Lately I have been reading articles and posts by people who say they love vinyl records better than CDs or downloads. Usually they say that is because they prefer analog to digital. In this post I will first talk about the differences in the ways analog and digital sound information is stored, then I will compare the advantages and disadvantages of analog and digital recorded sound. Some elements of this post have been covered briefly in other Audio Penguin articles, but I want to clarify the differences more directly in this article.
The nature of sound.
Sound consists of continuous compressions and rarefactions (the opposite of compressions) of molecules in a medium. We will be concerned with sound moving through the air to strike our ears or, in recording, to strike a microphone.
In most adult human beings we can hear these compressions and rarefactions – which we usually call “sound waves” – when they occur anywhere from about 16 times per second to about 16,000 times per second though some young people can hear sounds up to about 20,000 compressions and rarefactions per second. We sometimes call a single compression and rarefaction a “cycle.” When we mean “cycles per second” these days we usually use the term “Hertz” (named after the 19th century physicist Heinrich Hertz). Since the measurement is named after a person it is capitalized. Thus we can say that the average adult hears sound at frequencies of about 16 Hertz (usually abbreviated Hz) to 16 kHz (‘k” representing “kilo” or 1000). Thus the average adult can hear frequencies up to about 16 kHz with older people not hearing the higher frequencies and very young people hearing even higher than 16kHz.
In modern recording we convert these sound waves into electrical waves for recording. This is the function of a microphone and since the corresponding electrical waves are analogous to the original sound waves in the air, we call this form of sound capture “analog.”
Analog (or Analogue) sound recording.
In analogue recording the electrical signal representing the original sound is captured on some medium such as magnetic tape or an acetate disc. Since the original sound is continuous the analog recording is continuous. On tape the sound signal is converted to varying magnetic flux which can be stored on analog magnetic tape. In playback the varying magnetism is converted back into an electrical signal which then drives loudspeakers or headphones that convert that electrical energy back into sound.
In an analog disc the sound signal is converted into varying grooves. In playback the stylus in the photo cartridge moves in accordance with the variation of the grooves producing an electrical signal which can be amplified and converted to sound using loudspeakers or headphones (which are really just miniature speakers).
In analog recording the complete continuous sound wave is captured and reproduced, though perhaps not always perfectly due to imperfections in the recording and playback gear.
Where the Rot Sets In: ANALOG
Fanatic devotees of vinyl will exclaim how smooth and warm a vinyl disc sounds played on a good turntable with a good cartridge. But there are some defects.
- A vinyl disc is prone to being scratched or picking up dust.
- Every time a vinyl disc is played the grooves are slightly worn, thus shortening the overall life of the disc. (Just buy a disc at a flea market and you may well see this problem.)
- Loud low bass sounds have to be mixed for cutting with those sounds in the center (not extreme right of left in a stereo disc) to keep the playback stylus from jumping out of the groove.
- Longer music has to be cut at a lower volume level on a microgroove record in order to get more time on each side. (Note: Discs produced for dance DJs using loud beats are often very short — often less than 15 minutes a side — to allow for larger grooves at a higher volume and to keep the playback stylus from jumping out of the grove while “scratching.”)
- Purchased analog discs are vinyl but the disc recorded by a cutter is acetate. Usually these are used to make masters for press but the wear problem is worse with acetate discs if played directly from the cutter. Some DJs pay for special one-off beats to create special signature sounds. [Note: Acetate discs are soft acetate on aluminum. In high school I rebuilt a discarded disc cutter and used it to impress my friends with special one-off recordings. This was before cassettes came into use for music and few people had open reel recorders. I learned a lot but it was not all that practical a medium for most use.]
Digital sound recording.
Digital data is always stored as discrete values that represent the settings of a series of non-mechanical electronic switches in a computer that are often referred to by numbers. When I type the letter “s” on my computer it is stored as 115 and the letter “o” is stored as 111. Thus if my computer uses eight switches (often called “bits”) to store each letter, the value 111 might be stored as switch settings 01101111 where 0 represents a switch in the off position and a 1 in the on position. Letters of the alphabet are themselves discrete objects so they are easy to digitize (though there is more than one system in use).
But how about a continuous sound signal? There is no way to directly convert a continuous signal to some form of continuous digital information since digital data always consists of discrete values. Thus a process of sampling is used. The electrical value of several parts of a sound wave are captured and stored as values, just as letters are in word processing. This is known as analog to digital conversion. To play back sound elements stored this way a digital to analog converter is used to look at the discrete samples and reconstruct an approximation of the original analog signal in the same way that a connect-the-dots drawing is created by connecting the space between the dots when creating the finished drawing. Thus in digital recording the sound output is never an exact copy of the original analog signal.
If one converts a drawing of a dog into a connect-the-dots puzzle of 20 dots the finished drawing after connecting the dots will be very crude. But if one converts the drawing of a dog into 6000 dots the resulting drawing, after connecting the dogs, will be much smoother and closer to the original. This same concept is applied to digital sound recording.
The sampling rate, the number of digital “snapshots” taken of small portions of a sound wave (converted to an electrical wave by a microphone), determines the highest frequency that can be recorded. Usually that means the sampling rate must be at least twice the frequency of the highest sound. The CD standard dictates a sampling rate of 44,100 samples per second. Since almost nobody can hear sounds having a frequency of 22,050 Hz that would seem to be all one needs. Clearly that was the thinking of the inventors (Sony and Philips) of the CD. We will see later (“Where the rot sets in”) that there are reasons to use a higher sampling rate for some music recording and reproduction, and some lower quality apps (like voice mail for phone calls) can use lower rates.
Bit depth, or resolution, is an important factor in digital recording. The CD standard dictates that the value of each sample (i.e. the loudness) must be expressed by a whole number between 1 and 65,536. Thus each sample can be stored in a 16 bit “word.” This generally makes for a smooth sound, even for fadeouts and soft reverberation. But there are reasons, which we will see, to use a higher resolution in recording and convert, if necessary, for a CD. If each sample is stored in a 24 bit “word” then the loudness range, in whole integers, can theoretically be 1 to 16,777,216.
This may seem excessive, and is probably more than necessary, but the real benefits of 24 bit resolution show up when one records, say, a concert and isn’t sure what the final audio level will be. If one feeds a signal into a digital recorder that is too high, it will seriously distort (worse than with an analog recorder). If one records a 16 bit file but keeps the volume low to be sure of not over-driving the recorder, then the final recording will have far fewer “steps” than a CD is capable of. One can expand the loudness but then the space between each step, in terms of loudness, will be exaggerated. By recording 24 bit one has plenty of loudness “steps” to use when converting to CD standards. Also, note that the sound on most DVD movies has a sampling rate of 48,000 (48K) and generally 24 bit resolution. The world is changing. I often record classical concerts at 96K sampling rate and 24 bit resolution. I use very good “dithering” software to make CD compatible 44.1K, 16 bit files for CD, but I still have a master that is better than CD quality.
There is still another digital sound format that some people, like me, think sounds more like analog. It is DSD, Direct Stream Digital. This is a file type used in SACDs (Super Audio Compact Discs) invented by Sony. But Sony is a very money-oriented company. They seem to resent the move from vinyl records (which the consumer could not easily duplicate) to cassette and later recordable CDs. They felt “wronged,” it appears, by the fact that a consumer could duplicate a recording at low cost. I can see their frustration over lost sales of recorded music, but they seemed to ignore the fact that cassettes, and later recordable CDs, have been used by millions of musicians and sound producers to make demos and small run distributions of their own work. Thus when they introduced SACD they designed the format so a person could not record a one-off SACD, they had to have a glass master made and hundreds or thousands pressed. There is nothing in the DSD standard itself which prevents duplication. Sony simply made it so DSD recordings that were not produced as SACDs would not play in a SACD player. Bad move. They essentially killed their own format (though one can sometimes still download DSD recordings for playback on a computer or certain very high end portable music players).
Where the Rot Sets Sets In: DIGITAL
- The biggest issue is that since a digital recording is made up of samples that are then used to construct an analog signal used for playback, the final signal is never exactly the same as the analog original (coming from the mics or mixer). Purists claim this is why digital cannot be as good as analog though I think in many cases they are reacting to flaws in the original analog to digital converter (especially in older CDs) or in the digital to analog converter in their own CD player system. There is a reason why some audiophiles pay several thousand dollars for a special digital to analog convertor while other users play less than $100 for the whole CD player. Also, digital-to-analog and analog-to-digital convertors are not perfect devices, so sometimes they introduce other artifacts.
- The next major issue is that over 50% of music purchasers these days do not buy physical media but simply stream or download digital files. With a small number of exceptions these are data reduced audio files with as much as 80% of the original data deleted so a smaller file can be downloaded fast or many songs stored on a phone. Some people say they cannot tell the difference between full quality sounds and data compressed mp3 or AAC files. I can, but maybe others can’t. However if those compressed files are later loaded into an audio editor for some reason, then re-compressed, the sonic errors compound.
- Sometimes a person buys a CD of a recording that was made, and originally released, as a analog recording (vinyl or cassette). Then they say how much better the recording is because it was analog. Once an analog tape is copied to digital it is simply a digital recording, with all of the advantages and defects of that medium. If the sound seems “warmer” it is usually not due to the original being analog, but a result of the microphones, mixer, etc. used in the original. Modern digital recordings can be made directly with warm sounding gear (such as vacuum tube mic preamps, a hand built analog console, etc.)
It is probably obvious in this post that I support digital recording though I have considerable experience with both. The recording gear can be made smaller (great for live concert recordings), there is little generational loss in duplication, editing is much simpler, etc. And if one uses high end gear the result can also be very good. Sometimes a musician will go out and buy a handheld portable digital recorder with two mics that costs maybe $200-$400 and records his/her concert. Then they might say the digital recording is not as good as an analog recording that may have been made of them in a concert in music school years before. But the difference is often the quality level of the gear. Small handheld recorders are very useful for many purposes (I own three) but for serious music recording, serious gear will produce better results. When I record a classical concert using only two microphones, I use a Sound Devices digital recorder or a DSD recorder made by Korg and either a microphone pair of Schoeps mics or a pair of Danish Pro Audio microphones. These are in the $4000-$6000 range. That is the difference between strictly professional gear and consumer gear. You use the gear that is required for a specific job (and that you have access to).
If you want to learn more about digital recording technology and don’t want complicated math I suggest the following site: