“Analog versus digital” – the discussion, it seems, is everywhere. The problem is, many people simply don’t understand what these terms mean. In one 25-minute video – engaging and entertaining to watch straight to the end – the biggest myths all get busted.

In short:
1. 16-bit, 44.1 kHz really is okay for many tasks. (You’re saving that data for the computer and processing rather than your own ears. Hope to talk about this question in more detail soon.)
2. Digital audio doesn’t involve stairstepping.
3. Digital signals can store and be used to reproduce sound that’s identical to what’s stored in analog form.

“Choosing” between analog and digital, as categories, therefore doesn’t make any sense at all. Now, choosing between individual filters, for instance, or caring about the physical design of electronic instruments, or recognizing that you can screw up a digital or an analog recording – all those things do matter. In fact, they matter so much that obscuring them with misinformation is a very bad thing.

The video is the work of Monty Montgomery at xiph.org. (See also: http://xiph.org/video/vid2.shtml and http://xiph.org/video/) Watch the video, but here’s some discussion:

Digital technology is fairly simple to define. A system using digital signal simply represents information as discrete, sampled values. An analog signal use a continuously-varying electrical signal. Both are means of encoding – neither is the literal sound. A digital system is so-named because those discrete values are akin to counting (hence “digits,” as in counting on your fingers), whereas an analog system uses an electrical signal that is analogous to – though not literally – the original, in that it varies in the way that (for sound) pressure would.

The problem is, people imagine digital signals to be something other than what they are in reality. Ubiquity can breed ignorance. Before digital became so widespread, recording and photography were the first revolution. And perhaps part of the problem is that our society has become so comfortable with those processes – ones that would have seemed magical to someone just over a century ago – that we have failed to distinguish between the representation and the real. But any photograph, any recording is distinct from its original subject.

Analog and digital signals are, like words and numbers, a means of encoding information. Each has limits. In sound, those limits are measured in the dimensions that measure audio, frequency and amplitude.

Digital is no less “real” than analog – and because we listen, in the end, to sound and not the signal, the two can achieve the same results. That means that you don’t have to choose. This is not a religious matter. It’s an implementation detail.

That’s not to say that the difference between analog and digital itself is irrelevant – implementation details can be very important. If you realize that fundamentally, digital and analog signals can create and capture the same sounds, then you turn instead to all of the other potential decisions a designer might make. There are many varieties of different filters, for instance, each with different characteristics. The choice of analog or digital circuitry then becomes dependent on what is most economical, most logical, and what desired sound and usability characteristics a circuit would have.

And as people over-emphasize the difference in signal and fundamental sound characteristics, they also ignore everything else.

The choice of analog control or digital control, for instance, is significant. (In short: without smoothing, digital controls can cause stair-stepping effects, and likewise analog controls may be more limited in terms of features like automation.)

This also puts in sharper relief the other reasons people favor analog technology. Analog’s “warmth,” for instance, may not be a fundamental characteristic of analog signal, but it is characteristic of other tendencies of analog designs. It tells us in part that having more literal data fidelity is not always better. Analog gear also behaves in unique ways, susceptible to variations in climate, age, dirt, and other features – something that can be positive in some cases and negative in others, but that is harder to model in digital form.

Most of all, it’s unfortunate that the term “analog” has substituted for “physical,” particularly outside sound contexts. No hardware is truly “digital.” All of it incorporates some amount of analog circuitry, for one, and it’s also the sum total of many design decisions. The fact that we think of computers as not having physical interfaces is perhaps itself a critique of the physical interaction design of computers – we’re in a way used to the mouse and keyboard that we may forget we’re having a tangible experience at all. The advantage of other designs can be to remind people of that experience.

When people describe the appeal of vinyl records, hardware synths covered in knobs and switches, patch cables and modulars, and other “analog” experiences, what they’re really saying is that they like the physical qualities of these things. And there’s no reason digital technology can’t be involved. Increasingly, it is: music is now very often digitally recorded, mixed, and mastered before being pressed to vinyl, and digital instruments are making use of more knobs and switches and even patch cords, rather than focusing on “virtual” experiences of screens and the like.

Instead of getting stuck in meaningless debates like whether analog or digital is “better,” in other words, we need to have very meaningful debates about design, sound, music, and art. But that sounds, by contrast, like a good use of time.

Spend the 25 minutes – you won’t regret it, even if this stuff is review. Thanks to Chris Randall for the tip.

  • Chip

    This video rules. In my opinion, though, your quick statement “16-bit, 44.1 kHz really is okay” may be true as a *final* format for distribution, but I’d caution over-reliance on this level of resolution for the original recording. My issue is with taking a 16-bit recording and performing a long multi-step chain of digital processing. There is the risk that each step of digital processing returns to the 16-bit resolution, which means that each operation (amplification, EQ, chorus, whatever) adds truncation/quantization noise. Modern DAWs are all 32-bit or 64-bit internally, which avoids many of these issues, but your plug-ins might not be so smart. If you use newer plugins (at the 32-bit or 64-bit level), you’ll be fine. But if you use old software, be aware that if you apply lots and lots of processing with these old guys will slowly degrade your audio. (Don’t get me wrong, it’s still *way* better than the generation losses that occur with the old cassette-based portastudios!)

    • PaulDavisTheFirst

      just for reference – it has been at least 12 years since the VST standard specified a 32 bit floating point data processing pathway for plugins on native CPUs. if you have plugins that also use integer or fixed point math internally, they are either VERY old or do a wonderful job thanks to their internal sample representation format, rather than in spite of it :)

    • Chip

      Thanks for the info on VSTs. But, starting in on this stuff in the 90s, outboard digital processing *hardware* was and is a literal and figurative black box where you have no idea what internal digital representations are being used. So, back in the day (and extending into the 2000’s because the 90s hardware was so cheap on the used market) all the outboard digital processing boxes could accumulate a lot of gunk on your sound, despite the S/PDIF and ADAT and whatever digital audio interface you were using between all the gear. Going all-in with the PC, once that became feasible, was an improvement on that front, because of the deeper bit depths and the standardization. But, then you get the problems of the quality of the Windows drivers, unknown behind the scenes re-sampling, latency, MIDI sync, etc. So, while I feel that this video successfully makes the case for debunking many of the negative myths of digital, there are still reasons to go analog….some of which are not related to perceived sound quality.

  • Softcore

    As I had said when I first watched this video:
    Finally, a guy who knows what he’s on about!

  • cool story bro

    cool for a sine waves,

    but not actual music

    which relies on many many harmonics and micro harmonics

    all of those frequencies stack up

    once over-processed by crappy plugins developers

    your song/audio is ruined.

    Dont get me started on digital summing…

    use some tube pre’s and some outboard compressors

    you’ll be happy you did 😉

    • Kurt James Werner

      Bro, do you even FFT?

      What exactly is a “micro harmonic?” Something non-LTI, surely.

      I wonder, is there any reason to use outboard analog summing except a preference for the analog non-linearities/distortion that is induced?

    • http://dinside.no Øivind Idsø

      You ARE kidding, right? Please tell me you are.

    • just passing

      given woo’s nature
      it is perhaps best expressed
      only as free verse

    • http://www.pro-tools-expert.com/ Russ Hughes

      I would love to read your comment but it’s been ruined with pixel conversion.

    • jmej

      You can think of all of those harmonics as sine waves if it helps you understand this conceptually. Also – tube pres and outboard compressors won’t help you with accurate signal reproduction.. you’re talking about a completely different thing there.

  • plgDavid

    My grudge with this video is that, while of course the stairsteps don’t exist “virtually” , you still have to convert these discreet values to Analog, and this process is totally dependent on the design/implementation of the DAC you are using, and THAT he completely skips over!

    Would be nice to open up his Emagic EMI 6|2m and see what kind of DAC is in: oversampling/PDM/Delta Sigma or whatever, and make another video which explains this process for the casual viewer, not brush it off as if its a non issue.

    He only briefly mentions Zero Order Hold inherent to ‘older’ DACs, without any precision. Its not that far ago! Lots of CD players and equipment (including many Digital Synths and FX) used Burr Brown PCM53/PCM54 chips and needed an analog reconstruction filters to remove those stairsteps. Every DAC has its own qualities, and signal/noise ratio.,

    There is a reason people still think stairsteps are there and in some cases (including playback on pretty much any early 80s’ vintage stuff), its still there.

    • just passing

      I suspect the reason people still think in terms of stairsteps is that that’s what’s shown in pretty much every single representation of a digital waveform. We do tend to believe what we see…

      As to DACs, the simple fact is that the lousiest sigma-delta DAC of today will beat the pants off all but the verymost expensive DACs from 30 years ago – and except for the Synclavier, those DACs just didn’t make it into synths. The current fad for NOS is, as far as I can make out, audio woo based on an absolute failure to understand the concepts involved and an insistence upon magical thinking.

    • just passing

      Further to my other reply, he was actually completely precise – just not entirely accurate. The “zero-order hold” signal is, as he says, the result of an *incomplete* digital-to-analogue conversion; the conversion has been done, but the reconstruction filtering has not.

      Except that what he doesn’t say is that the process which converts an amplitude-coded impulse into a zero-order hold signal, usually a sample+hold circuit, *is* a reconstruction filter. It’s just *lousy*; it’s an all-pass filter where a really steep low-pass filter is required. (Of course, it won’t be a perfect all-pass filter, and if you zoom into the waveform it outputs closely enough, it won’t be a perfect stairstep either; the Gibbs effect will always be found at *some* level of zoom.) But here’s the thing: it’s STILL not going to be audible on its own, because all the stepping is STILL going to be above Nyquist – and at 44.1kHz, above the human hearing threshold. (That’s not to say it won’t have an effect on what’s within that threshold, though; it may not be audible itself, but it can still wreck a speaker if the amp doesn’t properly filter it away – a phenomenon which will quite certainly be audible!)

      There is one instance where it might bleed into audibility – if the clocking of the DAC isn’t jitter free, it will manifest as audio noise. Oversampling helps with this, in that the faster a clock is, the farther away from audibility the sidebands of its clock jitter will be. But that’s not the primary rationalisation for oversampling…

      So as regards oversampling and sigma-delta conversion: Steep low pass filters, especially steep lowpass filters that leave your signal’s phase alone, are really easy to design in the digital realm; in the video you can see Monty switch in a filter with a 100dB rolloff at 21kHz. (That filter might be a FIR hundreds or thousands of steps long.) But a lowpass filter that steep would be somewhere between utterly impractical and practically impossible to create with analogue circuitry, and god alone knows what it would do to signal phase. The answer is to use simpler, shallower analogue filters, oversample at the point at which those shallow filters have sufficiently bandlimited the signal, and then use digital filters – which are simple enough to be spun into hardware, even, which is why the first oversampling DACs were designed to be paired with their companion digital filter chips – to brickwall filter the signal into an absolutely bandlimited form. And what he said about noise shaping can be expanded upwards to apply to sigma-delta convertors, too; it’s a gross oversimplification, but if you can use noise shaping and dithering to distinctly encode sounds at 1/4 of a bit, you don’t necessarily need to stop there – if you oversample enough (at frequencies into the lower MHz), and noise shape well enough, you can reveal sounds down to an arbitrary 2^(-n) bits – and all your noise is still shunted off to the point where a very simple analogue reconstruction filter will get rid of it. (The principle behind dithering is exactly the same principle a cat uses when it wiggles its butt before pouncing on its prey; the movement reveals additional detail that its eyes aren’t precise enough to see in stasis. I’ve been known to call it bit-wiggling.)

      Anyway. This is all stuff you probably know already… but if someone doesn’t, I hope it helps.

    • plgDavid

      This is indeed the kind of stuff that the video should have explained because, as the narrator said, the stepping was the main question people originally had, and his reply IMHO didn’t address it enough (at least give pointers to sigma-delta/oversampling DAC evolutions).

      Yesterday I was searching for a document that would give some sort of “Timeline of DAC designs”. While this is not quite it, the closest I could find was http://www.ti.com/lit/an/slaa510/slaa510.pdf

      In any case my own research (resisting a shameful plug for my latest product here) has been focused primarily in the low end methods to encode digital audio because I like having a variety of dirt to choose from.

    • just passing

      *gets around to googling plgDavid*

      Ah. Yes. Erm… yes, you will have known all that stuff, then. Probably rather better than me. Um… *tries very hard to be invisible*

      I’m guessing this is the product you mean: http://www.plogue.com/advanced-bitcrusher-speaker-simulator/ I’ve not used it at all, but it certainly looks fun. :)

    • David Viens

      I just done the opposite, full visibility, lol

  • Alvaro Sandoval

    great vid

  • KID

    I´d love to see this guy explaining the difference between lossless and encoded audio files like mp3s.

    • lala

      i’d like to see the stupid comments on that one 😉

  • Will

    I think it’s always been interesting that one of the reasons so many prefer analog technology is precisely because of its inability to perfectly recreate sound. The mushy blurry goodness. Similar to why people love Instagram filters. :)

    • Graham Metcalfe

      Except those Instagram filters are digital…just saying. And did I really want my photos looking like they came from my Instamatic from the 60s? The only thing I like about my Instamatic pictures from that time is that they are FROM that period of time.

      I’ve never had access to really fine analog recording equipment, and most of us never will, so I’ll take my digital recording rig and skip the PITA that working with tape, noise reduction and all the rest entails. The reality is that most digital systems of low or moderate price will handily beat out low or moderately priced analog recording solutions in terms of sound quality. And yes, I’m old enough that my learning time was spent splicing tape, recording on modular analog synths, etc.

    • PaulDavisTheFirst

      actually, if you watch the video again, you’ll see that even a low priced digital system will far outstrip even a costly analog one in terms of sound quality, unless by sound quality you mean qualitative terms like “warmth”.

    • Graham Metcalfe

      Yes, exactly. And to really get that warmth you are talking about a huge investment in the finest of analog consoles, decks etc. and really for most listeners it just doesn’t matter when they’ve ripped the latest tune from a you tube video and are playing it back over the tiny speaker on their cell phone.

      This was a great video. Succinct and informative.

    • bilbo

      Seriously everyone looks at it in a financial/cultural/technological aspect than an artistic aspect (Which is kind of a big part of recording imo). It’s about the craft and being versatile with your skill. The way you talk is just so depressing. This is the reason people don’t understand what good quality audio is all about, because they have no idea what goes into making the record and the technology being used to play it. A person’s home theater was expensive, but it’s what they wanted to do with their money despite the fact that a regular LCDtv gets the point across just fine for less money. For me, I dream to learn how to use a multitude of recording equipment because AE is something I like and I will bust my ass trying to fund it or at least die trying. And for me, the feeling of accomplishment will be worth the journey. can tell you to use your digital gear to try to get a sound just like the beatles (And they did write their shit down) but it won’t because every little thing matters right down to the atom.

    • Graham Metcalfe

      Well, I don’t think its depressing really. I think there is always room in certain markets for the high end of analog. I think orchestral recording and film audio are good examples where you have an incredible dynamic range and subtlety and nuance to the sound. There was a very interesting video about the production of the soundtrack or The Incredibles that talks about the increased dynamic range and nuance to the sound of tape when recording brass that I think is very relevant.

      However, when talking about recording modern idioms (DNB, Dubstep,”EDM” etc) I think that nuance and subtlety imbued by analog gear becomes less important because the overall quality of composition and timbrel mix of the programming is less subtle and dynamic by nature.

      The digital domain is an incredible boon to folks like me who are primarily concerned with the compositional process. That doesn’t take a way from applicability of high-end analog audio in the appropriate circumstances. It’s just a bit irksome to have people whine that analog is just soooo much better, and that people working in the digital domain are just a bunch of hacks.

      For me it really just comes down to the quality of the composition…Beatles songs are just as beautiful sung and played on an acoustic guitar sitting by a campfire…you can’t really say that about Skrillex.

    • foljs

      “””Except those Instagram filters are digital…just saying.”””

      Beside the point, they are used to recreate analog phenomena. Like “tape saturation” DSP.

      “””And did I really want my photos looking like they came from my Instamatic from the 60s?”””

      You might not, a lot of people do.

    • Bilbo Baggins

      Dude, this is like saying I’m not going to try to emulate the sound of vintage analog gear for my led zepplin tribute band because I’m not recording in the time period from which they came. Of course there are a lot of practical uses for recording technology but one of the dominant ones is music. You’re not taking a family photo on christmas, you are producing art in a tangible format. And then you go on to say that digital is cheaper and easier to use and yet you are old enough to use tape. Wouldn’t working with the gear that is more of a hassle to use be a better tool for learning? People seem to learn best when facing a challenge. I’m not saying one is generally better than the other because of course they both have their strengths and weaknesses. I honestly would feel more accomplished after undertaking a seriously difficult challenge than just taking it easy. The only real issue here is prices IMO which is honestly pretty funny since there is probably more demand for digital gear rather than analog because (And I am quoting my generation…) “old technology is for old people”.

    • Graham Metcalfe

      No, I think emulating that sound is cool. I’ve done that with live recordings…running it through tape emulation to add warmth etc. I have to say your generation has definitely got it backwards…most of the folks in my generation have fully embraced the new stuff. Having lived through the old crap I’d never willingly choose to go back to splicing tape manually and dealing with the headaches of noise reduction etc.

      There is something to be said for suffering through the limitations of old technology. Learning the process from the ground up. For a lot of people learning the process sometimes the new tech makes it a little too easy to just turn out a lot of mediocre stuff without putting thought behind the process. Part of that is just applying discipline and patience.

      For me, I’m coming it at from a different direction than you are. I’m very focused on composition more than engineering so I tend to look at what helps expedite and ease that process. I found that working in the digital domain, and mostly “in the box” really simplifies the process of realizing a composition. I don’t miss trying to manage a studio with a whole lot of MIDI modules
      and having to remember all the interconnections every time I wanted to
      work on something.

      Sometimes I like to go really analog and write stuff onto notation paper. And I also find that working within the limitationns of trying to record with an iPad refreshing.

  • ZooTooK

    Wonderful video!
    Keep in mind he’s explaining digital to analog processing but he’s not explaining how analog recording equipment is producing non-perfect signal processing adding noise an non-linear behavior at different signal levels and and timescales that many people find symphatetic and pleasing. Just using straight ADA sound cards will then not sound the same as using analog mixdesk and tape recorder unless those “distortions” are added in the digital domain, which they are more and more the past 5 years. This video tells us that digital recording will be as good as the analog emulation algorithms invented.

    • PaulDavisTheFirst

      monty isn’t discussing “good” in this video. he’s discussing the basic misconception that you cannot reproduce an analog waveform from a digital representation of that waveform.

  • Use your ears

    If there is no difference, then why do digital only devs spend millions of dollars on emulating
    Something that does not exist?

    • Space Captain

      That makes no sense. They are clearly emulating something that existed and that people have an attachment to for various reasons. (e.g. analog synths). Companies need to sell products that are desirable in order to make money, pay their employees, do R&D, sell more stuff, repeat. See “capitalism” for more details. They use digital because it saves them money and chips can be used across various products. As an example of a multi-use chip search for “SHARC” processors which are used by many companies in their product development. Digital in not inherently bad. Results depend on the quality of the algorithms which are usually determined by programming skill and development budgets. See “ValhallaDSP” for some skillful programming of reverbs.

    • Use your ears

      That was a long exercise in distraction, was that question too simple?
      Why are people spending allot of money on the difference between analog and digital if there is no difference?

    • PaulDavisTheFirst

      i think you need to read some contemporary economics, particularly stuff oh human decision making in the economic realm. there is absolutely zero reason to think that customer choice expressed via “spending a lot of money” has any real relationship to quality, or even the existence of difference.

    • http://mboverdrive.tumblr.com/ Ifthenwhy

      “Why are people spending allot of money on the difference between analog and digital if there is no difference?”

      Because in “commerce”, there is absolutely a difference between analogue and digital: it’s called emotion, and industries are built on it.

    • foljs

      “””That makes no sense. They are clearly emulating something that existed and that people have an attachment to for various reasons. (e.g. analog synths).”””

      Yes. Those things are not the same things as what he discusses in the video (which are about recording and reproduction).

      He didn’t say “analog synths” are the same as a “digital synth”.

      He said that a ***recording*** of something (e.g an analog synth) will sound the same, whether the recording is digital or analog.

      Consider a piano. There’s no difference between a tape or a digital recording of a piano performance (some artifacts aside). But, still, producing an instrument (sampler etc) that sounds like a piano is a million dollar business.

      It’s another endeavour altogether — don’t conflate the two.

    • just passing

      Oh dear.

      They’re not seeking to emulate the vagaries of analogue *reproduction*. They’re seeking to emulate the non-linearities of analogue *processing*, which is expensive because introducing non-linearity to digital processing means having to worry about the Nyquist limits, and at which point factors like the effects of component tolerance and sensitivity to heat kick in, and feedback delays are limited only by the speed of electronic transmission in a wire.
      Of course, the more obvious answer to your question is: because idiots like you will insist that it does exist beyond all reason, and ultimately you have to build what the customer asks for, even when the customer is a woo-spouting pillock.

    • http://mboverdrive.tumblr.com/ Ifthenwhy

      “…because idiots like you will insist that it does exist beyond all reason”

      “Use your ears” is actually asking a very reasonable question. In some ways it’s the most important question of the A vs D “debate”.

      Perhaps refraining from calling him/her an”idiot” would help remove the unnecessary hostility that this topic invokes?

    • just passing

      I took a flier based on the writing style and quality in evidence. Shoot me.

    • foljs

      What they are emulating are artifacts that DO exist (saturation, distortion, wow, flutter, etc).

      Those are not in the items he discusses as being no difference in…

  • why so unserious?

    25 minutes to make up some crap being wrong in the simplest basics already? sorry, waste of time….

    • PaulDavisTheFirst

      name one piece of “crap” that you think monty gets wrong …

    • foljs

      Your comment is the biggest waste of time on this thread. And content-free to boot.

  • Sharon Kathleen Johnson

    I notice that changing tempo, volume, pitch, and EQ all change my sound quality somewhat.

    • just passing

      Yes, that’s because you have to do more than simply reproduce the recorded audio at that point. DSP *is* difficult. Resampling – the essential process behind pitch shifting – is impossible to do perfectly, so we settle for compromises, and historically those compromises haven’t always been that good; the result is that playing a 44.1kHz file at 48kHz will inevitably result in a sound of lower fidelity. Likewise, EQ is filtering, and filtering is again a compromise. Changing the tempo of audio without changing its pitch is even harder; even the best, most computationally expensive methods (phase vocoding, for example) will smear transients across a few milliseconds.

      The odd one out? Volume. Human perception reliably equates louder with better, whatever the source and even for very tiny increments, and that in itself is enough to render objective determination nigh on impossible. The effects of simply multiplying a signal by a relatively constant attenuation factor – *the* most transparent thing you can do to a digital signal – will absolutely pale into insignificance compared with that.

    • just passing

      Actually, I’m wrong here. You *can* convert 44.1kHz to 48kHz perfectly, as long as your signals are *exactly* 44.1kHz and 48kHz in frequency – or more precisely, that the signals’ frequency ratio is exactly 147:160. The principle is that you use an interpolation FIR to exactly insert 159 samples between each original 44.1kHz sample, maintaining the bandlimiting perfectly, and then simply drop 146 out of every 147 samples that come out of the other end (which you can do because the signal remains bandlimited to 1/160 of the FIR’s operating frequency).

      But when you start thinking about how to generalise that to arbitrary sample rates, you begin to see why resampling is quite difficult unless you limit yourself to easy ratios. (Which, to link it back to the discussion above, is why oversampling is always done at an integer ratio to the desired sampling rate, and preferably a power of 2.)

  • itsnowhipsternoottoseecolor


  • adam cote

    Thanks again, Peter, for another informative, geeky video! So simple it seems to have confused some here.

  • Mark Eats

    Super-interesting video, and very clearly presented, hopefully he does some more.

  • Random Chance

    The third bullet point of the article’s summary is wrong (as well as unfortunately phrased, e.g. signals can not be used to store or reproduce anything, they are stored or reproduced if we so choose). Perfect sampling and reproduction is only possible under the conditions mentioned in the video. I’ve only watched the video because I didn’t think that somebody would be dumb enough to put out a half hour video just to claim things that are physically and mathematically impossible just by omitting the assumptions made by the sampling theorem. If we were in school, you wouldn’t get full marks on this summary, Peter. :-)

  • Ken Steiger

    Does this mean I can finally feel guilt free in my belief that vinyl is not inherently “superior” to digital?

    • lala

      digital is true stereo – vinyl isn’t

    • just passing

      Again – wat

    • lala

      vinyl can’t do bass from the left then bass from the right – the needle would jump of the record

    • foljs

      No, but it can be argued that vinyl was culturally superior to digital:
      — the patina (crackle etc) was part of the aesthetic process
      — large covers that often featured masterpieces
      — tactile feel and tangible record collection
      — more difficult to skip around
      — more difficult to treat a vinyl record as the one in 10,000 pirated downloads you’ll never get around to listen to

    • Robin Parmar

      There is no such thing as “culturally superior”, only your preferences. The “patina” was not part of the aesthetic process, which presumably the artistic agent instigated. It was an unnecessary limitation of the medium. Only much, much later did it become an object of fetishistic fascination amongst certain recording artists themselves. (But, yeah, I love Philip Jeck.)

      The large album covers much more often featured visual atrocities than masterpieces. There is as much good design in digital files, cassette covers, etc.

      The way one treats one’s music is entirely a personal choice. I tend to get one download album and listen to it as an album over and over… much as I did vinyl. I do not feel limited by digital files or coerced into listening any way other than how I choose. Then again I have never purchased a crappy Apple product that would so dictate my listening process.

      It could well be argued that the relative rarity of a vinyl release (outside of top 20 products) and the difficulty in distribution only created false castes of listeners based on access and privilege. This was hardly an egalitarian outcome, or one that was about “the music” in any valid way.

      The tangibility of my record collection meant I abandoned it 12 years ago when I moved across the Atlantic. I did take the essential music however, on a single hard drive. This didn’t reduce my love for the artists and their creations one iota.

    • Bilbo Baggins

      Dude you’re wrong. In a way there is. Asian cultures were culturally superior to others at one point because they had a concept of hygiene while all the other civilization thought good hygiene was watch-craft. Modern digital downloads and (Especially!!) Mp3s are cheap plastic goods compared to vinyl. CD’s are technologically superior to vinyl because: 1. CD’s (And CD players) are more accessible/common right now 2. the sound a CD makes while you are listening to it is around -10dB and inaudible to most people while vinyl makes sound that is audible at 70db (hiss). Mp3s are probably the worst format for digital music to exist yet at the same time it is the most popular among our society. It is a part of our culture and it is kind of a bad one at that. I used to use Mp3’s until I discovered Lossless files and unless you are from a big city and have hearing problems, you CAN hear a HUGE difference because Mp3’s eliminate a large percentage of the original file for the sake of hard drive space. It made me appreciate the composers I love as well as giving me a glimpse about how serious AE is. Mp3’s are like a cheap agathis beginner’s guitar VS a 60’s fender stratocaster. Id say the stratocaster is DEFINITELY more culturally significant than a shitty cheapo guitar someone might buy at Walmart for $80. Hell the first electric guitar is more culturally significant than the Stratocaster.

  • http://melodiefabriek.com/ Marco Raaphorst

    we push everything through speakers which can only resonate in sequence. going back and forth in a certain speed. they are monophonic. and a great way to sum many audio channels :)

    I love digital. and we’re getting better at tweaking it to achieve mind blowing sounds. I’m glad to living now.

    thanks for the article, video + extra text. really good stuff. I learned a few new things and still believe in one thing: let your ear be the judge. it all comes down to listening.

    • just passing

      Not that great a way; if you sum audio channels in a mixer and they all peak at once, the voltage is capped by the power supply voltage. If you do so in a speaker, the cone bounces off the magnet. Which is, er, sub-optimal. :)

      I still can’t get over the fact that when all’s said and done, we’re pumping sound into an organ which basically depends on frequency-sensitive tendrils resonating. That’s some strange dark magic there. In fact, given that the ear is by far the least reliable component of any audio system, I can’t wait until we can finally bypass it altogether. 😉

    • http://melodiefabriek.com/ Marco Raaphorst

      hehe 😀

  • Will Mayner

    This is totally baller. Rare to find such good exposition.

  • Joshua Winget

    I’m an audio professional (music editor and writer/producer) with 25 years experience in digital audio. He’s really missing the point. Technically, yes, the conversion back to analog produces a sine wave that looks identical, but there are VAST differences in the quality of the sound of that conversion. By his logic, you’d get the same fidelity from the headphone output of your laptop as you’d get from an Apogee Symphony DAC. Anyone with even a slightly critical ear would know the difference in a second.

    • just passing

      How can he be missing a point that as far as I can see he never sought to address?
      And can you please quantify exactly which part of the difference in sound between an Apogee DAC and a laptop output his basic explanation of the Nyquist theorem and its more interesting negates?

      I’ll tell you something – the Apogee DAC will work in more or less exactly the same way as the laptop DAC – or, perhaps more to the point, a random USB soundcard that also works at 24/192; they’re out there. The difference in sound quality is WHOLLY explicable by the difference between soldering the output of a 50c codec straight to a 3.5mm jack and hoping for the best, and carefully designing and arranging a few hundred dollars’ worth of discrete components into the most stable, transparent configuration possible; there won’t be anything radically different about the way the DAC converts, only about the quality of its components. And that’s all on the analogue side – and therefore explicitly beyond what Monty is talking about in this video. Which you’ll be able to verify the next time you’re next to an Apogee Symphony, by turning it down to 16/44.1 (if you can) and noting how very much better than the laptop output it *still* sounds.

    • just passing

      Here we go – a random USB soundcard that can do 24/192.


      Nobody would rush to give up their Symphony for this, and surely nobody could fail to tell them apart in an ABX. But as a straight stereo DAC, its digital specs are the same, right?

    • lala

      i wouldn’t be so sure of that
      there was this listening abx test that compared some 4000$ ad/da to a soundblaster thing
      i prefered the soundblaster ad/da to my big surprise
      and I’m not deaf or something

    • just passing

      Yes, but you could tell them apart, reliably. That’s all an ABX does.

      After that, it’s probably not surprising you preferred the Soundblaster. Like hi-fi equipment vs studio reference monitors, a Soundblaster is probably designed to flatter sound, whereas the $4K a/d/a will be designed to reproduce exactly what it’s given in unflatteringly accurate detail.

    • lala

      yes i could tell them apart
      but that doesn’t say a thing about the perceived quality of the audio
      and i don’t care if i can see a better s/n ratio in an analyzer
      because i won’t hear it

      after that i sold the apogee duet and got a 100$ soundcard…

    • just passing

      Yes, but you’ve already said two hopelessly dim things in these comments. That this is the third, at this point, isn’t terribly surprising.

    • lala

      lol, prove me wrong

    • Concerned Webizen

      Lala – Please go away. You are dragging down this whole topic with your inane trolling.

    • lala

      no, i won’t shut up, there is to much bs around
      besides your comment doesn’t add anything to the conversation
      so what was your point again? &_&

    • lala

      check this out, then we`ll talk again


    • lala

      ah right
      so u can hear noise @ -70db while something else is playing with -10db, rofl
      are u dumbo? u know the elephant with the big ears …

    • lala

      thats a lot of bs
      so u say the soundblaster sound card is more linear than the expensive stuff lol and because of that the expensive stuff is better because it doesn’t have a flat response
      did u eat a clown?

    • foljs

      No manners, no grammar and no punctuation.

      Overuse of inane internet catch-phrases like “wat”.

      No understanding of basic audio/DSP science.

      You are below twenty years old right? (At least in mental development).

    • lala

      sorry to hurt your delicate sensibilities
      so what did i say thats technically wrong?
      ah i c u cant tell &_&

    • lala

      and i never wrote “wat”
      thats not my style

    • lala

      u cant compare a headphone output to the output of a sound card
      they are different

    • just passing


    • lala

      gee, do i need to explain everthing to u?
      look @ the sepcs of the things then think for yourself

    • foljs

      “””He’s really missing the point. Technically, yes, the conversion back to analog produces a sine wave that looks identical, but there are VAST differences in the quality of the sound of that conversion. “””

      If it looks identical, the quality of the sound will be identical.

      It’s science, not magic.

      “””y his logic, you’d get the same fidelity from the headphone output of your laptop as you’d get from an Apogee Symphony DAC.”””

      More or less yes. It’s the preamps that make the difference.

    • lala

      im repeating myself:

      u cant compare an headphone output to the output of a sound card

      > impedance – distortion

      does that ring a bell? no?

    • foljs

      “””u cant compare an headphone output to the output of a sound card”””

      u (sic) very much can.

    • lala

      yeah and i can compare the number of nonsense comments i read with how much aspirin i have to take because of that

    • PaulDavisTheFirst

      no, he’s not missing the point. you’ve missed the entire point of the video, which monty explained in the first couple of minutes of the video. what was the point? debunking the claim that you cannot reconstruct an analog waveform after digital sampling (a mistake inspired, he believes, largely by the “stairstep” representation of the results of sampling).

      as for the quality differences between DACs, that is explained very trivially by the quality of (a) the clock (b) the analog components, neither of which are particularly relevant to the point monty was trying to make.

      in addition, i’d also point out once more that double blind tests don’t support your belief that “even a slightly critical ear would know the difference in a second” (assuming that sample rates and bit depths are the same).

  • Marc Doty


  • heinrich zwahlen

    The adulation of analog sound today is mostly retro hype by kids with too much money to spend on equipment. I bet most of them would not even hear the difference to the good digital sound in a blind test This comes from a guy who has used a lot of (expensive) analog gear for the last 30 years…the only thing i could really miss at this point would be the sound of tape.

  • Chris Stack

    I was chatting about this video last night with Paul Vo, inventor of the Vo-96 Acoustic Synthesizer (intense DSP controlling real-world guitar strings) and the technology behind The Moog Guitar (100% analog). He had some interesting observations:

    “This is exactly the kind of issue to provide endless debate… yes of you don’t do it right it doesn’t work right. And yes in the old days they did it wrong and people heard things that formed opinions which persist unto this day. I have two of those emagic blue boxes. They used early sigma delta conversion. Perfect. All the bad is traded for about 1500us delay analog to analog, that’s the only prob.

    I think the difference between Digital and analog lies outside of the box entirely. When you design with analog you are designing with components that individually have complex behaviors they are larger grained building blocks. As such the analog design process tends to guide your thinking towards what those blocks want to do. This imposes conceptual limitations that tend to lead to a certain class of result.

    Designing with digital pretty much removes these limitations. You can do anything and that can be a problem. You get a kind of blank sheet paralysis sometimes, whereas with analog you have more limited choices and it’s easier to move forward. Digital is harder in that sense; you have to think very clearly about the vision you want to achieve because if your vision is distorted, digital lets you achieve it.

    Analogue tends to correct that for you because of the meta-behavior of the building blocks. Not to say you can’t get it wrong too, but the errors tend to be more gross and therefore more obvious and correctable.

    All of this leads to the creation of things with perceivably different characteristics. People favor analog still for good reason, The cause is inherent in the characteristics of engineers not in the techniques being used for the engineering.”

    • Karl Weiß

      Now, it’s getting interesting for me. Thank you!

  • Frydac

    I learned all these concepts in school, nevertheless I really enjoyed this, thx!

    Though I think for the uninitiated it will still be hard to really understand as he uses some concepts without explaining. (just watched his other video.. he eplaines them there..)

  • Yanakyl

    I came to the conclusion that those are two different levels of the same thing.

    Where analog is discrete to the level of molecules of air, or electrical charge (I am making that up now, but that’s my perception) and digital is a higher level of abstraction of the signal.
    When you make abstrations you can control the information in a new way, like when you create words that explain complex concepts, you can then easily talk about those concepts interacting with each other with just a few words. Some thing like that.

    More importantly I get this sensation that buiding with analog is a bit like modeling clay, you get this electricity and try to shape it from “outside”, while in digital you tend to create your signal from “inside”. So it ends up being somehow a difference of “point of view”.

    Then everyone get his own point of view of those and you get a mess!! 😀

  • gLOWx

    And once again, explained or not, it is going to be a mess…
    Because most ppl don’t understand…but even more because ppl want to loose their life on internet arguing on useless things.
    And that’s why i stopped using KVR forums : a mess of useless posts.

    Make or listen to music on everything you can get your hands on. And skip those useless analog vs digital, hardware vs software… “debates”.
    Informations are already accessible for decades, nothing is new.
    Until you just want to loose your precious time on Earth…

  • http://melodiefabriek.com/ Marco Raaphorst

    I’m doing some tests. I’ve recorded something on -23 db and it’s super easy to hear the dithering noise. And without dithering I hear quantisation noise. Super obvious.

    I also hear differences between noise types.
    Will write about this on my blog soon. It’s very nerdy stuff but it is obvious.

  • Claudio Ferrari

    One of the bests videos I’ve seen about AD conversion. And your text about it too! Simply amazing. Thanks a lot.

    Cheers from Brazil :)

  • brianmoore

    Wat kitty haz cheezburger (sp) – Inane internet phrases aside, I LOVED this video! Very clear, and to the point – like engineers tend to be. I learned stuff for sure

  • squeem

    Analogous means similar, not exactly the same. You used the world literally incorrectly. Analog Sound is “Similar Sound”, not “Just like the actual sound that was recorded.” Digital Recording sounds even more similar (in it’s replication) to the actual audio as opposed to Analog, but of course it too is not exactly the same. How can something be similar, but not literally?

  • squeem

    A big factor as to why analog gear is more expensive is because I’m thinking a good amount of the gear people want is vintage. Since this gear is vintage, it has the benefit of being made in a good time period where things were durable and built to last. Now, it is more productive to sell cheaper products that completely shit out after some time so the customer has to come back and either get it repaired or buy a new one. Make more income that way. Analog gear being produced today may still retain that durable quality as well. Of course this is all conjecture so please do not take what I am saying to heart and start the name calling and ad hominem.