First things first: Analog and Digital
I took a break from tax prep to research and purchase a solid-state digital recorder (and a microphone and some cables). I had narrowed my choice down, but at the last minute, I veered in another direction and bought a digital recorder with a 20 GB hard disk.
Throughout that long afternoon of research, I felt as though I was EveryShopper, caught in the tug of war between the eminently reasonable “Just tell me what I need to get!” and the ever-hateful “Well, it depends on what you want to do.” (ever-hateful, but correct, dangit!) As I weighed this feature and that aspect, I thought, “Ah! I’ll write an article about making this decision and put it on the website.”
But once I began scribbling the permutations of feature A versus feature R that were a part of my decision-making, I realized that I would first need to write about portable recorders in general. No. Wait a second. While we’re getting into first-things-first, I think I’d best define “analog” and “digital” and get that out of the way. (Plus, I’m thinking I’ll probably need to create a glossary for this site.) I’ve spent the last couple of years researching audio matters. I find myself wishing it were simpler than it is. I hope that this attempt at explanation succeeds at making it a bit more straightforward. So here’s Purchase Decision, part 1: Analog and Digital.
Analog and Digital
What’s all this “analog” or “digital” people keep talking about? You’re apt to hear them in statements like these:
Blah blah blah analog audio connection.
Blah blah blah, but even so, you can transfer it to the computer by analog method.
Blah blah blah digital storage and transfer.
Blah blah blah audio is created as digital.
Blah blah blah, the digital audio’s better when sampled at a higher rate.
Why in blazes should it even matter to you to know what digital and analog are? Well, remember. You’re recording Grandpa’s voice and putting it into digital form, so that you can make tons of copies of it and each one sounds as good as the first. One hundred years from now, there will be descendants who will wonder what life was like back when, and they will be able to hear this recording and listen and be enlightened and ponder and reflect. So the goal of this whole exercise is to somehow get the sound of Grandpa’s voice into digital form. Which implies, of course, that Grandpa’s voice does not start out digital, but something else.
When you sit in the same room with Grandpa, his voice is transmitted through the air in sound pressure waves. The first era of audio technology was devoted to capturing that sound and storing it for later playback (think Alexander Graham Bell, think Thomas Edison, think of early gramophones, of the first tape recordings, and the first phonograph albums). A microphone translates those sound pressure waves into an electric signal, which is subsequently translated to a storage device (minute vibrations of a needle into a phonograph record, tiny charged particles on surface of magnetic tape). Later, the recording is played and the storage medium offers up its vibrations or magnetic charge to be translated back into an electric signal, which is amplified and driven through speakers (or headphones), which vibrate to create sound pressure waves that travel through the air, and you hear the sound of Grandpa’s voice.
The process I just described is analog. The word analog is closely related to analogy and analogous to: Two things are not the same, but are related to one another. The sound pressure wave pattern of Grandpa’s voice is not the same thing as the electrical signal, which is not the same thing as either the set of tiny grooves carved by a vibrating stylus on a phonograph album, nor the pattern of magnetically charged particles on a strip of audio tape. One is not the same thing as any of the others, but each one is closely related, or analogous to, the others. This is analog in analog audio.
Now, about that analog electric signal that travels from one piece of audio equipment to another: it can be expressed as a wave form. You might have seen other, similar, analog wave forms: on a seismograph, on an oscilloscope, at the beginning of both the movie Fantasia and the TV show The Outer Limits (assuming you’re that old, or that you have a thing for classic 60s television shows), on an electrocardiogram (EKG) that measures the body’s signal directing the heart to beat. That wave pattern is what the analog electrical signal looks like. (though for the seismograph, the wave that creates that pattern is not electrical, but a measurement of the physical movement of the earth’s crust.)
To understand how an analog wave signal becomes digital—- and thereby understand what digital audio is, we’ll keep thinking in terms of seismographs and electrocardiographs. Both of the instruments for measuring those wave forms have a spool of paper and a pen that draws a squiggly line on paper. (Here’s a page with a short QuickTime movie of a seismograph, though the result isn’t drawn on paper.)
A roll of paper spools out continuously. If there’s nothing to measure, the pen (on a moving needle) draws a straight line down the middle of the paper. As soon as an earthquake hits or as soon as the machine picks up the heartbeat signal, the pen wiggles back and forth, drawing wave forms that reflect the movements of the earth’s crust or the signals telling the heart to beat. The paper moves at a continuous rate, showing how the signal changes over time. This wavy squiggle is also analog (the movement of the pen is not the same thing as the signal directing the heart to beat, nor the actual movement of the earth’s crust, but the wiggly pen movement and resulting waveform is analogous to each.)
Suppose you were to mark points on the graph paper where the wavy line intersects the cross-points of the grid. The horizontal axis of this grid is time. Think of each vertical line as a tiny metronome, marking off time in short, regular intervals. You will need to plot a point at each time interval.
In addition, this grid forces its exactness on you. You must use the grid’s vertical scale to measure the height and depth of the line (which is also called its amplitude). Therefore, your points must be at the intersection on the grid that’s closest to the location of the wavy line. This process of marking off points at regular intervals and at regular heights is called sampling.
Since each of the axes of the grid have numeric scale to them, you can represent each of those points by a number. This set of plot-points and the numbers—or digits—used to represent them, is a digital waveform. If the waveform started out as audio (the ones in this set of illustrations did), then what you’re looking at is digitally sampled audio. Take away the analog waveform and leave only the plotted points, and you’ve got a graphic representation of digital sample.
Here is the final “digital file” for this wave. If you were to make a statement at the outset that declared the time value (in this case 50 samples per second), then you’d have made a shorthand statement telling what the horizontal axis is. Now all you need are numbers to represent the height (or amplitude) for the waveform. And here they are:
What you’ve just done (in a crude form) is sampled an analog waveform and created a digital file of it. And it is very crude. There are sampling errors (see where the points miss portions of the wave?). It could be more accurate if you were to take more samples more often, and if the vertical measurements were made in smaller increments. In other words, if the graph had more lines, you’d get a more accurate sample of the analog waveform.
In the illustrated example, the vertical (or amplitude) measurements on the graph are very arbitrary; +10 to -10. But for a digital sample that works in computer-ese, it’s not a matter of tens or sets of tens. Computers work on binary, which boils down to ones and zeroes. (If you find this “ones and zeroes” stuff vague or confusing, I heartily recommend this essay by Rick Garlikov, who taught a 3rd grade class all about binary numbering using the Socratic Method—-that is, only by asking them questions. It’s brilliant stuff and worth reading in its own right.) The vertical axis of the graph is expressed in binary numbering system, starting with 2 (also known as 1-bit), doubling it to 4 (2-bit), doubling it again to 8 (3-bit), and continuing to double the numbers higher and higher. This image shows how the vertical axis would look for 1-bit up to 8-bit, where there are 256 lines on the graph. Showing 9-bit and higher would be too hard; it would require a large image size to show all the lines in the graph. But this should give you an idea, and also show you what’s being measured when people say 8-bit (256) or 16-bit (65,536) or 24-bit (16,777,216).
These graphs and their arbitrary scale have (I hope) illustrated What It All Means. But what would a graph look like for digital audio that’s Actually Being Used? Back in 1980, when Philips and Sony published the Compact Disc Digital Audio standard (also known as Red Book Audio), they declared this for the sample rate: Audio is sampled at 44,100 samples per second. The sound wave’s height (or amplitude) is measured on a vertical scale that has 65,536 lines on it. In shorthand, 44,100 samples per second is expressed as 44.1 kilohertz or 44.1 kHz (hertz means one per second, and kilo = 1000). And 65,536 is 16-bit. So whenever you hear someone referring to Audio CD, or 16-bit 44.1 Khz digital audio, that is what they mean: The analog audio signal is sampled 44,100 times per second, and the signal wave has 65-thousand height increments. That level of detail makes it possible to faithfully reproduce the New York Philharmonic, John Coltrane’s saxophone, Sgt. Pepper’s Lonely Hearts Club band, and the sound of Grandpa’s voice.
Now, instead of looking at hypothetical graph-like sound waves, here’s is the image of a soundwave created at Audio CD resolution. I created it by saying, “Hello, world!” into the microphone of that lovely new field recording setup that I just bought. (The original was in stereo, but I looked at only one channel) In this illustration, I examine an eentsy slice of the wave, magnifying it more and more until it reaches the point where you can recognize the individual samples.
You won’t find any graph lines here. But you can see how deeply we need to zoom into the Hello World waveform to reach the magnification where you can see individual sample points.
Getting ready to move on
Digital Audio is mostly a storage format. In day-to-day use spanning the last couple of decades, the Audio CD has replaced the phonograph album and cassette tape as the preferred method to store the audio recording. (For the moment, I’ll sidestep more recent developments in compressed audio—-including MP3—-and all the permutations of portable storage and playback.) Here’s how digital storage fits in with the capture and playback of the sound of Grandpa’s voice:
From Grandpa’s voice to microphone to electric signal to digital sample to digital file. The digital file is stored; let’s say it’s stored as an Audio CD. The CD is played back, and the CD player converts the set of numbers to an audio electric signal which is amplified to drive speakers or headphones and converted back to sound pressure waves, at which point, you hear the sound.
The next question to consider is what happens in between the step where you and Grandpa and the microphone are all together in the same room, and the step where the audio signal becomes a digital file. That in-between step calls for a discussion of different portable recorder options, which I will discuss in the next article. (and then—-then! I’ll tell you about the twists and turns of equipment decisions)
I’m so excited about your site and the issues that you are/will be tackling!!!
I have been recording my parents onto a digital voice recorder but have tons of questions about the quality of that recorder and what to do with the files - I’ld love to do some editing and to add some music on another track.
Look forward to future additions.
What kind of digital voice recorder?
And how’s it been, recording your parents? What’s the experience been like?
Yes, I have ideas about what to do with the recording. First off is to convert it (if necessary, which, for digital voice recorders it probably is) to an uncompressed format, and, at the very least, create an audio CD of it.
i have been looking for an simple expalnation about what is a analog signal and why do you have to convert it into a digital signal? when they do the same job. Your article has helped me immensely.
Thanks a lot
Ms. Kitchens, You’ve got yourself a new gal fan here! Not to get off your subject, but I wanted to let you know what drew me here. And that was the little tid bit you posted on the Scobleizer Blogger about a memory with your mom. What a heart warming and endearing tale it was. Gave me the biggest smile, and filled my eyes with the cheerful sort of tears. I read it over twice. Just so cute, and wonderfully told, that I had to get over to ya and tell you for myself! And I am glad I did. Because there is ever more great reading here! One heck of a noodle up there in that Kitchen! All by accident, you got me hooked now! Take care. - Rene’e
Thanks, Renée, for coming by and commenting.. and all the more for telling me why. That’s incredibly important. Stories have power ... I’m starting to cotton on to that fact. It’s what got me into family history. I don’t have kids, so it’s not a “pass on to the next generation” kind of thing for me, as I imagine it is for others. It’s the stories: Where do I come from? What happened? And there’s a variation, too—What the hell happened? Who are these people and what did they do? What were their challenges? How did they get through them?
But for a web site getting off the ground, I’m focusing on stories. Because boy oh boy (or girl oh girl) are they ever compelling….
The timing of your comment is especially wonderful: I just now finished reading Malcolm Gladwell’s The Tipping Point. Which is about how movements—fads—epidemics get started. Gladwell discusses What sets them off. Your timing, Renée, is impeccable. How did you know? You’re a genius! :D
What sort of equipment are you using for transferring analog recordings to the computer? I bought something that was supposed to do that… but I don’t get the squiggly lines.
I also have some interviews on reel to reel tapes…can they be transferred in the same way (assuming that I can find a reel-to-rell recorder that works)?
I’ve tried to interview the oldest person at any family get-together, so I have a variety of stories. Many are the same story - but from differing viewpoints. My 93-year-old mother is always ready to tell some more, so I’d like to be more efficient with the process.
Betty Jo, thanks for stopping by. When I do analog conversions, I connect the audio device to the computer using a stereo-mini to stereo-mini plug—one end goes into the headphone jack of the device (the recorder) and the other goes into the Line In port of the computer.
(I totally want to write about this in more detail on the site. My perpetually coming soon section, entitled “How To,” is wanting some love and attention in a big way)
If you’re connecting from a stereo system or a stereo-system component, then you probably need to have an RCA stereo to stereo miniplug connector cord. (RCA stereo is that double-headed set of plugs, one of which is usually colored red)
For software, I use Audacity. The lack of “squiggly” lines means that the signal isn’t strong enough… which may be for a variety of reasons:
The recording was too soft to begin with (you can amplify the signal using software)
If you’re working with a portable recorder, you can adjust the playback volume.
Again, I want to write about all of this in more detail here. And even provide a demonstration movie where you can watch how it’s done.
The explanations are all on target-“Bull’s eye” except for this.
1)From microphone output the electric signal,(I presume) travels to the Next Electric component say, a particular IC or transistor or for that matter some diode or triode of solid state grid comprising several of these triode/diodes. 2)Am I right so far? Because this diode/triode business is my supposition.
3)Then the signal output from this particular IC happens only when “switched” -Right?
4)I am totally hazy from this stage of signal processing.
So pl reply to help me understand