I’m sure you’ve heard things like this before; it’s a point of view heralded by young hipsters and aging hippies alike — you may have even said it from time to time. “Nothing beats the sound of vinyl.” “Vinyl just sounds warmer.” “CDs don’t get the sound quality of an old-school record.”
If you haven’t said this before, you might be quick to dismiss it: they’re either pining for a past that never was or hating something just because it’s popular (much like your humble author and her aversion for all things Harry Potter). Is this true? Is it false? I WISH THERE WAS SOME PROCESS WE COULD USE FIND OUT THE QUALITIES INHERENT TO VINYL AND DIGITAL RECORDING THAT MAKE THEM DIFFERENT!
Oh wait. Science.
The main difference between CDs (and MP3, and AAC, and WAV, and silly, silly WMA) and vinyl is that the former is digital and the latter is analog. It’s very similar to the difference between a digital photograph and an analog photograph: the digital version takes a million different tinier snapshots of the sound or of the image, then recreates them in an entirely different format to the best of its ability. A crappy camera shows that it’s crappy by pixelating the image; a crappy recording shows that it’s crappy by cutting out or sounding distorted. In audio, this is a sign of the particular devices’ sampling rate.
Analog, on the other hand, makes an almost true-to-life recreation of the sound or the image. Analog cameras essentially blast the light coming from the lens onto a bit of film. There’s the image. Similarly, the grooves in a record are the waveforms that were recorded — not a series of snapshots or samples, but a true recreation of the sound the recorder heard (think about the tickertape on a seismograph. The needle messes with the line on the paper as the vibrations affect its position. Same deal).
Therefore, the difference between digital and analog recording is the simplest and perhaps the biggest difference between the sound of a CD and the sound of a record. No matter how high a digital track’s sampling rate, it cannot get the true-to-life recreation of sound that vinyl does. This isn’t the end of the matter, though: there’s the tendency for a record’s dust to add scratchiness to the sound, the fact that the physical act of a something scratching a material year after year begins to degrade that material, a fluctuating temperature’s effect on vinyl, and the fact that the heavy-handed nature of a phonograph needle often leads to a “roll off” of the high frequencies (leading to that “warm sound” people talk about). In these aspects, digital wins.
But wait, there’s more!
It’s called the loudness war. You know how what TV considered scandalous in the 60s was a bit of midriff, and for something to be considered scandalous on TV now, it needs to be a full-fledged mammary gland? Whether it’s the audiences who demand it or the content creators who shove it down their throats, time has made the stakes higher for everything. Music is no exception, of course.
When a series of songs play on the radio, record executives want their own songs to stand out. Loudness is a great way to do just that. Make no mistake about it: early on, they were doing this to records as much as they could, but there were limitations. As a result of vinyl’s innocent analog disposition, it has a variable limit for how loud it can get, one which depends on a variety of finicky factors. Digital, on the other hand, has all sorts of tricks up its sleeve in order to turn up the volume. Once the loudest parts of a digital track have reached the file’s loudest possible level, audio engineers can still do more to make the whole track louder. Often, this involves dynamic range compression: bringing up the quiet parts to sound louder. Sometimes, nearly as loud as the loud parts. Sometimes, just as loud.
As CDs became the format of choice, more and more songs started using compression to sound louder. As the technology got more advanced, the dynamic range (the quiet quiets and the loud louds) got blurrier. Until now, when you can hardly hear quiet sections in any of your music. This is why we can wear earbuds and never adjust the volume, sure, but this is also why we can listen in our living rooms and never be blown away by a glorious, climactic musical passage.
This video demonstrates the difference perfectly:
Though there are many facts that objectively determine each format’s strengths and weaknesses, this is by no means the end of the debate. Which do you think is better?
Photo Credits:
Blue Vinyl by Peter Organisciak
Pixelated Pink Ladies by Clyde Robinson
We Can Only Turn the Volume Up by TheGiantVermin
So, I disagree with this article (other than the loudness war, where I’m in full agreement. Dynamic range compression is a terrible thing.)
Firstly, analog devices do not necessarily make a “almost true-to-life recreation”. Noise in the input chain (microphone, cabling, recording equipment) can all conspire to reduce the fidelity of the recorded signal. There are other limits as well, such of that of the physical media. Cassette tape, for example, despite being analog is a really poor way to record music.
Secondly, you can chop up a signal and get the original out – it’s called the Nyquist–Shannon sampling theorem. It’s why CD Audio is 44.1 kHz – it’s just about twice the frequency of human hearing. There are caveats to this, of course, but the idea is that you can also get the “almost true-to-life recreation” that you argue only analog media can perform. Now, digital can be poor-sounding (8-bit, 22kHz audio is hardly high fidelity – but we’re talking CD Audio or modern digital media, which is usually 16-bit, 44.1/48 kHz)
Given (1) and (2), I think I can argue that you could make a digital signal equal that of an analog signal. At the very least, you could do so to the limits of human audible perception, which is what really matters here.
For the comparison of vinyl and CD … things aren’t so good for vinyl. Its dynamic range is inferior (typically around 80 dB, compared to around 90 dB for CDs). There’s also the rolloff you mentioned, which seems endemic to vinyl. Ironically, it is the superior fidelity of the CD that permits such terrible things as compression to be done to it! Vinyl is literally not good enough to be compressed!
To be fair, Phong, it sounds like you simply disagree with my claim that vinyl gets a true-to-life recreation and digital does not. The other things you’re saying are exactly what this article says.
You’re also not backing up some of these claims: You’re saying analog recording does not make a true-to-life recreation, yet you explain this by faulting the other parts of the playback chain that have nothing to do with the recording itself. I’m also curious as to what exactly makes a cassette a bad way to record music, seeing as it was what was done for decades.
I would elaborate by saying that an analog recording simply cannot, by definition, avoid degradation due to faults in the playback chain. By attempting to maintain an isomorphic representation of an audible sequence in a different medium (and then translate it back again), it must necessarily expose itself to noise and entropy. A digital recording has the advantage of minimizing, as far as possible, the time (on both ends) that the representation is in analog form, and so massively decreases the opportunity for noise to be introduced.
I’m not a musician, but the people I know who hated recording music on cassettes generally complained about all the extra degradation that they were vulnerable to (e.g. print-through, drop-outs, stretching, breaking, ambient magnetic noise) as compared to digital recording. Also, when you compare a cassette tape to a typical 1/4-inch reel-to-reel recording at 7 1/2 ips, a cassette offers roughly one-eighth the magnetic real-estate for recording, which is reflected in their overall fidelity.
Just realized I equated cassette with magnetic tape back there. I s’pose when one says cassette, they mean those cheap little things people used to stuff their gloveboxes with, not the high-tech device my band director had me record All-State auditions on back in the early 2000s.
Seriously, you guys, I love these comments. I started out researching this topic without knowing much about it, and every dissenting viewpoint makes me learn that much more.
“By attempting to maintain an isomorphic representation of an audible sequence in a different medium (and then translate it back again), it must necessarily expose itself to noise and entropy.”
Could you explain that further, breadbox?
There are a few invalid assumptions in this article. But first, a cultural note.
Jonathan Berger at Stamford has been playing various music formats to students for years. Basically people prefer what they grew up listening to:
http://www.macworld.co.uk/ipad-iphone/news/?newsid=25288
In the post: “Analog, on the other hand, makes an almost true-to-life recreation of the sound or the image.”
Unfortunately simply being “pure analog” does not mean something is better. The phone system was analog for years; it clearly sounds worse than an LP due to bandwidth constraints. FM radio is analog, so is AM; which sounds better? Two cans and a string is analog. Is that better than a CD?
If we’re talking about recorded music, it’s important to consider the full signal path. Nothing sounds close to sitting in a room with a great musician and instrument so let’s ignore that particular (I hate to use the term on this site, but it applies)… “magic.”
Typically music is now recorded digitally. This is sort of amazing in that you can put a mic in front of someone singing and playing guitar, play it back in the studio, and ship the EXACT SAME NUMBERS on the CD that were on the hard drive in the studio.
In contrast, going from analog tape to LP to cassette requires lots of additional electronics — just to make the stuff “stick” to the different format. Each step degrades the original.
Lots of tricks are performed with digital files as well — but no analog medium is able to deliver what the musicians hear to your home like digital can.
But… is a CD copy of a Beatles master recording better than the 1960s original LP record? Frequency response, channel separation, transient response, signal to noise, dynamic range? All vastly superior on CD.
Will you prefer the emotional experience to what you heard sitting in a lounge chair as a kid? Possibly not. There’s more to life than specifications.
Is the output measurably better along every scientific axis? Yup. The problem isn’t scientific measurement, it’s the emotional component of “better.”
Not sure which part you want me to elaborate, but okay.
So we start out with a very complicated pattern of air waves. Some of those waves hit the transducer of a microphone and vibrate a magnetic element of some type, which results in a pattern of electrical impulses. This pattern won’t perfectly correspond to the pattern in the air waves, naturally. Still, it’s pretty darn close.
If this is a digital recording, then this pattern will then be converted from analog to digital, which will then be written out to digital storage. This digitized pattern is again an imperfect representation of the original, but it has the advantage that it is much easier to work with without introducing further imperfections. No doubt this pattern will be intentionally altered further (changing the volume, adding reverb for ambience, all the things that people do in recording studios). Since they’re working on a digital representation of a pattern and not the pattern itself, these manipulations will also be imperfect, but (unless they’re done wrong) the imperfections shouldn’t accumulate — little imperfections won’t get magnified by these manipulations.
Eventually this digital pattern is stored, copied, and given to someone else who plays it back. At this point the reverse of the recording stage occurs: the digital pattern is turned back into an analog pattern of electrical impulses, which are used to vibrate the diaphragm of a speaker, which vibrates the air that enters someone’s ear. (The brain then digitizes this pattern into a waterfall of nerve impulses, but that’s another story.)
If we go back and assume instead that we are making an analog recording, then we get to skip the analog-to-digital conversion stage. That’s good, right? But instead, the electrical impulse pattern will be transferred to a storage medium of some kind, such as magnetic tape. So we still have to have a conversion step, and while it’s characteristic imperfections will be different from the analog-to-digital conversion, there will of necessity be imperfections in the process. Further imperfections can be introduced just by playing back the tape. If any conversions are done to the pattern (such as adding reverb after the fact), that involves another playback and recording step, so two more places where entropy can and will add noise to the pattern. Eventually a master recording is assembled and used to cut the vinyl records: another conversion step. Like the tape, the record cannot be played back without being subjected to entropic change, as the needle is dragged across the pattern in the grooves.
I remember back in the 90s you would occasionally see audiophile vinyl recordings that were “direct-to-disk”, which resurrected the record-making process from back when tape hadn’t been invented. It had the advantage of absolutely minimizing the conversion steps (from the microphone directly to the master record cutting head), but of course can get pretty expensive if you wind up needing multiple takes.
(I really hope this hasn’t been a whole lot of verbiage that completely missed the point of your question.)
That was exactly what I meant, breadbox. Thanks. 🙂 These imperfections in the transferring stage were something I hadn’t realized.
This, everyone, is the only argument I have truly considered to be viable. Most of the comments (both here and through various facebook channels) only take the reproduction into account — the fact that a record is beholden to the wires and speakers and dust and such that it plays through. Perhaps I should have added the phrase “In an ideal world,” to the sentence “digital cannot get the true-to-life recreation that analog does.” I clearly point out the drawbacks of playing analog (while the technology itself would suggest that analog is indeed a true-to-life representation when recording, as I said), and to be blunt, most people are ignoring this fact and going along with their own preconceived allegiances.
Let me try again. 20+ years of high-end, fancy studio, analog recording studio experience here.
Record music to a perfectly-maintained, meticulously-calibrated $50,000 analog recorder. Now rewind and press play. It’s NOT AT ALL a “perfect, true-to-life representation.” It’s not even close.
If anything, it’s closer to painting on canvas! People fill up forums with opinions on which machine using which tape at which level at which speed makes the best drum sound, which of course will not give you the best vocal sound. And best rarely means “most accurate” — nobody’s striving for accuracy.
There are recorders that aim for pure, accurate reproduction. But they were only popular in the lab (and since abandoned for more accurate digital systems)!
Musicians hated the way they sound!
But it’s not just rock guys trying to get groovy sounds that evoke Zeppelin. Classical music actually went digital first. Dragging tape across a mechanical record/playback head is imperfect; the varying speed (specified as wow/flutter in the specs) is particularly audible on piano notes, which go out of tune.
Still: an LP is not capable of representing the sound of an analog tape — the specifications are simply different, and not even close. Cassette has different restrictions. The closest is 8-track, but even those were processed because they sounded too different from records, and people liked the “punchy” LP sound.
But again, you’re ignoring the cultural issue which arguably makes most of the nerd/tech stuff moot. The first time I heard Antonio Carlos Jobim it was an LP through a funky old tube console stereo. It was dark, muffled, distorted, and sounded absolutely amazing. The CD doesn’t replicate that experience for me, but it’s undoubtedly closer to what the musicians were hearing when they performed it.
This article, and the subsequent replies, are wonderful. Well done, Smash.
Take 1 “True to life?”
I’m not going to directly answer the question of which format sounds better, why will eventually, hopefully, become clear. 🙂
What I wonder about is the phrase “true-to-life” in your statement: “In an ideal world, digital cannot get the true-to-life recreation that analog does.” What does that mean in the context of recorded music?
As breadbox pointed out, the signal chain cannot be dismissed when considering these questions. Without that chain there is no recording regardless of the recording media. So let’s take a look at the first step in that chain, the microphone.
You could use a dynamic mic like the Sure SM-57. Of course it suffers from a lack of high frequency resolution and has a notable presence peak in the high midrange. Neither of which has stopped it being used on countless recordings. Also of note is its cardioid pickup pattern which means it doesn’t pickup much sound coming in from behind and less from the sides than the front.
You could choose a large diaphragm condenser like the Nuemann U87. Of course those don’t handle high sound pressure levels all that well, and while they are pretty flat in their frequency response they have a little peak and a quick drop off in the high end. A limit that means they don’t record the whole range most humans can hear. They do have a switch that can change the pickup pattern though, so you could switch it to omni to pick up the whole room.
How about a small diaphragm condenser like the Earthworks TC30? It has a dead flat, or at least as flat response as can be reasonably achieved, and the high end doesn’t roll off ‘tll 30K, well above most folks ability to hear, and it has an omni pattern so it picks up sound from all directions. Theoretically this is the most true-to-life of my examples, but despite that it’s not used any where near as often as the SM58 or U87 and not just because it’s expensive.
This is because most recording isn’t really about capturing a true-to-life sound. It’s about creating something that sounds good on playback. Engineers reach for the mic that will give them the best sound on playback given the circumstances they are dealing with.
Recording a singer for a distortion laden punk band? The SM57 or 58 will help punch the vocals through the din better than the TC30. Recording a tender ballad with a string section but the vocalist is a little reedy sounding? The Nuemann U87 will sweeten the sound where the TC30 would just highlight the warts.
And that’s just a small sample of the choices you have with mics and you have similar choices with mic placement, room baffling, compression, EQ, and every thing else between the sound source and recording media. All of which comes before the inevitable manipulations of those sounds in mixing and mastering.
My point here is that the end media used to record is almost never receiving a signal that’s true-to-life in any really meaningful sense. Not unless you’re the Cowboy Junkies recording their Trinity Sessions album live on a single stereo mic that was placed in the simulation of a human head. Or, more practically recording a band live through a stereo mic pair.
I’ve done this using a matched pair of small diaphragm cardioid condensers set up in the ORTF pattern to simulate human hearing. I don’t recommend it to anyone who doesn’t have the patience of a ninja on K. I had to threaten our bass player with gaffer taping the knobs on his amp to get him to stop fiddling between takes.
I never, ever got the mic pair in the perfect sweet spot where everything was clear when it needed to be. This is because that spot moves as the dynamics of the various instruments change between section of the songs. The Cowboy Junkies with their acoustic instruments got around this by having musicians move towards and away from the mic as they played. You can’t do that with JC-120 boat anchor of guitar amp let alone the human sized speaker cabinet with six 10″‘s that our bass player used.
Those recordings are as true-to-life as my budget could afford; the raw track on a flat response set of headphones, which I also own, sounds exactly like what it would have sounded like to have sat where the mic pair was. They sound like crap, not because the media was digital, but because none of the artistic artifice that a real recording engineer possesses was applied to it. They are in fact too true-to-life.
TL;DR The media that something is recorded on is really a non-issue with regards to true-to-life sound in the majority of cases because the recording process is rarely about producing something that is actually true-to-life.
Take 2 to follow: “It’s a Trap!”
Annnnd damn. While I laboured away at my screed, redacted perfectly pipped my point in a far more succinct manner. Ah well, there’s always Take 2, though at this rate someone else will make that point with a delicate Haiku before I get a chance to write it.
Heh, I had a feeling that the hardcore audio nerds would show up sooner or later. “Après moi le déluge.”
It’s not relevant to the discussion at hand, but it is interesting to occasionally remember that analog processes themselves are also, ultimately, digital. Right? Eventually you reach the Planck level; there are only so many different energy states a photon can be in; the magnetic surface of a tape is not perfectly homogenous but is made up of discrete magnetic grains … So what really distinguishes analog from digital is that one, analog is actually digital but with a much much much finer degree of digitization; two, because of quantum physics (not to mention the amount of storage it would require) we cannot access the actual digital
numbers of an analog recording, so we have to copy/manipulate the pattern’s elements en masse using gross, macroscopic processes.
You guys are amazing, and I have learned a lot. I sort of want to risk going off-topic and have you all recommend me what gear I should be looking for in my quest to put my saxophone through guitar pedals (friends are split between a keyboard amp, a guitar amp, a PA, and one of those little amp emulators), but that would be wrong. WRONG.
I’m a little disappointed that in all the fray, nobody jumped down my throat about my Harry Potter swipe.
To follow up Breadbox and the original post, film photography is also ultimately digital. If you look at film (or a print) with powerful magnification, you’ll see it consists of tiny grains of color (or gray for B&W film.) The size ranges from enormous for cheap color print film to amazingly tiny for the late, lamented Kodachrome 25. (The Wikipedia article claims 20 megapixels for a single frame of 35mm Kodachrome, though I think the resolution was highest for K25, and significantly lower for K200, so that figure may be an average.)
Oh, and Smashley, shame on you! Just because something is popular doesn’t mean it’s bad. Have you ever actually read any Harry Potter?
I was hoping redacted or someone with real knowledge was going to answer your amp question Smashley. For the record: I’m a hack when it comes to this stuff, no one in their right mind would pay me to do their sound engineering. Doesn’t stop me from having an opinion though. 🙂
Okay, there’s a few things to consider when choosing amplification. First off you have keep in mind that amps can have two separate functions: projection to the house and monitoring for the musician. In a venue of any size house projection is going to be handled by the house PA so if that’s the only gigs you ever play a guitar or keyboard amp are still options.
On the other hand owning a small PA can open up more venues. Separating those two functions means that you will have less issues with feedback at the volumes needed to fill an audience space.
Speaking of feedback, I’m not sure what problems you’re going to have with a sax running through an amp. It seems to me that because the sax is a loud instrument you’re going to have to run the amp pretty loud in order to hear the effected sounds. If all you want the audience to hear is the effected sound there could serious feedback problems. I don’t really know about this though, as the only times I gigged with horns I had nothing to do with their setup.
My feeling on this is that a PA set up would again help as you could run a monitor mix just loud enough to keep track of what’s going on while still feeding a decent volume out to the house. In the case where there’s a house PA they could even take a line out of your PA and you don’t have to schlepp extra speakers around.
Another thing to consider is impedance. Electric guitars for the most part are high impedance while most mics are low. This means that you’d either need to get a high Z mic or buy a matching transformer in order to run your signal through guitar effects. I’ve never used an impedance transformer that didn’t suck all the character out of a signal, but I’m sure good ones exist so some research would be necessary.
Again, a PA wouldn’t have this problem they’re primarily designed to handle low Z signals. But the catch is you’re then going to have to get one that has an effects loop in order to run your guitar effects. This shouldn’t be a problem, just check to make sure that that loop can handle high Z devices.
I’m not too familiar with keyboard amps. I suspect that they tend to be more flat in their response and I also suspect that they’re more likely to handle a low Z signal. If that’s the case you’re going to again need one that has an effects loop that can handle high Z signals. I’d also look for one with a direct line out so you can plug it into a house PA and keep the actual amp sound to a feedback free monitoring level. If my suspicions are correct and it has all those features this would be the most cost effective solution.
As far as guitar amps go you need to keep in mind that an electric guitar without an amp is useless. They are essentially two parts of a single instrument. Because of this most guitar amps are not in anyway neutral or flat in their sound reproduction. They all have different frequency response graphs to add a different character to the sound of a guitar. This is deliberate, it’s how the various manufacturers distinguish their products.
If you go the guitar amp route be prepared to spend some time sampling different amps to see how the characteristics of that amp effects your sound.
Also keep in mind that many guitar amplifiers are designed to break up into distortion when hit with enough signal. This can be tricky to control and the more effects you have between the source and the amp the harder it is to use that amp distortion consistently.*
From my point of view the best solution, because it’s the most versatile, would be a small PA, followed by the more cost effective keyboard amp and finally a guitar amp if you’re willing to put the time in to find one who’s sound you like.
*Because I like effects including distortion I use a Roland Jazz Chorus amp. It has a good, clean, loud sound that doesn’t break up like other guitar amps.
I built myself an effects switching rig out of the Craig Anderton (SP?) book Electronic Projects for Musicians so I could play around with the signal routing while performing. I always run a dry signal in parallel to the effected one and the effect loop has a volume pedal on it. (Actually, this and everything else is switchable between the two channels but the volume usually stays with the most effected channel.) This allows me to vary the amount of distortion, or whatever, as I play. It also allows me to EQ the two sounds independently for the best results.
It’s also a way to get around the desire for the consistent volume that a lot of sound folk want because makes it easier to mix a house sound. When I want to increase the intensity of emotion I make my sound more distorted rather than increase it’s volume. In the venues I played in, if you used too much dynamic range they were simply going to compress your signal to take care of it. It’s vital though to show the engineer your full on clean and distorted signals and make sure they are within his or her limits for volume fluctuation.
Another great trick with two signal paths is to put a delay on the clean line. What happens is you get the rhythmic pulse of the delay while having a sustained, fluid line over top of it. It can sound like two separate instruments. Fun.
Anyway, I hope that helps and doesn’t make things more confusing and as with everything else if someone with real knowledge of this stuff disagrees with me I’d go with their opinion over mine.
I just miss the emotional attachment of opening the LP, cleaning it, putting it on the turn table and putting the needle down. Then, putting on the headphones and gazing at the album art while listening to the record.
Personally after 30 years of the CD world I grabbed two nice turntables and got into vinyl got a great collection going I still have CD’s but I like vinyl I like the sound everything about it