using a good digital system (Pro Tools is starting to get past their claim to that title), you should be able to get very good results, and in some cases depending on the song better than what tape would do to it. As I mentioned before. The sound of both are very subjective, and one must take their bias out of the equasion, and use whatever will serve the song. I have used 2" tape for projects, and I've used nothing other than a computer with an interface... and I've used a mixture in a studio with a nice console and outboard gear... None of them were inherently better... some had better results for one certain sound, some had better results for another. There are some songs that I've recorded that would fall apart when introduced to the pleasant distortions of tape, it just wouldn't sound right. But then, there are some songs, that ask for, even scream for, some nice analog distortions, wether it be a channel strip, or a track of analog tape. There is not just one good recorded sound. I can understand this...this is why I am open to using ANYTHING, even a fucking cassette if god forbid it would help get a certain sound (have before) to make a song have more feeling behind it. There are many different production sounds, styles, and most of them sound just wonderful to my ears. Why is that? They are appropriate for the music being recorded. If I were recording a traditional jazz group, I'd reach for the tape machine before anything else... I might only use a DAW for some very late in the game stuff. If I were recording a death metal band, I'd probably split the duties... I might throw the guitars and bass onto tape, and the rest into a DAW... industrial? I'd go all digital, unless it were a very specific scenario, which is fairly unlikely in that genre... You see what I'm saying?

I haven't used any of the linux based DAW software, but I suspect they must not be very good, since I don't hear of really anyone using them professionally... there are plenty of mac people in pro circles, as well as windows people in pro circles, but you don't hear of linux based studios. I don't know how advanced Ubuntu studio is, but I doubt it's as good as Logic, Sonar, or Cubase.. which have a very demanding user base, who at times can even go overboard on how much they expect their software to do... I can't imagine, with the lack of professional users of any given linux DAW software, there would be the same pressure, as well as the open source/free nature of linux software... people tend to expect more from something they paid good money for, and/or have no direct control over the way the program is put together.

And as far as experimenting with different types of production on the exact same tracks, done it...results...all over the map. I took some tracks recorded to 2" tape... digitized them, did a full studio mix from tape with lots of great gear, nice NEVE console, the tape machine was a really nice Studer, some songs were phenominal, and some songs the in the box Logic Pro mix was better (and even sounded, more "analog", than the tape) I wish I still had copies of those sessions... I'd post them without saying which ones and have you choose which the analog or digital ones were... keep in mind I was digitizing and mixing in Logic at 24bit (32bit floating point) 96kHz.

 

[Beck]

Since I don’t have time to answer every point on the thread right now (Today) I will just share some general observations. Except for a few members that seem to be really trying there is a lot of skewing and twisting of words and meanings and attempts to discredit. That’s not productive.

Although well-informed, qualified people may disagree on technical matters, it’s not necessarily rising to that level in response to my posts.

Statements in many of the above replies are simply inaccurate… not the least of which are some comments about the bloody Nyquist Sampling Theorem and the implications some are reading into it. I find most of the time on these forums people invoke Nyquist to appear they know something others don’t… often just repeating things they’ve heard on the web.

Again, Nyquist doesn’t address subjective quality of sound as perceived by the listener, but only the sampling frequency necessary to prevent aliasing. That’s it folks. It’s rudimentary recording 101 material… no government secrets here. When we examine digital in light of Nyquist's theorem we are talking in terms of adequate, not optimal.

And besides, it was proposed by Nyquist in 1928, so that makes him an “Old guy.”

If anyone wants to read something concerning the Nyquist Theorem…

http://www.digital-recordings.com/publ/pubneq.html

Frankly though, going down this road is a waste of time when discussing human perception of music.

Hmmm… how to correct misconceptions without stepping on toes and dooming every thread to a spiraling pissing match? I don’t have an answer to that. If we could bottle the solution and sell it… oh wait, how ‘bout a little Captain Morgan?

Glen, what, pray tell is an “Analo-phile.” Speaking for myself, I’m a pragmatist that will use whatever tools work to achieve sonic excellence in recording. Why some people have to treat the issue as though these are opposing political parties, I don’t know. I have to call you on the comments that put people in boxes and dismiss them for some status. Earlier you were dismissive of Scholz, citing some vague reference to his obsession with a FFT spectrum analyzer and that it took him too long… a year and a half on the first album. He actually worked on the album for more years than that, but what does this have to do with anything? It’s akin to saying, “Oh, he doesn’t know anything because he’s black, or a Republican, or from The South, or an American, or rides a Honda instead of a Harley, etc.” Stop that now, will ya?

You can easily find something quirky about most artists and musicians. I have a long list of recording icons that prefer analog and a long list of studios that use it. Most also use digital where it works. I’m sure you can find something about each one that discredits them in your mind… maybe even a criminal record if you dig deep enough. But that approach is not reasonable debate and doesn’t help in sorting out a complicated issue. This doesn’t make you a bad person; it’s just bad argumentation.

Real quickly, about the soundstage, which by the way mshilarious, it’s a term that has been in use relating to stereo imagery since stereo was invented.

“Soundstage” refers to the width of the stereo image. Imaging and soundstage are fundamental to the study of psychoacoustics. Stereophonic music with a wide soundstage will sound as though it extends outside the physical placement of the speakers. Digital narrows the soundstage somewhere within the physical placement of the speakers. 16/44.1, 16/48, 24/48, 24/96, etc… it doesn’t matter.

This has been a subject of discussion in professional journals since before I was aware of it (1989) right up to the present. A narrowing of the stereo field in digital has always been recognized. It’s not a controversial issue.

So, I’m afraid it’s a big negative on the idea that hard-pan is hard-pan regardless of format. It is not. Some have theorized that it’s because a CD digital stereo image is drawn by an algorithm from a single soundfile, while analog stereo consists of two physically separate tracks containing individual electromagnetic signals, not binary values representing a signal.

A song on a CD is one soundfile. Values are interpreted by an algorithm, including relative placement of points in the stereo image. The same song on cassette, open-reel or LP consists of two separate physical tracks. The relative position of an instrument is dependant on raw amplitude in each (left-right) channel.

"If you want my advice, with all the available digital technology you still can't beat the sound of a good analog mixdown.... The effect on your sound can be dramatic. With an analog mixdown, you have a much wider, deeper sound with greater stereo imaging. An analog mixdown has a texture that digital cannot produce. And, simply put, to my ears it sounds better ... that's it. No more explanation needed."
--George Graves, Chief Engineer - Lacquer Channel Mastering, Toronto
Professional Sound Magazine, April 1998

To be fair, not many are able to A/B analog and digital sources these days. For those who can it’s quite perceptible. There are many things one can discern when able to A/B the two.

It brings to mind the experiences of Steely Dan during recording of their 2003 release Everything Must Go. The group was an early digital pioneer through association with Roger Nichols. Most albums from late 70’s to 2000's Two Against Nature were tracked digitally. For Everything Must Go they used the services of Elliot Scheiner and the facilities at Sear Sound.

"When we got to the studio I said to them 'Look, this is not for your record, let's do it on analogue. Let me give you a taste of that again.' They said, 'fine', not expecting much. But when they heard the first playback, they went wild. They had completely forgotten how good analogue sounded… It did sound amazing. When they realized that it was great to work with live playing and analogue again they decided to record the whole of Everything Must Go this way."

"When you are able to A/B analogue and digital, which we could do in this case, there's simply no comparison. The top end is so sweet and beautiful. I've never heard anyone say about digital, even at 24-bit/96kHz or 192kHz: 'Isn't the top end as sweet and beautiful as you've ever heard?' You don't because digital just doesn't sound that way."
--Elliot Scheiner - Sound on Sound Magazine, Aug 2003