Saturday, January 28, 2012


A degree of simplification is often the key to grasping a concept. It is perhaps a lot like the argument about learning to swim. Swimming, experts will tell you, is the art of learning and mastering the many strokes. But the knowledgeable will also add that the first step is to learn to be comfortable in the water and then learn to float. Jumping in at the deep end and launching into the standard beginner's stroke is likely to earn you a stomach-ful of water and a holy terror of swimming.

Audio is at best tricky waters, and learning to maintain your balance in a strange new medium is the key to expertise and enjoyment of the many delights of sound reproduction. May I urge the 'pundits' to close their eyes, with an indulgent smile on their faces, as the newbies are invited to a frolic in the shallows of audio in the following discussion.

Sound propagation --  just 're-imagine' the concentric circles as concentric spherical waves in  3-D space

Anybody who, as a child, had thrown a stone into the still waters of a pond knows how sound travels in the air in ripples or wavelets. (If you haven't yet done that, it is time you hurried to do that experiment!) On a rainy day, the many raindrops that cause a series of ripples on the surface of the water demonstrates the interaction of the wavelets. If you have disturbed the surface of water in a large container, then you are sure to have seen the reflection of the wavelets from the edges of the container, and a complex interaction with the original wave. It is a fascinating sight.

Now perhaps the first reminder for the sound enthusiast is that sound is not confined to the two dimensions of the surface of water or that of a sheet of paper on which the wavelets could be drawn. From the source it spreads all around in the three dimensional space in a spherical pattern. Imagine soap bubbles, one within the other, expanding as they are blown, when you think of the expanding wavelets of sound around the source. The main thing here is to imagine that this is what happens when a bird tweets, when somebody speaks/sings, when a musical instrument is played, when a cracker is burst etc.
( the loudspeaker here is idealized as a source )

This spherical spreading out of the wavelets of sound occurs because most often the source of the sound has smaller dimensions than the wavelengths it produces. However, in a real room or hall, the propagation is more hemi-spherical as is illustrated with this idealized speaker output. (Idealized because in real life speakers do 'beam' a lot, and do not always launch the sound equally well around them.)
(idealized source)

Even a grand piano, when played can be heard with equal fidelity from any angle around it, though musicians tell us that concert pianos are meant to be played with the 'lid' up and the audience facing the reflecting lid. However, inside a room or a hall, the immediate reflections also contribute to the feeling of hearing it 'fully' from any angle.

The second reminder for the audio enthusiast, and especially the one who is more interested in the reproduction of 'canned' (recorded) acoustic events, is the fact that reflections are many even in a large auditorium, to speak nothing of the average living room. It is lucky that we do not get confused in the midst of a medley of reflections, thanks to the ear-brain combo 'processing engine'. A sobering demonstration would be to push your finger into one of the ears while listening, and immediately the 'clarity' of the sound image collapses into a confusing and irritating mayhem. Now be careful to not jump to the conclusion that this is the magic of having two ears and so stereo (stereophonic sound reproduction) is the answer to all the ills of perception. Sadly it is not, and it is a complex issue that has not been fully tackled. Just understand that reflections are a part of the real auditory scene as we know the sound wavelets propagate all around the source; the question is what are the 'needed' reflections and what are the unwanted reflections, and do we have any control over them while recording and reproducing acoustical events. Your understanding will grow as you progress with many aspects of the audio art.

The third set of reminders have got to do with frequency, tone and timbre and perceived realism. Though elephants are known to communicate over vast distances using sub-sonic frequecies, the human range of hearing, by common consent of experts, is limited within 20 Hz to 20,000 Hz. The unit is the Hertz and it is one cycle per second, and a cycle is a comlete vibration from the rest position to and fro, initiating a compression and rarefaction of the medium (air, in the case of speakers, which are easier to understand). Here are representations of a tuning fork vibrating to create the compressions and rarefactions of the spreading sound waves.

Tuning forks and sound wave representations

 When a speaker cone reproduces say, 50 Hz, it is vibrating back and forth fifty times a second. It will be educative viewing the cone in the light of a neon lamp or a fluorescent light with some flicker and varying the frequency up and down a bit. Often say, when reproducing a drumbeat, when the loudspeaker cone jumps out at the first transient, many think that this is what produces the 'thump'. Yes and no. The jumping forward of the cone creates the high amplitude sound wave, but the frequency and character of the sound per se is created by the vibration of the cone which is not visible to the eye. Remember. movies 'move' because of persistence of vision, and anything more than 10 Hz (no, not the sound, but the movement of the speaker cone!) is difficult to see.
Frequency and wavelength relation

Now speaking of the frequency range of human hearing, don't be in a hurry to swallow all that about 20-20,000 in a simplistic manner, and try to listen for the extremes of the range. Extreme low frequencies may be aplenty in the home theatre stuff, but they are very rare in real life, except perhaps when there is a thunderstorm or when you are near a huge waterfall or when the sea is stormy-or maybe near a fireworks display! As for the high frequencies, they are again only the modicum of 'garnishing' that gives character to the real world sounds. An extreme dose of HF, as is put out by a modern loudspeaker driven by an amplifier with crazily set tone controls, playing an 'unnaturally recorded' track, easily brings on listening fatigue and ear damage in a short while. So much to discourage you from joining the 'tish-boom' brigade. The key here, as it is always in audio reproduction, is to try and get as close to 'natural' as possible.

It is time for us to familiarize ourselves with the frequencies--and how they sound. The Web has today given the layman many advances tools and facilities to advance the knowledge of his hobby. Here is one of the many links ( ) that lets you download free samples of audio files at various frequencies; download and play them in the computer itself or write them onto a CD, and remember NOT to convert them into low-fi .mp3 files, but to preserve them in the original .wav format itself. It would really "open your ears" as you start listening and discovering many new things for yourself. For example, how difficult it is for you to hear very much above 10,000 Hz as you become older. So take all that you read about with a large pinch of salt and listen and learn for yourself. Remember not to drive the amplifier and speakers with high volume levels of extreme low and high frequencies as it is not good for their health and also for that of your ears. Also remember though you might think yourself to be familiar with the various frequencies, it is not easy to 'remember' them and compare them with the component frequencies of real life sounds--it would take years of practice to have a discerning ear like a trained musician. You might often have seen and heard musicians using pitch pipes and sometimes tuning forks while tuning their instruments. But it is not practically possible to hear even one pure note in isolation in natural sounds.

Take a look at the frequency distribution of the various musical instruments and the human voice, the male and the female.
The audible sound range

The pipe organ, the King of instruments or the Master instrument, covers the full range of audibility (please also note that the larger organs can go way below 20 Hz!), while the concert piano comes a close second. The male and female voices have a very limited range. Researchers of the early 20th C, while studying telephone circuits have concluded that in order to have intellible communication, the frequency range needed is even smaller. The sounds that you encounter in Nature too have a somewhat limited range. Today, after considering many factors, one could say that a reasonable degree of fidelity could be had within the range of say 60 Hz to about 15,000 or 16,000 Hz. Fidelity, as you will soon discover, depends not on frequency response alone.
A sample frequency response curve

That brings us to frequency response, a term bandied about by audio enthusiasts. Is it just the range of frequencies that your amplifier or tape recorder or CD player or speaker is capable of handling? Hardly. To give it its full name, it is actually frequency/amplitude response. And when you say that your amplifier has a flat frequency response, what you mean is that it is capable of handling the specified range of frequencies without altering the relative levels of the frequencies. Suppose it is fed a signal with a freq at 200 Hz of an arbitrary level, along with another at 2,000 Hz half as loud, and also a third at 8,000 Hz with say one-fourth the level of the first, though the amplifier might be called on to raise the overall levels to drive a speaker, the relative levels of the three signal would remain the same, provided the amplifier has a flat response. In other words, an amplifier or other audio component, should preserve the loudness relationships between various instruments and voices in the input signal and should not over- or under-emphasize any frequency or tone. This then is known as flat frequency response.

But then, remember, there is no ideal amplifier capable of doing such a precise job and amplifiers are usually rated to have a "flat frequency response" within say plus and minus a small figure, usually expressed in Decibels (dB). Decibels indicate ratios of voltages and powers, and it is not easy for the layman to have a non-math understanding of unit. The reader is sure to be familiar with units like the Volt, Ohm, Ampere, Tesla etc (each honouring a great scientist), and the Bel is a unit of measurement that honours Alexander Graham Bell. It is a large unit and one-tenth of that is the decibel, called a "deebee" and written as dB. You are likely to find the dB a lot in the specifications of audio equipment and it is easy to remember a few things about the dB.

The sensation of loudness is detected as a logarithmic function of  the sound pressure levels at our ears, and the dB scale indicated that easily. A difference of 1 dB is taken as a minimum change in volume/loudness detectable by ear, while 3 dB is a moderate change. A difference of 10 dB means a doubling of volume or loudness. By convention, 0 db is the threshold of hearing. Other examples inlcude a soft whisper at about 15-25 dB, general background noise at about 35 dB, noise levels inside a home or office is around 40-60 dB, a normal speaking voice goes up to 65-70 dB. The climax of a Western orchetra is known to reach about 105 dB, while rock band easily top 120 dB. There is the onset of pain and loss of hearing from them onwards, and jet aircraft are known to be as loud as 140-180 dB, with an "unhealthy mix" of frequencies at the upper and lower registers. And while on the topic of loudness levels, a good reminder is that a 4 W amplifier can easily sound twice as loud as a 2 W one, other things being equal, while it would take a 100 W amplifier to sound twice as loud as a 10 W one, and as you move up the ratio of loudness, the figures soon become ridiculous and dangerous to your ears!

The average "hi-fi" amplifier claims to have a frequency response that is flat to within plus/minus 3 dB. That is, with possible wild variations, a possible change of 6 dB and it is not a moderate amount by any measure. To understand why this kind of imprecise response could play havoc with fidelity, one has to take a look at real world sounds. The natural world presents us with hardly any pure tones. Every natural 'tone' is a mix of a basic frequency, and its 'overtones' that are multiples of the base frequency. Take the same musical note coming from two different violins or from the mouths of two trained singers able to precisely vocalize the same musical note, your analysis will tell you that the ratio of the overtones and their nature will be slightly different. This is what gives 'character' or 'timbre' to the sound in real life. Musicians can easily distinguish the sound of many particular similar instruments as their ears are trained to recognize the subtleties of the timbral differences relating to the overtones and their levels and ratios. Please note that timbre is defined as the 'quality' of a sound that distinguishes it from other sounds of the same pitch and volume.

Imagine an amplifier that has a wildly fluctuating frequency response curve (very common in the real world!), though it is still within the +/- 3 dB range in its specification, and so qualifies as a moderately good hi-fi instrument. When this amp is fed with a real life sound, there is every  chance that, if the vagaries of response are in the critical middle frequencies (approx: 2 kHz to 5 kHz) particularly where the ear is most sensitive, the output that emerges from it will have altered the timbral quality; that is to say, the relative levels of the overtones are altered, and the signal sounds like "something else"-- and fidelity is lost. To sum up, it is not the 'correct' specifications that can hide the 'truth' very much that matter, but rather, measurements that will certify that an amp has an over-all smoothness of FR that is more important from the angle of fidelity. A timely reminder here is that the same criteria could be applied to every component in the audio chain.

[ More to follow ]